> If we had cheap clean limitless power, what would we use it for? Someone asked this question in a presentation, and the cheap clean limitless power advocates didn’t have a canned answer ready to go. They eventually proposed things like improved public transit, supersonic flight, carbon capture, AI compute, geoengineering to prevent hurricanes, and city-wide air filters (really).
Seriously? No one mentioned the obvious? The thing that would immeasurably improve quality of life for whole civilizations, if only it weren't so energy-costly?
De. Sal. I. Na. Tion.
If you can make water purification cheap and abundant, at least a dozen major problems just vanish in a puff of vapor.
If you want to use solar power to remove CO2 from the air and fix it in a more solid form, you don't need high technology or cheap power to do that. Just plant a tree! Or a few thousand trees.
On a related note, the biggest contributors to the "urban heat island" effect are asphalt (turning cities almost literally into Black Bodies) and the lack of vegetation. Adding a few shade trees can reduce the scorching temperatures of a hot summer sun by more than 40 degrees F.
"Efficiency" necessarily takes expense into account. I think this is one of those situations where trying to get everything at once is the wrong strategy. Imagine if we damaged our ability to fight global warming because in order to get one billion pieces of fruit. We should let private industry create as much wood and fruit as there's demand for (which they're currently doing, in orchards and timber farms), but also pay for carbon removal at some socially meaningful rate (which would incentivize private industry to capture carbon directly, and then the industry could choose between trees and machines, and if I'm right and machines are more efficient even when counting the extra profit from fruit, they would choose the machines).
Perhaps. I guess some of it is just my own pro-fruit and pro-shade biases peeking through.
When my wife and I bought a house a while back, one of the first things we did, before the previous owners had even moved out, was go to a local nursery and buy a young fruit tree. We waited until they were gone to actually take delivery and plant it in the yard, of course, but that's how important it is to me.
This only works long term if you then cut down and bury the trees deep under the earth, otherwise the carbon is released when the tree dies and decays. Of course, cutting down, transporting, and digging mines to bury the trees is going to take a lot of work with heavy machinery that probably runs on fossil fuels, and will require a massive amount of land dedicated to the task since trees grow slowly relative to our rate of emissions. If your high tech carbon machine can make something like graphite bricks, that would be much easier to deal with logistically, and possibly even useful for some purpose, and is relatively stable so even if it ends up in a landfill won't just turn back into atmospheric carbon. I'll also note we already farm trees for wood and paper, most of than ends up in a landfill and decays at some point or gets burned, re-releasing all the carbon.
That all depends on just how long of a long term you're talking about. Building long-duration objects like houses and furniture out of wood does a pretty good job of carbon-sequestration for a decent length of time.
And trees lower surrounding temperatures. Nothing like photosynthesis to absorb energy! Just walk from barefoot from a grass lawn over to an astroturf lawn, and you'll feel the difference! The grass lawn remains coolish in the afternoon heat, while the astroturf will burn your feet.
Actually, how good and how expensive are purpose-built carbon-removal machines right now? (As Bob Frank says, there are other factors to take into account, but I'd like to get the right order of magnitude.) Compare them to, I don't know, a particularly efficient tree (or bush or weed or what have you).
To get back to the pre industrial revolution levels of CO2 we'd need to remove 1e15 tons of CO2 from the atmosphere, so that's about a quadrillion dollars. To be fair your number is probably for mass of carbon rather than mass of carbon dioxide so let's call it 300 trillion dollars.
It sounds like an untenable solution, in that you can throw an unfeasibly large slice of global GDP at the problem and still fail to make a meaningful dent in it.
Grinding up CO2-absorbing rock, and dumping it in some warm part of the ocean to speed things along, could be done with current tech at a cost two orders of magnitude lower: https://worksinprogress.co/issue/olivine-weathering/
They might be eventually, but they are not cheaper yet. We're going to need both, for a lot of reasons (industrial point source capture, higher output per acre, ability to site capture where we need the carbon for utilization, and on the other side of course we still need wood and fruit).
Unfortunately, a lot of what gets sold as “reforestation” is actually trying to force a forest to exist on land that people didn’t find useful, and which can’t support trees without lots and lots of inputs and attention.
Certainly when it goes by the name of “tree planting”, which has a nice ribbon cutting, and then no one pays attention when the trees all die.
I do volunteer work with a wildlife trust (I know its not very effective altruism) and its sad how true this is. Just this week I was at a session where we were mostly undoing the harm done by a well meaning landowner trying to reforest land poorly. Boring charities like that which just do tree planting and maintenance as part of their regular operations rather than "adopt a tree bring your kids to plant a forest!!" seem to be the only ones who actually do tree planting effectively. My trust manages about a 95% survival rate with the only maintenance after the planting being removing the tree guards.
99% of that problem is planting the wrong trees in the wrong soil. Most EU for example has a fetish-like fixation on growing vast spans of pine, on soils that do not support pine at all, and for variety, force poplar, linden and decorative apple-trees in the middle of a city where they instantly get sick and die.
Meanwhile, the ecosystem of continental Europe is mostly geared towards various types of oak.
It's not that they're less efficient. It's that there isn't room for the needed amount of trees. (We can't even keep the ones we still have.)
I've got no idea whether they're less efficient, or even how you would measure efficiency in that context. Personally I find just having trees around is a net benefit. So by that measure trees are extremely efficient.
We is humanity. And I'm not sure the US is reforesting at all. I see lots of forests of dead and dying pine trees. Canada isn't the only country that's experiencing a increase in forest fires. (OTOH, it's been about a decade since I rode along 5 and saw all the dead and dying pine(?) trees. (I assume they're pine, because they're some sort of conifer, pine trees are common in the region, and I'd just been reading about the pine beetle infestation. But the exact species is irrelevant to the point. The area was still called "forest", but about half the trees looked dead or dying.)
I once calculated that the entire world’s net carbon emissions could be neutralized by planting paulownia trees in an area the size of Ukraine and burying all the resultant wood. This would be considerably less expensive than the dozens of trillions people want to spend on other methods, and not require hypothetical future tech.
It would be even simpler to plant the correct species of trees in the correct environment and let it do its thing. Sure, paulownia is the most efficient carbon grabber, but its not the easiest to grow on any type of soil. The beat solution is just to replant the logged areas with the types of trees that used to grow there before humans cut them down.
Direct air capture costs $500/ton now, and the EPA SSC at a 5% discount rate is $12/ton. It doesn’t make any sense to use discount rates lower than nominal gdp growth, because 1. any public good that increases productivity by a fixed amount forever would have infinite value. 2. Other investments can easily beat the rate of nominal gdp growth.
perhaps, but an average tree absorbs approximately 22 kilograms of carbon dioxide from the atmosphere each year, so a copse of trees can make up for one car. If you add all the other economic, social, psychological and environmental benefits of trees, there is no excuse why we are not spamming saplings and seeds at every available surface.
I am not a botanist/ecologist, is it true that there is tons of land where trees would naturally flourish without further human intervention if seeds were planted, and the problem is just that no tree seeds have ever gotten there through normal ecological processes?
That seems like a stunning L for evolution and ecological equilibriums, but maybe it's trivially true?
Also - since bamboo grows so fast, is it a more efficient way to capture carbon than slower trees/plants?
I think for many purposes, bamboo counts as a “tree”, because it is a tall woody plant. Basically every plant family separately re-invented the “tree” body plan, and bamboo is the grass family’s version.
A lot of tree planting initiatives fail for the reason you mention - they figure “trees are good, we aren’t using land, let’s plant some trees here” and don’t stop to think about why there aren’t currently trees on this land if we aren’t using it.
There is a more relevant plan though of finding land where humans cut down forest to turn it into agriculture or coal mines or whatever, and get people to stop using it for agriculture or coal mines and let the trees come back. Planting seedlings sometimes accelerates the process a bit compared to just letting the trees seed it (particularly if there are apex trees that used to live in the region but now only exist far away).
The number one region where replanting is needed are urban areas. Trees wont naturally replant on carefully maintained asphalt no matter how long you wait.
Yeah urban trees are great. They also have the advantage that they are often actively maintained, so they can survive even if the local climate wouldn’t support them. They aren’t that significant for carbon reduction though, since there’s so little urban land and the possible tree density is limited. They are good for human benefits though.
Hmmm, just had a thought: bamboo is carbon negative whereas concrete is carbon positive. So have a building material consisting of mostly bamboo with concrete pouted in between it.
There's an enormous amount of land that used to be old-growth forest and has been cleared by humans over the millenia. IDK how suited it necessarily is for trees *now* - in a lot of cases we've probably made the soil less fertile - but in principle it's no knock on evolution that trees aren't growing in fields we deliberately cleared of trees.
With that said, land is ultimately finite, and there's an awful lot of fossil fuels in the Earth's crust. Also, we're using some of that land for stuff.
Another area that is ripe for replanting are man made cattle ranches. I don't know what happens to feed grass growth under a canopy of trees, but there's probably some mutually exclusive dynamics going on (else, why clear the land in the first place?).
Generally a closed forest canopy doesn't allow enough sunlight for grasses to grow. And if the forest floor is carpeted thickly enough in herbaceous material like grass or ferns, it impedes the regeneration of tree seedlings.
the main issue is the water table, not the soil itself. As long as the land is the right level of moist, you can replant the trees that used to grow there since the last Ice Age, and they will beat the odds eventually.
The main problem with reforestation is the fetish-like desire to plant the kind of trees we want, not the kind that can survive there, and drying/watering the land to achieve that. Europe takes the cake here, we destroyed countless square miles of perfectly fine, soggy-soiled oak forests and now try and fail to replant them with pines. We buried miles of alder and willow-rich swamps in sandy soils to force more pine there, which promptly rots at the roots from too much groundwater.
There's factors like soil ph, soil nitrogen content, slope, drainage, and sunlight. Different trees have different preferences, and might do poorly in a non-optimal environment. I can imagine a situation where a change in the local geography means that trees suited for environment A can't get across a stretch of environment B to an uncolonized patch of environment A.
That said, birds and deer and other similar animals often "carry" seeds around, maybe even depositing them with a bit of fertilizer. But then, animals like deer also eat young trees. So the animal factor works both ways.
Trees are actually generally not an effective carbon fixation method in the present environment. They used to be, in the carboniferous period, before microbes evolved which could decompose wood. But now, dead trees generally release their carbon back into the atmosphere over the course of their decomposition. Reforestation is carbon-negative, but forested land is carbon-neutral since it's in an equilibrium where it's releasing as much carbon through decomposition as it's taking in through photosynthesis. Although, you can sequester the carbon longer term if you cut the tree down and treat it in a way that makes it resistant to decomposition.
But over a long time frame, carbon sequestration is not usually a meaningful benefit of planting trees.
The point is that trees only sequester the carbon temporarily before releasing it back into the atmosphere. Forests are carbon neutral because they're in an equilibrium between trees sequestering carbon and releasing it. If we want to sequester carbon, trees don't offer a very long-term solution, unless we cut them down and store the wood under conditions that offer long-term protection from decay. Whatever other benefits trees provide, other methods come out far ahead in terms of effective carbon sequestration.
You have to bury the trees and make sure they don't decompose. But you're probably going to have to bury whatever hydrocarbons you get from carbon capture anyway, so I'm not sure how big a difference it will make in the end.
People did the calculations on that, and basically, trees don't remove enough carbon.
I ran some rough numbers. And the amount of forest we need to return CO2 to per-industrial levels worked out to be 2x earths land area. (Rough numbers.)
wonder, if there is a cheap, environmentally friendly, and technologically durable alternative to asphalt that would be white (or light gray/beige/pink/light blue whatever) instead of near black? Or maybe its just a matter of adding some white agent to existing asphalt to make it lighter in colour?
The first thing that comes to mind is concrete. It's light gray in color, abundantly available, and just as good as asphalt at most of the things that asphalt does. Unfortunately, it has well-known problems with cracking under thermal variance that asphalt does not have, making asphalt the clear winner in terms of durability. I've read about a few different groups working on solutions to this, but nothing that's commercially viable quite yet AFAIK.
I mean, the correct answer to "What do we do with clean, limitless power?" is "Literally everything". Everything that isn't a physical plot of land becomes cheaper and/or more accessible.
If the question is narrowed to "What do we do with clean, limitless power that we don't already do a lot of?", then space flight, routine terrestrial flight, desalination, and massive reshaping of the Earth's crust are good options.
Fair enough. But a lot of those would require other advances in addition to cheap energy. Desalination really wouldn't. The only major other thing that would be required is to ship the purified water around in bulk, and we already know how to efficiently ship liquids in bulk; we do it with petroleum all the time.
Remember that we can apply purification technologies to wastewater as well as salt water.
1/3 or so of water today is used for thermoelectric power generation. That goes away in a nearly-all-solar world. Residential consumption usually goes through wastewater treatment facilities, and with current technologies implemented at scale we could get 80-90+%recovery (a lot of these systems can also be implemented in a home if you want!). Farming is the more difficult case. Can't really recapture transpiration water unless you're farming indoors... which gets a lot easier with super cheap energy. But even without that, a mostly-solar-power world can cut fresh water demand by nearly 2/3 long before you start having to worry about shipping water from coasts to farmland.
This is true, but that's mostly cutting freshwater demand from places where freshwater is already abundant enough for farming. But places like Australia, huge swathes of the Middle East, etc. would become vastly more habitable with access to cheap desalinization, much more easily than we could divert water to there from other more arable locations around the world.
Could do halfway-indoors agriculture with inflatable tents, like an enormous transparent air mattress: https://caseyhandmer.wordpress.com/2019/11/28/domes-are-very-over-rated/ Natural shape doubles as a funnel directing both rainwater from above, and any condensation on the underside, toward some sort of combined irrigation / HVAC machinery at the anchor points.
We know how to ship petroleum for a price of several dollars per barrel. That’s not going to be a viable price for water unless there’s a very special place that has a reason to exist even with really expensive water.
Does it? I would have thought most of the cost was actually building and maintaining the pipelines (and more importantly, tanker ships), not the energy for pumping. In any case, no one has ever transported oil at the scale that water would be transported here. The entire world’s oil consumption is only a bit more than New York city’s water consumption.
Well, I don't think you'd be shipping much desalinated water by tanker ships, you'd produce it at the nearest coastline.
In terms of flow rates, the trans-Alaska pipeline can ship 340,000 cubic metres per day, which is about 4 cubic metres per second. The flow rate of the Thames River (to pick a medium-sized river that most people will be kinda-familiar with) is about 100 m^3/s so you'd need 25 trans-Alaska pipelines to create one Thames. The trans-Alaska pipeline is only 48 inches in diameter (smaller than most TVs) so it's not hard to imagine scaling this up by a factor of 25 at less than 25 times the cost.
How much of that is because the existing desal plants were optimized for energy being expensive? A setup designed for intermittent operation, taking advantage of electricity prices down around "we'll pay you to take it," might be able to get away with far lower capital costs relative to throughput capacity.
> The only major other thing that would be required is to ship the purified water around in bulk, and we already know how to efficiently ship liquids in bulk; we do it with petroleum all the time.
Quantity scales with water are quite a lot bigger than for petroleum,
Humans already move large amounts of water around with dams, canals, pipes ect.
It currently costs a lot to move enormous piles of dirt from one place to another, especially if you want to do it in a precise way. (e.g., not exploding a mountaintop). This is mostly because dirt, rock, and clay are heavy, which means a lot of force is required to move it. With boundless energy, you can generate the force more easily and cheaply, meaning we'd probably do a lot more customizing of medium-to-large swaths of land. Things like building local mountains, valleys, and the like. Like how in Sim City 2000 you could start with a "blank" city and then just create the rivers, oceans, and mountains you wanted. Applications include aesthetic appeal, preservation activities, and for large scale industrial efforts.
I'm sceptical that sufficient cheap energy is the main limitation towards doing things like building mountains and valleys. The engineering challenges regarding f.ex. geotechnical, environmental, hydrogeology, hydrologic issues etc. etc. would be daunting.
Cheap energy means cheap steel, allowing more and taller skyscrapers. Cheap lighting, potable water, ventilation, imported manufactured goods, so the spaces below remain pleasantly habitable. So, yes, to some extent it could.
The greater obstacle is land speculation, solution to which is a land value tax.
And the difficulty of that planning permission is deeply entangled with incentives from land speculation, which I already agreed is the greater obstacle.
abundance of free energy would do wonders solving global poverty, improve literacy and education worldwide, thus making Western Middle Class mostly obsolete, and thus, irrelevant. WMC are a bunch of specialists that we only need until we can automate them away or have developing world's lower middle class do their jobs remotely.
The gold isn't worthwhile, but the uranium might be. It seems right on the edge of practicality. (But what' you'd use it for if electricity were really cheap isn't that clear.)
Uranium would still be critical when optimizing for watts-per-kilogram, mostly meaning space travel. Gold might be worth setting aside a channel for if you're going to be thoroughly sifting and sorting all the rest of the sludge regardless.
Like Bob Frank said, it can be used. The chemical industry will guzzle it up in a lot of cases, certainly for all of the desalination plants that are geographically close to chloralkali plants that need brine as a raw material, and potentially for the ones that are further away as well depending on demand.
a pool a brine is essentially a giant battery - we could do interesting things with it.
But realistically, the amount of brine that would result from this is negligible compared to the size of the ocean. Just tow it a 100 miles off shore and dump it into the Pacific. All the post-salination brine we could ever produce would be like a single grain of salt in a bathtub when thrown into the ocean.
That would be my immediate answer too. There is hardly a more fundamental issue than access to potable water. Even in regions like Central Europe, which have never been in danger of permanent water scarcity, falling groundwater tables have been causing concern in recent years. Climate change and changing rainfall patterns will only compound that problem.
Well the point was more that water is or can be an issue just about anywhere. If it really came to a need for desalination for Central Europe though, it would be solvable in principle, for example through desalination on the coast and transporting the fresh water through pipelines.
Huh. Where did you see that? I'll admit I'm no expert in the field, but what I've heard from people who are is that energy costs are the largest factor by far. If you've got solid data that says otherwise I'd love to see it.
We could have home appliances that never break. Right now, our dishwashers, refrigerators, etc. are designed to be as energy efficient as possible to the detriment of being long-lasting. *This is an excuse I've heard, planned obsolescence being unreasonably financially successful is probably playing a role here too.*
The biggest problem with dishwashers isn't energy; it's that they can't *scrub.* It's the classic "we have to wash the dishes before putting them in the dishwasher" complaint: if you put dishes (or especially cookware) in with non-tiny amounts of food stuck to them, it's more likely to end up baked on in the drying phase than cleaned off.
Adding more power won't fix that. More likely would be the outcome (admittedly exaggerated for comedic effect) memorably demonstrated in the pilot episode of Home Improvement, when Tim soups up the dishwasher with More Power™ and it ends up exploding.
Eventually you'd have to. At some point it has to be easier to recycle landfills to get at various metals than to dig up increasingly less profitable mines, right?
Seems to me the question of what we'd do with cheap or near/limitless power is similar to someone in 1970 or 1980 wondering what consumers would do with a personal computer.
I agree that fresh water is likely one of the commodities which will become much cheaper if the price of energy goes to zero.
An other example would be aluminum, where 34% of the costs are electricity costs, and another 37% of the costs are for the raw materials (alumina, etc), which would likely also drop if the prices for electricity and metals would drop.
For some plants, it might become economically feasible to grow them under optimal artificial light 24x7 (the way cannabis is sometimes grown in areas where it is illegal), thereby decreasing prices of raw food.
I'd be grateful if anyone could explain the costs of solar to me. On the one hand I see this 'Our World in Data data', which are "without subsidies", and which show that costs are already among the lowest (lower even than nuclear).
On the other hand, I keep seeing people post what we (in the UK) actually pay solar providers via "contracts for difference" (https://www.iea.org/policies/5731-contract-for-difference-cfd) and it is usually dramatically higher than any other source.
Prima facie, the figure with subsidies seem like it might be the more relevant, but it depends on to what extent the subsidies are actually necessary in order to continue receiving the energy from this source.
The purpose of CfDs is to drive investment into renewable generation by making sure they generate enough revenue, meaning the strike price will be set at a figure high enough for payback and profit generation. But in any case the UK is an awkward market to look at investment in generation assets because the planning system is so terrible.
Lazard recently published their 2024 analysis of LCOE, including a slide on costs with and without US tax subsidies. Here's the headline chart:
The chart that Scott provides shows the "Levelized Cost of Energy". For solar it is the cost of electricity delivered locally by the solar panels when the sun is shining. It includes the amortized cost of producing and installing the panels. It is not the cost of so-called dispatchable power that is available on demand 24/7 at distant locations via "the grid."
It provides an excellent, detailed, technical analysis of both renewable and nuclear energy production, looking at everything from costs and cost-curves to tax policy to land efficiency. (My own aesthetic preference is for solar to be deployed on rooftops and in urban areas and not to cover hundreds or thousands of acres in solar panels for commercial generation. Nuclear is *much* more land-efficient on a energy-unit basis. See fig. 13 of the report.) There is an interesting discussion and critique of the Levelized Cost of Energy (LCOE) metric. There is also a counter-intuitive graph showing a cost-analysis of decarbonizing California's grid using renewables with storage only versus renewables with storage AND nuclear. (Fig. 6.) I know most people won't read the report, so here is a helpful and, I think, accurate summary:
"Nuclear has an essential role in the energy transition as a clean firm complement to renewables. Nuclear provides clean firm capacity; modeling shows including nuclear and other clean firm resources with variable renewables reduces the cost of decarbonization. Nuclear can help address the power needs coming from load growth, where much of the demand is disproportionately for 24/7 electricity, e.g., data centers. Nuclear does not “displace” or “compete with” renewables; decarbonization requires both nuclear
and renewables. Nuclear provides clean firm generation that enables the increased deployment of variable renewables like wind and solar."
This is something of a battle of assumptions that will only be resolved by time. There are some reasons, however, to think that nuclear will play a critical role and that Big Tech's recent embrace of nuclear electricity generation is not an ignorant mistake.
First, aesthetics of solar farms aside aside (and the fact that one person's "cheap low value land" is another person's beautiful open space with various endangered species), transporting electricity long distances from massive solar farms on "cheap low value land" to urban and industrial centers requires large, expensive upgrades to the electricity transmission grid. Those projects are time-consuming, regulatory nightmares that rival the construction of nuclear plants before the federal government decided to promote nuclear energy. Check out the history and costs of the SunZia Transmission Project for an eye-opening example.
Second, as noted above, the Department of Energy itself has concluded that "decarbonizing the grid will be very difficult and expensive without 20-40% clean firm power. Firm power refers to power or power-producing capacity intended to be available at all times during the period covered by a guaranteed commitment to deliver, even under adverse conditions. With an increasing portion of the grid supported by renewables, the value of dispatchability provided by firm power increases. A variety of technologies including nuclear can help maintain grid stability via synchronous inertia, reactive power, and other benefits."
Third, many of the growing use cases for electricity are especially sensitive to even small fluctuations in power that make even solar-with-battery-backup infeasible. Although some Big Tech companies have claimed to use 100% renewable energy before their open embrace of nuclear, this is merely a financial claim; their data centers, especially those devoted to AI, are actually powered by the most stable "non-renewable" sources of electricity-generation (including taking power from "the grid," which is stabilized by utilities using firm power sources).
Casey Handmer's arguments are interesting, but the people at the Big Tech companies who are responsible for securing large amounts of electrical power for real world uses have looked closely at the same issues and the same straight-line graph forecasts and have decided to embrace nuclear electricity generation.
EDIT: This isn't a perfect comparison since the nature of CfDs is you pay the generator even when you don't need the electricity, which means the actual price you're paying for the electricity you use is higher. You're also paying for a lot of electricity you don't use depending on the make up of the grid. See my last paragraph for how this can work like a subsidy if you build too many renewables without sufficient storage.
The nature of electricity markets in the UK (and most/all? of the developed world) means gas will almost always be the most expensive source of electricity when purchased since it has a high marginal cost (the price of gas) so generators won't produce until there is demand that can't be met by other generators. Sources like solar have negligible marginal costs of production and will outbid gas generators as long as they have capacity available. This means gas tends to set the price of electricity during high demand periods and tends to set that price much higher than when zero-marginal cost sources have sufficient capacity to meet all demand (in the case of excess capacity, you even get negative prices which have begun appearing in the UK with some frequency).
But CfDs look kinda like subsidies under current policy which is supporting renewable construction for climate reasons. At a certain point, you've used CfDs to support overbuilding the renewable system hugely (to deal with periods of limited sunshine/wind and high demand with limited storage capacity) and have lots of excess capacity/curtailment. CfDs will almost never save a grid operator money at that point while forcing up rates for customers in the long-term.
Levelized cost of electricity (LCOE) is basically what you have to pay someone per-MWh to get them to build a project. It's the best we have to compare the costs of generating electricity from gas, solar, wind etc where the costs come at different times in the project lifecycle: for solar and wind nearly all the cost is that of building the plant (capex) while for gas most of the cost is the fuel. The main inputs are capex, opex and the cost of capital.
Since LCOE is a theoretical calculation, it makes sense to publish "without subsidies".
Average global LCOEs aren't particularly illuminating because, for example, the same solar panel will generate twice as much electricity in southern California as in the UK. However, it is true they have come down everywhere, mainly because solar capex has fallen drastically. In markets like the UK where solar panel costs aren't affected by import tariffs, a Watt of solar panel costs about 10 US cents today, compared with over $5 in 2005.
However, even in the UK (which is not very sunny) the Contract for Difference prices in the last round of auction were slightly lower than for wind:
(Personally I think this is not great, because the UK needs more power in the winter - it's a highly seasonal climate. But it does also have a lot of wind).
LCOE is particularly poor for judging the cost of renewable power because there are system level costs associated with it that do not exist for traditional generators. Mainly, solar and wind are not available on demand so additional backup capacity or storage is needed to support them (there are other minor costs as well, a more distributed grid which solar tends towards has higher interconnection costs for instance). You should assume the actual costs of solar/wind are anything from slightly higher than LCOE to much higher depending on the grid. At the moment, this still makes them the most economic option for additional power generation in many places (hence the solar boom). Though as more solar is added, more backup is needed and system costs increase.
Batteries do a lot to solve this problem and costs are coming down. I don't full understand why we haven't seen wider deployment of grid-scale batteries yet given the cost data I've seen and I suspect it's a lot of it is inertia. But batteries don't completely solve system costs. The current economics of batteries make storage for 1-2 weeks (think cloudy, no wind and cold in the winter for several days in a row) very expensive. Without batteries, an entire separate generator system needs to be maintained as a backup for these situations which is a huge system cost for renewables.
This seems like a solvable problem to me, although slightly more nuclear would help most grids. In the meantime, 60-70% renewable grids (excluding hydropower) seem very achievable in most places but getting much higher than that could be very expensive.
>Batteries do a lot to solve this problem and costs are coming down. I don't full understand why we haven't seen wider deployment of grid-scale batteries yet given the cost data I've seen and I suspect it's a lot of it is inertia. But batteries don't completely solve system costs. The current economics of batteries make storage for 1-2 weeks (think cloudy, no wind and cold in the winter for several days in a row) very expensive.
Any suggestions on where to look for the cost of solar + batteries, as a function of how long the battery backup is good for? I tried doing a BOTE calculation, and it looked like solar + 12 hours backup for night approximately doubled the cost, with a 10% interest rate for the capital cost of the batteries - but this was really crude. Are there good numbers available for this?
In Massachusetts I regularly got offers in the mail to switch power providers to 100% renewable sources. The quoted rate per kWh was often lower than the default National Grid rate. Solar, unsubsidized, as produced, does in fact have the lowest cost of any power in the world in most places. This is not the same as the price charged to residential customers of a utility.
Often the problems here in practice include the fact that grid operators need to provide dispatchable power on demand, and right now the cost of "firming" solar power supply is very high in many places. There are quite a lot of rabbit holes and quagmires of technology and economics and infrastructure and policy to go down to untangle that mess, and they vary a lot by location.
I also get those offers, and have been told that the gotcha is that the quoted lower price is subject to change at any time, whereas National Grid rates are guaranteed stable. OTOH, can confirm that unsubsidized solar is cheaper than National Grid, based on the solar panels on our house.
Well, in my experience from 2012-2021, those rates were not any less stable than National Grid's rates, and when they did change, it was very easy to switch providers a second time.
I do not know the law in UK well enough to explain it, but I have listened to enough people complain about the laws in the UK to know that the energy grid is subject to extremely weird and bad regulations, and nothing that happens there is anything like a natural price.
Levelized Cost of Energy is a highly imperfect metric that obscures as much as it reveals.
What is needed are comparisons of total network cost of making reliable 24/7/365 electricity ubiquitously available. LCoE does not enable such a comparison.
Solved the Twin-Nuclei problem and then post Bretton-woods. I think Eric Weinstein coined the term EGO (Embedded Growth Obligations), but I think that sums it up pretty well.
> somebody told me that the San Francisco legs of the BART - the Bay Area’s light rail, infamous for being noisy, dirty, and violent - had become comparatively safe and clean over the past few months, after the city installed fare gates that actually worked and couldn’t trivially be jumped over. Apparently the people ruining the BART for everyone weren’t even paying the fare. I always would have guessed there was a correlation between bad behavior and nonpayment, but am surprised at exactly how high the correlation has turned out to be
Psst! Don't look now, but California is about 2 steps away from rediscovering the principle of Broken Windows Policing. And they think it's a new thing they just came up with!
Also, couldn't the *Grants Pass* Supreme Court ruling have had something to do with it? I've definitely noticed a serious improvement in Philly over this past summer.
Reading the article, it seems like Grants Pass is having continued problems due to restrictions specific to Oregon and the 9th Circuit. That's really frustrating.
Meanwhile, on the east coast, I noticed improvements within a month of the decision. The Philly ACX meetups are in the old town/tourist-y part of the city near Independence Hall. There used to be lots of homeless sleeping on the sidewalk and panhandlers all over certain intersections. Now the sidewalks are all clear. There's one spot coming off the Ben Franklin bridge that still has panhandlers, but it's usually just 1-2 guys who are only there during the day, instead of an entire tent encampment under the overpass.
Downtown Atlanta was similarly cleaned up when I visited over Labor Day weekend.
(Side note - there are no apostrophes in US place names. Places like Harpers Ferry, Pikes Peak, Grants Pass, etc. are all spelled without them. There's an act of Congress banning them.)
That's good to hear. I used to live in the general area, and have been to Philly many times. It's encouraging to hear they're improving things out there.
The ambiance is improving for some, but the homeless people are still there. I believe in progress deeply, and I fully believe the Grants Pass decision is the opposite of progress. It's like sweeping guts under the rug and calling it a day.
Unfortunately, sometimes that's all you can do when too many people who believe too hard in "progress" make actual solutions impossible. The first and most significant problem by far in dealing with "homelessness" is Progressives cynically redefining a problem that has nothing to do with housing as "homelessness" so that they don't have to face up to the consequences of their failed policies in the areas of drugs and mental illness.
“We all want progress. But progress means getting nearer to the place where you want to be. And if you have taken a wrong turning then to go forward does not get you any nearer. If you are on the wrong road progress means doing an about-turn and walking back to the right road and in that case the man who turns back soonest is the most progressive man. There is nothing progressive about being pig-headed and refusing to admit a mistake. And I think if you look at the present state of the world it's pretty plain that humanity has been making some big mistake. We're on the wrong road. And if that is so we must go back. Going back is the quickest way on.”
I was in the core part of downtown for DragonCon - the Hyatt, Marriott, and Hilton. Andrew Young Blvd and Peachtree Street. We stayed up in Midtown and took MARTA back and forth every day.
There were some wandering homeless/mentally ill guys without their stuff around MARTA, but way fewer panhandlers or campers on the sidewalks than in past years. I lived in Atlanta for college from 2013-17, so I'm calibrated to those years. It seemed about steady for the last ten years, and then dropped off a lot this time around. I wasn't accosted by a single panhandler this year.
But I was only there for a few days, so maybe it's different in other parts of the city/different times of year. Maybe DragonCon beefed up security outside on the streets this year?
I was very amused by all the ads on MARTA this year telling people to move to Philly. I saw those and was like, yup! Philly's great!
Aha. Thank you - I live in the East Atlanta area and don't go down to the Central Business district that much. East Atlanta has seen a marked increase (and so have interstate areas from what I've seen). Greater enforcement would explain the shift.
Philly was not effected by the Grants Pass decision, so any improvement is unrelated. The ruling that the Supreme Court overruled in Grants Pass was for the 9th Circuit which covers the western US, and it never affected the east coast.
Frankly, that Grants Pass supreme court case was so disgusting that is still baffles me that some effective altruists like it. It solved nothing and only increases pain and suffering. The homeless people it affects are being offered no services, just a temporary jail stay or even just a violent eviction. It's effectively social murder. As a society we don't want to actually address homelessness because it would be expensive, but to make the Grants Pass decision make sense morally, you have to actually address homelessness.
I agree with most of this, but not the part about it solving nothing - I think that criminalising homelessness and driving all the homeless people to move to other towns probably will solve the problem of "we have lots of homeless people in this town, lowering quality of live for others".
Good objections to criminalising homelessness in order to force homeless people to move to other towns include 1) it's unethical, 2) it makes life worse for the inhabitants of those other towns, and 3) it doesn't scale, because if everyone does it you're back to square one except everything is worse.
But if your goal is purely "solve my town's homelessness problem here and now, without worrying about any of those things", I think it's probably a really effective way of achieving it.
One of the ways we address homelessness is by criminalizing homeless encampments. Most of these people are dug addicts and will slowly kill themselves unless something stops them first. Allowing these people to sleep where they want and die slowly in peace is not compassion, and making that not a live option for them is. The Grants Pass repeal will lower homelessness and help people by forcing them into the recovery and support programs we already provide, and that addicts aren’t interested in because they’d have to get clean. Well, get clean, sleep in a ditch, or go to jail: setting up a shack next to the playground at the park is no longer an option.
I think the broken windows version of this would be attempting to arrest everyone who peed on a bart and making the punishment relatively harsh. I think preventing the peeing in the first place is an importantly different thing. I'd be interested in hearing if anyone involved in this decision even believes the broken window thing that doing this one small thing will have larger knock on effects.
In the original Broken Windows program in New York City, turnstile-jumping/fare-beating in the subways was one of the things they focused on cracking down on, with significant effects on downstream crime reduction.
Un-jumpable fare gates is equivalent to removing the fare-jumpers from the population of people riding BART. Arresting the fare-jumpers would be removing them from public society. So right now only BART gets the privilege of a social environment with a significantly-reduced jackass population.
I don't think it is equivalent. There's a set of people who will jump when they can, and that set is also more likely to commit other crimes. But they're capable of paying a fare, just like they're capable of using a taxi/uber if they actually need to.
I don't think "fare-jumper" is necessarily a persistent, intrinsic property of individuals. Lots of things work more by perceived value, or back-filling excuses for decisions made on some entirely different basis.
If you're able and willing to ride for free, that sets a tone, a social context, where everyone who DID pay the fare must ipso facto be a self-sabotaging coward. Why, then, play along with any of the rest of society's rules?
On the other hand, if jumping is inconvenient enough that it's easier to pay, riding the BART becomes a privilege you've willingly paid for, so the space itself must be worth something, and thus ought to be treated with respect.
That sort of context-sensitive morality is more extreme when drunk.
The real success was that NYC discovered if you arrest fare jumpers, a lot of them already have outstanding warrants (petty theft, assault, etc). So they'd get a big unit of cops together with a paddy wagon and would sweep through the subway, arresting every fare jumper they found. Then they'd process them right there at the paddy wagon (they brought all the necessary paperwork and equipment for processing with them) and everyone they had with an outstanding warrant would go the jail and the rest would get a ticket for fare-jumping and be released. It really helped clean up the streets.
Its that thing about how arresting a tiny percentage of people would end an absurdly high percentage of crime. People who flagrantly disregard the minor laws tend to be bad people who also flagrantly disregard the more important laws.
I don't think that's actually a good example of "broken windows" because it leaves no visible effect. Littering, on the other hand, does. Arresting jumpers is a good way to grab a lot of people who turn out to be carrying illegal items and have other charges against them though (similar to traffic stops).
I think it does have a visual impact, fare jumpers are noticeable when they literally jump the turnstile and in this case they are clearly causing problems on the mass transit and removing them from it ensures a better environment.
Fare jumpers are phenotypically different as a class on average from non fare jumpers. Based on just how someone looks you can come up with a decent idea if they fare jumped. Therefore removing fare jumpers changes the average phenotype of a BART rider.
Not really, no. What we see in the Wikipedia article is commonly known as "lying with statistics," particularly the last bit, where it says:
> A 2017 study found that when the New York Police Department (NYPD) stopped aggressively enforcing minor legal statutes in late 2014 and early 2015 that civilian complaints of three major crimes (burglary, felony assault, and grand larceny) decreased (slightly with large error bars) during and shortly after sharp reductions in proactive policing. There was no statistically significant effect on other major crimes such as murder, rape, robbery, or grand theft auto.
You have to be careful here or you might miss the trick. First, there's the big obvious one: the decrease in reports of crimes is "slight[] with large error bars." But then there's the much more subtle one: they're looking at the rate of *civilian complaints of* crimes, not the rate of crimes being committed. And keep in mind that this was happening in 2014/2015, at the height of Obama-era anti-police pressure tactics!
It turns out that, when people have little reason to believe that calling the police to report a crime will do them any good, they're far less likely to call the police and report the crime. (See also the broad divergence under the Biden/Harris administration between the FBI crime statistics, which are based on police reports, and National Crime Victimization Survey numbers, based on asking people if they've been the victim of a crime.)
The only crime statistics that can be taken at anything close to face value are homicide and auto theft (including smash-and-grab of stuff in automobiles). Murder, because basically nobody tries to hide the bodies, literally or statistically speaking. Auto theft, because even if you know the police aren't going to do anything you need the police report to file an insurance claim.
Everything else, reporting bias makes the error bars uselessly large. So if someone is counting "burglary, felony assault, and grand larceny", but not murder and not auto theft except as buried in the larceny numbers, yeah, suspect statistical skulduggery.
I think vibes beat statistics in this particular case. New York in the 1980s had a reputation as crime-ridden -- I remember a lot of Mad Magazine jokes about the impossibility of crossing Central Park without getting mugged. By the 2000s it was had a reputation of being remarkably safe by US big city standards.
Depends entirely on who you ask. As far as I can tell, there's still a substantial minority of people who believe in it, but they're almost exclusively conservatives who believe other things that make me distrust their judgement.
I would like the BART gate thing to be true, but I question it -- the gates are really new, only 8 stations on the entire BART system have them so far -- and one of those 8 is at the airport, which is probably not a major location for sketchy people to board. In San Francisco, only 3 BART stations have these new gates. One is Civic Center (which is less than half a mile from Powell, which doesn't have the gates) and another one is 24th and Mission (which is less than a mile from 16th and Mission, which only got the gates as of October 12th). It seems surprising that the gates could be having such a dramatic impact on the BART this early in the installation process.
Oh that you had returned to the Gender Studies comparison because in fact gender studies has been progress studies. Progress studies should include how gender has progressed, including (not to wade into anything) gender technologies.
There’s a lot of gender technologies! Everything from stiletto heels to mom jeans to the manbun to testosterone injectables to surgery is an invention that enables people to present their gender in new ways.
There's something missing here, perhaps because it is difficult to quantify: at what point do you have enough?
It is clear that, up to a certain point, economic growth in the US went hand-in-hand with progress at a more concrete and human level - becoming "richer, healthier, safer, and better-educated than our ancestors", as Scott put it. Scratch that - it's not as if it were just a coincidence - we can probably agree that the former was a prerequisite and the main engine of the latter.
Ditto in the developing world today, *up to a certain point*: you need a certain level of production if you want a middle-class society. No, it's not a sufficient condition, and yes, GDP growth can be very deceptive, particularly in extractivist economies (high GDP growth can mean that resources are being extracted at an accelerated rate by foreign company X, which got an amazing deal from subornable and/or naïve officials, and only has to pay a pittance in royalties). Nevertheless development remains a crucial part of, um, development.
But do most problems in the US really have as their necessary root a slowing-down in productivity? GDP per capita in the US is now 50% larger than it was in the mid-90s. (Note: yes, in constant dollars, obvs.) Are people really that better off? Life expectancy has barely budged. Are people much better educated than then? There are some positive trends (some cities are safer) but their relation to the rest is less than clear. Now, if a wizard were to replace the GDP per capita of our days with that from the 90s, *perhaps* the immediate result would be immense suffering, but not because we were miserable in the 90s; this is what 'growth addiction' would mean, I suppose.
There is some point (probably well before the early 70s) at which people in the US could sensibly talk about an "affluent society"; that's actually the title of a book by Galbraith from '58. That really meant a society were more than half of the population were not in any obvious material need, and their main problems were of other sorts. Jacobs/Nader (who did make us safer)/Carson (whose concerns were valid even if she was wrong on many details) can be sensibly seen as a product of those times. Whether they had any sort of macro impact (for good or for the dastardly forces of anti-progress evil) is much more doubtful, at least outside the very particular fields that they cultivated.
Looking at the big trends is useful and necessary. Looking only at productivity figures alone seems reductive and potentially self-deceptive.
(On this last note: it would be interesting to have Emmanuel Todd's books come up in the annual Book Review here. I would get into that myself, except I don't feel particularly qualified in any way - I am not a demographer, or an anthropologist, or a historian, or for that matter an economist. Of course his stuff is often rather controversial and has to be taken with a grain of salt. Still, it's a good example of big-picture analysis based on a close look at quantitative data that does not stop at the level of dollars.)
My guess (haven't looked into it) is that some of the GDP per capita gains have gone to the top 1%, some of the rest have been eaten by higher housing costs, and some of the rest have been eaten by stricter regulation (or, more optimistically, been spent on safety rather than consumption). I think of the remainder - maybe 20% of the total - some of it *has* gone to things like bigger houses, more cars, more vacations, eating out more, getting DoorDash, and other things that do make us happier.
If somebody were to ask "what are the main things that make Americans unhappy?", how plausible would the following answer sound: "their houses are too small, they have an insufficient number of cars and they don't eat out enough"?
(An outsider wouldn't answer that, and an American wouldn't either. I know, someone will argue "but the fact that people spend money on that shows that's what their priorities lie", but no, really.)
Vacations in the US remain very short relative to the rest of the developed world (only Canada can compete in vacation shortness) - how much has that changed in the last 30 or even 60 years? The only thing in that direction I can think of is that there may be more people in the jig economy, but that's often not because they want to be.
I agree rich Americans probably don't worry about these things. But poor people are much less happy than rich people, and it seems like most problems that poor people have but rich people don't are problems with money.
Well, I'm neither rich nor poor (nor for that matter a US citizen, though I lived there for ten years), but my impression from living in non-rich areas of the US (mainly intercoastal; I've seen more than my share of square-shaped states) was that relatively few people in the US are short of *stuff*, as in, stuff you buy at the shop. Some have trouble getting to the end of the month (that's probably not very closely correlated with income level if you exclude the very top and the very bottom), but that's not quite the same thing.
Money does obviously help with access to many things that are in short supply - health care, a good education, housing. In fact what really helps there is having a lot of money. Is the root cause of that really that there's not enough money to go around? Or that not enough stuff is being produced?
I'm not sure what distinction you're drawing. Yes, there should be a good monetary policy where the Federal Reserve prints the right amount of dollars to denominate the amount of wealth, but the basic and harder problem is having enough goods that they're very cheap.
Let's get the money supply question out of the way (that's an irrelevant issue that human language has somehow invited - though of that as I was about to press Enter). What I dispute is the last bit: that the basic, hard problem, *in general* is having enough goods that they are very cheap. That may make sense at certain stages in development, but it seems countersensical in a developed economy, and even more so in the US in particular.
When it comes to relatively new technology, particularly when it solves a clearly demonstrated societal need, then I agree with you: obviously we need more and cheaper solar panels (keeping an eye out for their environmental cost, which is real, if much smaller than their benefits) and ditto for batteries (there the issue of environmental cost vs. benefit is less clear; if you disagree, I'd be thankful if you could convince me - I have actually just had solar panels installed and I am trying to decide whether I ought to get batteries). And yes, this is possible because we have such a thing as an industrial economy, mass-production, etc.
It's the claim that the slowing down in the rate of growth or qualitative change in the "world of atoms" was "one of the greatest historical tragedies" that has me unconvinced. At least it's not a claim that makes itself.
Having been poor, in the absolute poverty sense of living in a pre-electrified semi-subsistence farming community where there was no municipal water and three towns shared one phone line and then later in the American sense of having way more money than 70% of humanity but less that most other americans, I much preferred the former (unfortunately, modernity has reached that part of Costa Rica, and now they have wage work, cellphones, depression and diabetes).
The thing that was really awful about being poor in the US was how precarious life was, that your shitty dodge you bought used because you couldn't afford a Toyota could blow a distributor cap (I had three rattling around in the trunk), you would miss work, you would lose your job, you would lose your house, and you would be on the street three weeks later.
Now that I have a LOT of money, I still live like I did when I was poor (other than owning my house) but with more big vacations. The extra material possessions and services (to me!) are hedonic treadmill net negative distractions from what actually makes me happy: eliminating stress, and work that isn't completely alienated in the Marxian sense, and novel experiences.
I know I am much less wigged out than my friends who have bigger houses and nicer cars by having a bullshit MBA type job, but have to worry about losing them and dealing with the existential ennui.
>"their houses are too small, they have an insufficient number of cars and they don't eat out enough"?
Replace 'their house is too small' with 'they can't afford a house and their rent eats 50% of their paycheck and their apartment sucks and the landlord never fixes anything,' and yes, this basically describes the gripes of a large number of my working-poor friends.
"If somebody were to ask "what are the main things that make Americans unhappy?", how plausible would the following answer sound: "their houses are too small, they have an insufficient number of cars and they don't eat out enough"?"
Those are real problems for poor people in America who lack those things.
I'd say that a comparison to other developed countries makes it clear that those are rarely problems if the above is taken literally - or rather what Americans have can be a problem *because of the way much of America is set up*.
- I have not yet seen a truly small American house, but of course many Americans don't live in houses. (Note: for my first four years in [major European city], I lived in 23m^2, and that was considered normal for a single academic; I had lived in much less space while in college and grad school in the states.) Compared to just about any place in Europe, American houses are not small - they are, however, impermanent (not just compared to most places in Europe, but even to most non-slum housing stock in Latin America), though that's a preference that *may* make some sense given how cheap lumber is and how much people move. Much more to the point: there is a real problem, in that access to house ownership has become much more difficult, and being a renter sucks much more than in e.g. Germany (where, to boot, places are *usually* better built and insulated, though not always).
- It seems bizarre to think that Americans don't have enough cars (what is next - their cars are too small?). What *is* the case is that the working poor sometimes have a car in good condition and live in a place in poor condition, even if they own it - they depend entirely on cars to work, in that public transportation is pretty terrible outside a couple of major urban centers (indeed, it's less useful than the chaotic, polluting, unregulated 'public' transportation (really madmen with vans) that you find in many Latin American cities - no, I don't mean Mexico City, which has a good metro).
- Eating out often is not any sort of basic need - growing up, we'd eat out once a year, after we got and showed our report cards . More of a problem in the US: a lack of corner shops (you can't run out and across the street to get a clove of garlic), plenty of people get half-hour lunch breaks (they are lucky if there is a McDonald's nearby), etc. (In France, your place of work must either have a cafeteria or give you a 'ticket restaurant' valid at a restaurant. That, together with sufficient lunch breaks, hasn't been invented in the US yet.) Also a reason why eating out feels like a basic need: a lack of third spaces which are either free or close to free. Social life *is* a need.
There is a small percentage of Americans that do suffer from not having enough in a gross material sense - people who are poorly nourished, homeless, etc. It's shameful that anybody is in that condition. There's a much larger percentage of people who have very real problems - but they are not at the level of a gross lack, though it may look at first that way, looking from inside the US, without outside comparisons. Making more stuff available more cheaply at Costco is not going to solve their problems - and these problems haven't become less serious since the 90s, when the US was 2/3 as rich in the sense of GDP per capita; some seem to have become more serious. It's less obvious, then, that the source of these problems is that American affluence has increased more slowly since the 70s.
The American preference for larger houses is because Americans are richer. Rich people in Europe live in bigger houses too.
Lots of people would be helped by America getting richer. The family that moves out of the trailer park and into a house. The 19-year-old who can move out of momma's basement and buy a car.
More Americans than Europeans are (a) able to (b) interested in imitating the lives of the rich in some ways, at least in the sense that a stage setting imitates a house or a landscape. Hence - nothing follows.
Again, GDP per capita in the US has increased (correcting for inflation) 50% since the mid-90s and has doubled since the mid-80s (to give two points of reference at which many readers of this blog were already alive). Are there really many fewer people who live in trailer parks and worse? (I'm actually asking for statistics.) What I would believe more readily is that some trailers have got larger, maybe even 50% larger, but that doesn't solve things, does it.
And yes, Scott is almost certainly right that part of the issue is new wealth being sucked up by the very top, NIMBYism, etc. What is completely unclear to me is that the solution is for the country to become richer faster, in the sense of real GDP per capita.
There is probably a stage of development at which the most important thing is arguably for the country to develop quickly - the gains obtain thereby overwhelm everything else. Developed countries, and the US in particular, would seem to be well past that stage, almost by definition.
re American houses being impermanent. I was just thinking the other day that the old morality tale of the Three Little Pigs seemed to have gone over their heads. Climate change is increasing the number of big bad wolves...
Well, improving the energy efficiency of an existing 'permanent' house can be expensive (ask me); I suppose that's one good potential side of a place where people consider a 30-year-old house to be "old" - when they build, they could build to contemporary standards. It's more that having to build from scratch every 30 to 50 years (numbers I just made up) is not in itself eco-friendly.
Yes, without a doubt, houses could be either bigger or better furnished for ~99% of Americans. Look at the furnishings in the $10M California houses: any house that's not furnished as nicely as those houses could use an improvement. Even if you live on dirt cheap land in the middle of nowhere and only need ~1000 square feet/person, you'd still need to spend at least ~$1M on furnishings/renovations alone to get the "perfect house".
As one example: if your house doesn't have a heated pool where the water can stay at 85F even in the middle of winter, it could still use some improvement.
"Better" is subjective. It may be that to the owner of the $10M house, everything in it is necessary, or could even be improved. And the owner of a $50,000 1000 square foot house could have everything they want, and adding, for example, a pool would detract from the value for that person, either by making the walk around the pool longer to get somewhere (for an indoor pool) or cause a wildlife hazard (for an outdoor pool), or something else.
Keeping up with the Jones's is a fool's game. One needs to figure out what is best for oneself.
> "their houses are too small, they have an insufficient number of cars and they don't eat out enough"
I don't know that I would identify these as the main things, but I am already doing pretty well and yet I would still enjoy living in a bigger home, having a nicer car, and eating more luxurious food more often. Living, eating, and transportation are *huge* parts of everyone's lives, and making them better will increase your QoL for a very long time.
I mean, that's exactly the sort of tradeoff (with less made up numbers) that a lot of people make these days. For the same price, some people choose to live in smaller dwellings closer to the city centre, and others choose to live in big sprawling acreages on the edge of town.
There's no right or wrong answers to this tradeoff, just a matter of personal taste, and the important thing is to ensure that housing is being built to suit everyone's taste.
For me personally, I wouldn't accept a 30% smaller house closer to the city (but then again I don't commute).
There are many margins at which I would accept many tradeoffs! That's totally orthogonal to my point, which is that I think having bigger houses, more and better cars, and more and better food _would in fact_ make Americans happier. If you want to talk cost, that's a different conversation. I'm just objecting to parent comment's dismissal of the idea that lack of big houses, fast cars, and fancy food is keeping us unhappy. Those aren't the *only* things that keep people from sublime happiness. But they're good things to have, and I want everyone to have more of them.
(There are issues with how the poor eat in the US and with what they can eat, but a lack of calories is not one of them, not outside the very, very bottom.)
It may be that many people who are far from thin may want more food, in the sense that they are addicted to it. If their income starts being a limitation at some point, that is not a tragedy.
"how plausible would the following answer sound: "their houses are too small, they have an insufficient number of cars and they don't eat out enough"?"
Very plausible? At least for Americans who are not rich? I can recommend this very impressive article from residentcontrarian:
"I am always consistently shocked by how little people living at a decent-to-great income level fear their cars; it seems like they hardly lie awake at night thinking about their iffy alternator much, if at all."
"Do I spend 50% more on an apartment and stress my budget to death, or make my wife and kids in a place where I have 5 front doors within 15 feet of ours in the murder-iest part of town? "
I've addressed that point below. Neither of these would be a problem if the US had decent public transportation and cities were less murdericious. More cars and fewer front doors are not going to solve the real problems (even if having more money does help an individual).
> If somebody were to ask "what are the main things that make Americans unhappy?", how plausible would the following answer sound: "their houses are too small, they have an insufficient number of cars and they don't eat out enough"?
If we look at revealed preferences, then yes. People who get extra money do spend it on improving their houses and cars (let's leave the eating out aside as more of a personal taste.)
I do think that people also have a lot more *stuff* than they did in the 1990s though. TVs are bigger, kids have more toys, everyone has another dozen appliances in their kitchen, we fly around a lot more, etc.
Separate point: no, that does not tell you what would make people unhappy or happy, if scaled up to the level of a society; in fact, it barely tells you something about what they prefer, and that only at the level of things they can afford.
(a) Yes, the US is car-dependent - most of the working poor need a reliable car.
(b) What people need is things that can in fact be bought (in the US, where they are for sale), but at a completely different price scale: health care, good housing that they either own or that remains affordable and is kept up to a high standard (not: bigger houses), not living in a neighborhood with crime rates five or ten times what is typical for a developed country, etc.
Of course advertising is very good at making people want what they don't need, but one would think most people have smarted up by now.
>I do think that people also have a lot more *stuff* than they did in the >1990s though. TVs are bigger, kids have more toys, everyone has another >dozen appliances in their kitchen, we fly around a lot more, etc.
Here in Denver, a safe neighborhood in Superior with large, new houses is cheaper than a safe neighborhood in Louisville with crappy old small houses and a vibrant little downtown with lots of families and kids bustling around.
I think people do prefer larger houses over smaller ones, but their main preferences are for a safe neighborhood with good neighbors, a good community, a reasonable commute and interesting things to walk to. Larger houses tend to get built in new developments for practical reasons, and those developments don't usually have any nearby shops to walk to, but "lack of shops and restaurants" is not a revealed preference. It has more to do with regulations.
s/"house is too small"/"apartment is too expensive and an unreasonable commuting distance from work", and you're getting a lot closer.
Those are both manifestations of housing costs, but they affect different populations differently. You've got one large group that's happy living in the houses they bought decades ago (raises hand), another large group that despairs of ever owning a house and are dissatisfied with their expensive, poorly-situated homes, and some smaller groups that like their apartments or dislike their houses.
Higher GDP can be converted to more vacations and the same amount of stuff, if a society wants. The fact that we all keep working the same level of hard to get MORE stuff instead is a revealed preference, no?
Or perhaps it's a deeper societal ill around coveting what others have ... but that's orthogonal to GDP IMO.
A few other things can also cause big divergence between measured GDP per capita and real utility/well-being.
First, and simplest, is hours worked. Increasing labor force participation, longer work weeks, later retirement; all of these will increase measured GDP per capita but at the cost of both leisure and unmeasured home production, so the welfare gains are not as high as you would think from the economic statistics. (But while we’re specifically comparing the US in the 90s to today, this isn’t a big part of the answer.)
Second, harder to measure but potentially unbounded in effect, is status externalities. The extent to which your house is “too small” is partly about real consumption, like how much space you actually physically need to move around in comfortably, but also partly about social comparison. If we double everyone’s house size, this won’t double the utility everyone gains from their house, partly because some of the utility from having a “big” house is specifically that other houses are smaller.
Third, and more speculative, are meretricious Needful Things. A dollar spent on Doritos counts the same as any other dollar in the economic statistics. The feeling of regret I get after finishing the entire bag of Doritos is not measured on economic statistics. To the extent that improving technology makes vendors better at selling us Dorito-like goods, this could also create a increasing gap between measured GDP per capita and actual well-being.
These are great points. A labor-hour-adjusted GDP per capita metric is relatively easy to obtain, but it’s significantly more difficult to try and adjust for status or for Dorito-like goods.
>Third, and more speculative, are meretricious Needful Things. A dollar spent on Doritos counts the same as any other dollar in the economic statistics. The feeling of regret I get after finishing the entire bag of Doritos is not measured on economic statistics. To the extent that improving technology makes vendors better at selling us Dorito-like goods, this could also create a increasing gap between measured GDP per capita and actual well-being.
That's fair. There are also analogs in the sorts of things degrowthers push:
An hour spent guilted into a community activity that one would actually rather avoid, for instance.
Economics recognizes decreasing and negative marginal utility. Sure, it can't be in the statistics, because economics also assumes people behave to increase their total utility which, as you illustrate, isn't always true.
Many Thanks! I also agree with one of your previous comments, that:
>Keeping up with the Jones's is a fool's game. One needs to figure out what is best for oneself.
But I don't have a good intuition (let alone data!) on _how much_ of GDP growth is in purchases of items with negative marginal utility to the purchaser, essentially mistakes at the level of individual households.
There is also the problem of what a society could _do_ with such information, even if they had it. In theory, a wise and benevolent regulator could try to discourage such purchases - except that wise and benevolent regulators are as available as the Tooth Fairy and Santa Claus. ( I'm writing from the USA, where we are currently trying to decide between Horrible and Worse for POTUS. )
Certainly the case in the UK that GDP per capita gains have gone to higher housing costs. It's a sellers market of an essential so if you put more money into the system everybody ends up in the same house but they pay more for it. A big part of the rise in house prices has been driven by larger mortgages which take longer to pay off, and then impinges people's saving for retirement.
You might be interested in Easterlin's book on happiness. If your disagreement with Scott in the comments is over whether more wealth -> more happiness, his data suggests you're right that it doesn't.
Thanks for the recommendation - I'll take a look. I think what we are getting at (I'm not sure there will be a disagreement in the end) is something a bit finer than that. I think we can all agree from the start that
(a) it's good for a society to be competent at producing a wide range of things well, quickly and cheaply
(b) the life of a random individual poor person in a rich country would, on the whole and most of the time, be better if they had more money. (On the whole, the less money you have, the more difference a given amount of money makes.)
So the issue is not that (but I'll take a look at Easterlin).
The idea that more money doesn’t make people happier after a low threshold has been thoroughly, thoroughly debunked.
Affective happiness goes quite consistently up with the log of income in the US.
*Unhappiness* does seem to have a threshold where more money doesn’t help, which is one reason why people get confused.
The idea that more money doesn’t help with happiness is, to me, an irritating just-so story folks like to tell themselves to help them feel less jealous of the wealthy.
As it happens, my mother is one such person. She went from middle decile income to receiving an unexpected inheritance.
She is much, much happier. This didn’t surprise me, but it sure surprised her.
In the US, enough money will buy you out of many problems that non-rich people in some other developed countries generally don't have. That does not mean that increasing per capita GDP as rapidly as possible will make those problems go away.
>In the US, enough money will buy you out of many problems that non-rich people in some other developed countries generally don't have.
Could you give two or three examples of what you view the most important such problems are? It would help to make the discussion more concrete. Many Thanks!
You should check out Easterlin's book if it's something you have strong views on (or Scott has a post on it on SSC) he has responses to all these objections.
The log(income)->Happiness data is correlational among individuals, Easterlin's claim is specifically that there's no causal link between higher wealth and happiness for the society as a whole, if that makes sense.
It makes sense to me that someone successfully working toward what they want increases their happiness, which would correlate with higher productivity for that individual.
And there's no doubt in my mind that it's possible to have bad culture combined with relative prosperity, and trading some prosperity for better culture may well be worth it at some margins, for groups, not just for individuals. So I don't think I'm rejecting the thesis that it's possible.
I just don't think it happens in practice to large countries not involved in wars or truly giant natural disasters.
Said another way: holding culture constant, I don't see a case for more wealth doing anything but making people happier.
Regarding Easterlin... I found Betsey Stevens and Wolfers (many) responses to him quite convincing. His statistical analysis just seems shoddy.
"I found Betsey Stevens and Wolfers (many) responses to him quite convincing."
Hadn't seen that one. Giving it a quick look it doesn't seem convincing though.
The first part of the paper is just re-establishing the (income, happiness) correlations between individuals and across countries that are already a part of Easterlin's paradox.
Then they find small correlations with log(income growth) and happiness in a few individual countries that would be relevant evidence against Easterlin but just looks like random noise or a very small effect size at most.
I'm thinking about reviewing Easterlin's book for the next review contest so I'm looking for counter arguments. So far they've all had obvious weaknesses. Is there something more to the paper I'm missing?
Yes, the statistics are clear on that. If it hasn't made people smarter or happier, that's a problem with higher education. It doesn't mean more wealth is not being expended.
Never! Or at least not anywhere we can see from here. I want everyone to get richer and richer at increasing rates and enjoy science-fiction levels of comfort and luxury. I want our current fantasies of a rich life to eventually seem like quaint anecdotes of harder times.
That seems like entirely the wrong approach, particularly for Americans. There are (at least) tens of millions of Americans that have enough material comforts and more than enough tchotchkes, yet are doing badly in many other ways.
Sure, unhappy families are unhappy in their own way. Poverty is only one thing that can make people unhappy, but since it is one and we can address it, let's do.
Just temporary poverty can cause a lot of damage. Unemployment leading to cash-flow problems causes lots of divorces. If people have more money in the bank, they can ride through more storms.
My point is that American poverty is real but that a lack of stuff is not what makes it real. Yes, a financial cushion is nice - and plenty of people in the US who are not poor or would not be considered poor elsewhere have no cushion; compare its savings rate of 4.8% to https://tradingeconomics.com/france/household-saving-rate-eurostat-data.html .
there are zero humans who have enough material comforts by my standards.
it seems you might be holding some sort of context in your head where you are advocating for prioritizing one side of some trade off, in which case you should probably say so explicitly. it is true that at some margins some interventions might be better than simply prioritizing 'everyone gets richer' for addressing some problems.
nonetheless, overall i very much want everyone to get unfathomably richer.
This, um, is so extreme that it makes my point. Actually, this makes me think that there must be a character in myth or fairy tales created to warn others of this particular kind of pitfall, sort of like King Midas, the utility monster, the sandman, etc.
Centuries ago even the wealthiest people were poor by our standards because things we value weren't available at any price back then. Centuries in the future, we will also be poor by the standards of such future people.
There are ways in which a poor person in most developed countries is genuinely better off than a rich Victorian - a chance at getting cured of cancer, say. As for being able to have pineapples in a can, or Internet access: who cares, in comparison? The problem of Victorian squalor was of a different kind. Consider the poor in the Third World nowadays - their squalor does not go away when they get a TV.
No, it just sounds extreme if you don't a) read it carefully and b) take it literally, rather than as flag-waving.
It genuinely is always true that, *all else being equal*, making anyone better off is a good thing.
It's also true that all else often /isn't/ equal - as Alexander Corwin implies, there are usually tradeoffs, and the value of making someone who is already rich even richer, while technically positive, is often so small that practically any negative downside will outweigh it.
But "how rich is rich enough?" isn't really a useful way of thinking, because the answer is "the point at which the sign in the cost/benefit calculation flips depends entirely on what method of making them richer we're discussing them, and what the costs of it are", so there isn't some set level of wealth that answers that question.
Even if you clarify the question to "if you had a magic want that would give someone an extra cent and not change the world otherwise, so the only downsides are inflation and Spirit-Level-type inequality effects, at what point should you stop waving it at them?", I think the answer is that it will be sooner for someone who prioritises consumption than for someone who prioritises investment, and sooner for someone who prioritises investment than for someone who prioritises philanthropy, and for any of them it will depend on the state of the economy they're in.
My point is that's the real question that we should be discussing, or at least a real question! A net increase in 'available material resources' (available to whom?) would be nice, but the US has kept on getting wealthier since the early 70s, just more slowly (like every other developed country), and yet in some ways it has genuinely stagnated or got worse (e.g. access to housing).
That's kinda why YIMBY has become such a big deal now. Regulation has artificially made living so expensive that the extra money doesn't go to extra things, like allowing the poorest people to live in decent housing.
And I would add that we have a lot of new undeclared part time jobs. Last week I spent two hours and five phone calls trying to book a referred appointment for an ultrasound. This week I will be calling various offices about some weird medical billing that I don't understand and seems wrong.
"Paper claims that we are gaining 0.5% per year in terms of how much welfare we get from across a variety of categories from increased product specialization and variety. Households increasingly spend funds on specialized products that exactly fit their preferences, with the increased variety driving the divergence in consumption.
This is also evidence we are richer. Increased product variety requires people able to consume enough, and pay enough extra for quirky preferences, to justify greater product variety. This represents a real welfare gain. However, instead of making people feel less constrained and wealthier, it puts strain on budgets and competes with and potentially puts additional strain on raising families rather than making it cheaper to raise one.
I very much appreciate the product variety, but increasingly I think we need to consider three different measures of wealth:
The welfare value of the experience of the items in a typical consumption basket.
The combined welfare value including goods that remain unpriced.
The difficulty in purchasing the typical consumption basket, and what affordances that leaves for life goals especially retirement, marriage and children.
Or: The Iron Law of Wages proposes that real wages tend toward the minimum to sustain the life of the worker. So we can measure four things.
The minimum real wages required to sustain the life of the worker.
The welfare value of that minimum consumption basket.
The surplus available after that to the typical worker and what that buys them.
What else is available that is not priced.
When we either effectively mandate additional consumption, such as purchasing additional safety, health care, residence size, education or other product features, or our culture effectively demands such purchases, or the cheaper alternatives stop being available, what happens?
We do increase the welfare value of the minimum basket. We also raise the cost of that basket, which reduces everyone’s surplus.
What happens when things that people value, like community and friendship and the ability to raise children without being terrified of outside intervention, and opportunities to find a good life partner, are degraded?
Life gets worse without it showing up in the productivity statistics or in real wages.
The current crisis and confusion could be thought of as:
The value of the minimum consumption basket is going up a lot.
The cost of the minimum consumption basket is going up less than that.
Real wages are going up, but less than the cost of the basket, so the surplus available after purchasing the basket is also declining.
Key other goods and options are taken away, like those mentioned above.
Economists say ‘workers are better off,’ and in many ways they are.
People say ‘but I have little surplus and do not see how to meet my life goals and I have no hope and my life experience is getting worse.’"
Enough for what? If you are seeking to maximize total happiness then the correlation with GDP is very weak, if at all existent. e.g. Bhutan is frequently said to be one of the world's happiest countries, and it's GDP per capita is around $3500. That's about the point at which your population doesn't need to go short of calories or die from lack of access to shelter. Once you go beyond that point...does electricity increase long term happiness? Does education? Does indoor plumbing? I don't think any of these things is clear. Why is the 1990s American (or 2020s Western European) material standard of living the ideal to strive for, rather than the 2020s Bhutanese, or indeed the 1790s American?*
On the other hand, people generally prefer to be richer than poorer, and if you respect people's preferences, then being richer is good actually. Also, wealth is power. European economic stagnation in the last 30 years has translated into a loss of hard power. Maintain that long enough and you end up at the receiving end of an opium war or a bunch of conquistadores or something, which is definitely bad (at least for you, obviously it's very good for the conquistadores). Meanwhile, America's robust economic growth over the same period has helped it mostly maintain it's hard power. Of course, as long as Europe can shelter under America's security umbrella this isn't a problem, but that umbrella may not be extended indefinitely.
* Relative to 1790s America there is a big difference in life expectancy, and certainly it is better to be alive than dead. But most of the gains there just come from childhood vaccinations, antibiotics, and an understanding of the germ theory of disease. e.g. Cuba at GDP/capita of 7k can match US life expectancy, and the aforementioned Bhutan is only a few years off. You certainly don't need modern Western European (or 1990s American) levels of economic development to get modern-Western life expectancies.
Oh, but the meta-question here really is: what makes sense for presumably intelligent people to think about for free?
Want to spend time thinking about how to
a) make bigger and better tchotchkes for people who already have plenty
rather than
b1) alleviating poverty at a world scale
b2) making energy cleaner and cheaper, and our use of it more efficient (PS: yes, of course I know that "energy conservancy" should be really "low-entropy energy source conservacy" or "watch your increases in entropy')
b3) discussing the pressing and real needs (health, education, clean environment) of people who have enough tchotckes (note: other people also have those problems)?
Be my guest! But I hope you are being paid for it, preferably by the people concerned :).
>After a few disappointing years, these are finally coming into their own. The expert I talked to said...
Scott, it was a pleasure to meet you at the conference, and to speak with you about driving automation. To be clear: no firm, yet, is explicitly proposing self-contained pods within automated vehicles, but I do think the only way we're going to get shared-use at scale is if SOMEONE does this.
If the light-touch approach that Waymo and Zoox (and to a certain extent Cruise) are using works, and we manage to avoid a suffocating blanket of regulation on the sector, then this model would be a promising avenue for a new entrant: even if 'incumbent' firms fail to offer a combination of an optimized-for-sharing vehicle, some upstart will: they could compete on price all by themselves, and more so if they can get a licensing deal with transit operators.
(BTW I write about self-driving and innovative mobility at my own Substack, CHANGING LANES, and would be delighted if readers of ACX joined my audience)
Would you like me to link your blog? I was going off the rule that we shouldn't name names of anyone we talked to, but if you're commenting publicly here then I might as well.
Please do! I think many people at the conference have good reason to prefer a cloak of anonymity, but I am trying to grow my audience, and wish as much (good) publicity as I can get
Another thing that might help with that is that governments may be more willing to tax robotaxi companies for congestion externalities than people driving their own cars (purely for vibe reasons), which would make the cost difference sharper.
Yes, that would help shared robotaxi vs. one-rider robotaxi... but it would hurt ALL robotaxi vs. their #1 competitor, the private car.
Such a move would be a really bad one and would make overall congestion worse. That said, it's QUITE plausible that a populist government would go that way.
The best outcome would be that as electric vehicles wax, and correspondingly the gas tax wanes in its ability to finance maintenance of roads and other transport infrastructure, that governments move to a global VKT (vehicle-kilometre travelled) charge. Cars today are pretty-much always connected to the Internet, which means collecting VKT data would be straightforward for a government that wished to do so. Such a charge could nudge in all sorts of ways: less for shared rides, more for solo; less during off-peak hours, more during peak hours; and so forth.
I'm imagining this would only meaningfully happen after robotaxis already have enough market share to be a significant cause of traffic (at which point their other benefits and momentum are enough to swing them over into being the primary mode of car use anyway).
Re VKT charge, you'd also want it to be location-based (downtown trips cause more congestion than rural ones), but that's an easy enough adjustment. If by the time these rules start coming up most people in cities use robotaxis instead of private cars anyway then maybe local governments would be able to pass a universal VKT tax without riling populist opposition.
Location-sensitive VKT could be integrated with legislation defining GPS navigation as mandatory safety equipment, and avoid populist opposition by also integrating automated speeding tickets and similar traffic enforcement. Less paperwork all around, takes away a cheap pretext for police harassment, yet makes the more serious crimes (hit-and-run, carjacking, etc) vastly easier to investigate.
>Location-sensitive VKT could be integrated with legislation defining GPS navigation as mandatory safety equipment
Gaa. That's pretty much tracking the location of all the people, all the time. I suspect that, at least in the long run, this may well be unavoidable, but it pretty much would remove the last vestige of non-online privacy.
A car's license plate, and where it is when it's driving on public roads, already lack reasonable expectation of privacy. There'd be no need (nor ready method) to record who exactly is inside the car, so someone inclined to travel anonymously could likely still find plenty of ways to do so.
In fact, the transparent impossibility of thus-equipped cars being stolen to useful effect - or even negligently misplaced - might allow the formation of a robust subculture of peer-to-peer rental or borrowing, and thus a sort of VPN equivalent. Means investigators would probably need a judge to sign off on digging through relevant records when the vehicle's owner wasn't inclined to cooperate, which seems likely to be an all-around improvement in due-process protections compared to the many ways a traffic stop can go wrong.
"Cars today are pretty-much always connected to the Internet, which means collecting VKT data would be straightforward"
This will, unless somehow otherwise accounted for, make older vehicles without the ability to track this data get subsidized. One possibility is substantial increases in annual vehicle registration fees, but that penalizes those who drive the least. Maybe a gas tax which is reduced based on VKT data?
The straightforward solution is get a (verifiable) odometre count at time T, another one at T+1 year, take the difference, and charge accordingly. This would be a remarkable pain in the neck, but that might be a feature to encourage non-connected vehicles off the road.
This approach works as a straight-up VKT charge but fails to do any nudging regarding congestion, but there's no straightforward way to do this absent a connected vehicle (or massive infrastructure investment).
In many countries, vehicles need to have a regular checkup to stay licensed (for example the TÜV in Germany), so for those taking the odometer value is easy.
Could you briefly summarize what mistake you think Tesla made vs. Waymo, etc.? I'd have guessed cameras vs. LIDAR, but I thought that cost was a major issue for LIDAR, and I don't really know much about the topic. Or maybe link to an article you've already written on the topic?
Another possibility would be using technologies already developed for dating apps to allow prospective robotaxi customers to efficiently select among their potential fellow passengers. Lots of people can tell at a glance who they would or wouldn't be able to get along with, while clever route-planning algorithms could calculate and display appropriate fare adjustments, allowing tricky externalities to be priced into a simple, elegant swipe right / swipe left decision.
This would lead to all sorts of interesting scenarios, like single women who spend all day shuttling back and forth between the two richest parts of town in the hopes of meeting a wealthy dude.
On the topic of "Humanity is very good at mass manufacturing things in factories" here's a fun anecdote from Neolithic Africa: we have evidence of humans mass producing bone tools (presumably to trade) as far back as the Paleolithic era. Elephant long bones were systematically broken to make blanks appropriate for shaping tools in what amounts to a Neolithic factory efficient enough to drive Paleoloxodons to extinction. You can learn more about it here if you're curious: https://doi.org/10.1371/journal.pone.0256090
Without re-reading SB1047 to check … if it had become law, and the hypothesised mass casualty incident had happened, and some of the people who got killed were in California, the liability provisions would apply even if the negligence that caused them to get killed were in, say, Texas.
New version of the AI alignment problem: the AI is only allowed to kil people if they aren’t Californians. This seems likely to be at least as hard as “don’t kill everyone”.
Alternatively, the AI is allowed to kill people, but if any one of them is a Californian, it has to kill ALL Californians. That might be slightly easier.
This is all great until you talk about factory farming. Under some reasonable moral theories, we are therefore declining. But Tyler, for example, has pretty weird takes on why animals don’t matter (see his podcast on 80k for the full view). I would love to hear what others at the conference (Pinker, for example) would said about this.
What effect should you expect to have progress (let's say increasing wealth) have on factory farming? I would guess Americans are rich enough, and meat cheap enough, that we don't have the simple relationship that more money = more meat-eating. So shouldn't we expect that more wealth gives us more money to spend on moral luxuries like non-factory-farmed meat?
I guess this goes into one of my issues with the progress movement in general (while I’m certainly on board with a bunch of things) — they don’t define progress in a way that can be measured! Progress for what? Progress in what sense?
Also, if you don’t think that morals are aligned with more money and you think something like the signaling theory of altruism is correct, the incentives can be wrong. Perhaps the future can be better. Maybe we will live in a future where we do solve this problem, and the incentives are better aligned - for instance, there is social stigma around eating meat and the alt protein industry does very well. This seems plausible to me. Under many moral theories, however, it still may not be worth the 23 million animals that are basically tortured to death each day.
I think it's mostly fair to think of them as talking about economic progress / GDP. I also think it's a fair first-order assumption - obviously not the whole story, but better than not making the assumption - that whatever problem you have, having more money is a good start to solving it, on the grounds that if people are only willing to devote a fixed fraction of their income to solving it, then the higher their income, the more money goes to the solution. If everyone was a billionaire, then the cost difference between factory-farmed meat and humanely-raised meat would be trivial, and people would give into animal rights activists just to shut them up.
The main exception is where wealth itself causes more problems. I think this might be true in the developing world where more money might translate into more meat-eating. I'll sort of plead the Pareto exchange defense here, but I recognize it doesn't work that way in the real world.
I still think EAs have a better proxy for thinking about good or progress, so I’m not really sure why we would want an only okay proxy if it takes away counterfactual resources from the better proxy -
On the other hand: perhaps multiple proxies are good because they’re a good way to hedge on values that other proxies would miss (as long as one of the proxies isn’t just the terminal values which could be the case with EA).
In theory, some mix of total utilitarianism and prioritarianism (I agree that things get tricky when you actually try to design a function that fits the intuitions that you want because of the impossibility theorems in population ethics, but the approximate idea should do much of the trick). In practice, while I’m not very well read on the topic, from what I do know it’s probably something like total QALYs.
I do think that these are clearly harder to measure than GDP, but if we want a better proxy, it’s probably this. I still stand by my point about maximizing for multiple proxies in case one misses something (either for systematic, moral theoretic, or practical reasons).
There are some things that will definitely be missed if you only optimize for GDP (e.g. animal welfare), and while it may even be a good proxy for more stuff in the future (the reason you described seems plausible and maybe likely), it still may not be a better total proxy because of all the damage done on the way (i.e. a billion factory farmed animals being killed every year).
Even if higher income does increase demand for meat/eggs/dairy, it's also true that non-factory-farmed meat tastes better and works better in many recipes (eggs and butter in particular). We should expect, in the limit of high enough income, that people will demand the higher-quality non-factory-farmed meat, eggs, and dairy products.
It’s true. I think this will largely be a cultural thing, though. Seems to be a large factor in things like what people eat (traditional foods often stay common in countries when there are better alternatives, for example -- though I admit that there are other factors at play here as well). I just hope the culture moves in the right direction (which I’m optimistic about!).
Maybe, but it's unclear that we would the increased wealth for a reduction in meat eating, or use the wealth for even more meat eating.
The factory part I think is irreversibly tied in with the actual wealth creation. The greatest source of increased wealth in physical terms, including food, is in finding more and more efficient ways of creating the physical products. Factory farming is a natural result of the process that makes things efficient. Other than mandating non-factory conditions (which seems counter to the Progress mindset and goals), increasing factory conditions seems inevitable.
Sufficiently advanced factory farming hits a tipping point where it abruptly becomes cruelty-free by default. Instead of raising, feeding, and slaughtering whole animals, better to build a machine which turns electricity and distilled chemicals directly into a pure, continuously-extruded tube of the most desirable types of muscle tissue.
It would. But factory farms are also part of the history of progress. We should pay attention to the fact that some past progress has been moral monstrosities, so that we are thinking about whether some current instances of progress might be similarly monstrous so that maybe we should direct progress differently. (Some might say that by definition, moral monstrosities aren’t progress. Still, studying the monstrosities is going to be relevant to distinguishing them from true progress in advance.)
Earlier, West Indies sugar plantations were a font of wealth created largely by improved ships and navigation and transatlantic trade. This was all impressive progress, but the sugar plantations themselves were basically hell on Earth and were morally about one hop above Nazi death camps.
With the increase in technology, I'd personally expect us to decrease in meat-eating. The current alternatives are 1) worse-tasting in various ways 2) different-tasting, which can be a big problem in converting people who didn't grow up with it 3) more expensive and 4) politically polarized.
Still, once we have the technology to synthesize better meat in terms of taste & cost, I don't expect factory farming to linger for a long time. They'll latch on and try to make it harder, but I expect them to fade away.
Ex: Climate change. Oil companies have more power & state-backing, as well as the problem being harder to substitute. I can't just replace the oil in my car with electricity without a large investment, even if the electricity happened to be cheaper the car is not. This makes so the people who are most price-sensitive (buying used cars from 1990s, 2000s or early 2010s) can't just switch to electricity even if it was quickly cheaper.
Food is less like that. If there's a good cheaper meat substitute on the market, then that makes it far easier for people to individually switch over.
That, and the technology for synthesizing products in that manner likely being useful for a lot of other food, makes me optimistic about meat production.
I think there are good arguments for this (and I appreciate you bringing these up!), but like I said in response to Scott, the object level arguments miss what I’m saying. Technology in this sense entirely leads to progress always is clearly missing a lot. The goodness of technology and science is that they are contingent on other facts about the world. While it may be a proxy for values much of the time, it should also be viewed as a proxy that is quite removed from the terminal values we have.
One example of this might be that without technology we would never have been able to systematically kill animals - maybe the bad we do now outweighs the future good, even given that we stop at some point in the near future (20 million animals being slaughtered every day might be good reason to believe this).
Another thing here might be that giving humans more power might lead to them killing themselves with existential risks (see Bostrom’s Vulnerable World Hypothesis paper for a good form of this argument). While I certainly hope humans do not do this, the chances are not small enough to make it negligible or say that we should just rush technology and science without thoughts about what is it that we truly value.
Solar is amazing when it’s a small percentage of overall capacity. When you are actually trying to replace a modern power plant with it, like a coal power plant, it’s incredibly expensive.
Capacity of solar often depends on assumptions of great weather, or average weather, when we really want electricity whether the last week has been average, or particularly cloudy. You also need to factor in the monumental cost of batteries, as we want electricity at night too, and when we consume electricity is usually independent of when it’s produced.
In reality, while it’s fine to treat cost per kWh when dealing with small solar plants, as soon as solar gets to be above ~10% of the energy mix, or is used to replace existing fossil fuel plants, the actual cost to produce a robust system that’s almost equivalent (but still worse in many ways) to their energy generation capacity is 10-100x as much.
In Germany electricity prices are going negative more and more often. Meaning not only does solar not make money, fossil fuel plants that still need to run for the night actually have to pay to keep running, as they can’t easily shut down and power up without incurring significant cost. This raises fossil fuel prices, which may work further to incentivize green energy adoption, but the more negative prices it produces, the more we’ll have to pay for that throttle-able capacity.
It’s not an insurmountable problem, but a grid based on solar (or wind) would be 1-2 orders of magnitude more expensive than cost per kWh might make it seem.
"In Germany electricity prices are going negative more and more often. Meaning not only does solar not make money, fossil fuel plants that still need to run for the night actually have to pay to keep running, as they can’t easily shut down and power up without incurring significant cost. This raises fossil fuel prices, which may work further to incentivize green energy adoption, but the more negative prices it produces, the more we’ll have to pay for that throttle-able capacity."
Explain like I'm an idiot: why do electricity prices go negative? Is there no way to "discard" unused electricity?
They can discard used electricity in extreme circumstances, but generally it’s far preferable to either increase demand, or attempt to decrease the supply. The grid must be almost perfectly in balance at all times, which means keeping demand and supply in step.
The negative electricity cost usually only lasts a few minutes to an hour, around noon on sunny, windy days for example. For power plants, that’s a problem because it costs them a lot of money to shut down and start back up, and a lot of time to do so, which has its opportunity cost. Since they cost fuel to operate, their profitability is usually at a certain price above $0, although fixed price or minimum price contracts complicate that calculation.
Renewables basically cost $0 to operate, and maybe even cost little bit to curtail as that requires more complicated systems. Thus, the solar farms will keep sending the maximum amount of electricity they can produce to the grid so long as the price is positive. Unlike power plants, that will reduce production when prices are low, solar is completely demand-inelastic for all positive prices. Better to make $0.01 per kWh rather than $0 after all.
Negative prices are the way to de incentivize this oversupply. For the power sources that have high shut down/startup costs it makes continued operation less appealing, and for solar with no marginal cost it actually has an effect on that normally inelastic demand. Another solution power grids take is by literally paying producers to reduce production, which is less likely to cause economic troubles for the producer, but directly translates to higher electricity prices for the consumer.
It’s a signal that at times, we already produce too many renewables for our current demand. If we doubled the solar capacity, negative prices would stay for longer, and lower prices during peak solar production would also last longer. Profitability on solar originally assumed normal pricing during solar’s peak production, but as they actually get to be a significant portion of the grid, their $0 marginal cost will bring the price people are willing to pay for that electricity down to $0.01 when solar supplies most of the grid. This is good for consumers, but doesn’t bode well for justifying exponentially increasing solar investment.
The obvious way to get more demand in the system when the wind is blowing or the sun is shining is to reduce retail prices when this is happening and alert people via a smart app. Instead the market pricing is at the wholesale not retail level, so consumers have no incentive to use carbon free electricity.
There was a power company that just passed grid prices through to the customer.
But the impression I got was that they hosed themselves by continuing to provide power at grid prices to their customers, who were contractually obligated to pay for the power they consumed, but who didn't actually have the money to do that. This left the company with gigantic liabilities and no prospect of collecting the revenue that was notionally supposed to cover them.
So did every other power company at the time, the only difference is they nominally passed the cost to their customer while at the same time telling them not to pay it, while the others billed the usual $1 against the $1000 cost of power. Everyone had to swallow the emergency pricing. But if all the companies had actually been passing the high price to their customers, then price signals might have done their job and caused enough voluntary decrease in consumption to prevent the emergency.
In Germany, there's at least one retail electricity supplier offering dynamic pricing based on live market prices. The problem is: As a consumer, what are you going to do with this? Maybe you can schedule your dishwasher or dryer to run at 13:30, but the major loads like heat pumps or electric cars aren't as flexible: Heat pumps want to run in the early morning, to re-heat the house after a cold night, and electric cars have to be be recharged when they're at home, which is typically in the evening and at night.
I think the market still needs time to adapt. Employers can offer workplace car charging as a cheap perk. You can heat up (or chill?) a big tank of water as a kind of thermal battery if cheap electricity supply doesn’t correspond with home heating/cooling demand.
My point is that the biggest part of consumer demand is relatively inelastic and can't be shifted around much. If I charge my car while at work, I'll want to charge it whether the sun is shining or not, lest I pull an involuntary all-nighter.
If electricity prices were negative right now then I'd turn on all the heaters in my house. Then I'd turn on all the air conditioners in my house. Then I'd open all the windows just in case one turns out to be a lot more powerful than the other.
Professional arbitrageurs operate at the wholesale level by investing in batteries, then outcompeting peaker plants at grid stabilization services. https://caseyhandmer.wordpress.com/2021/05/20/the-unstoppable-battery-cavalcade/ Given a few more years, I'm sure many other intermittency-tolerant industrial applications will spring up to take advantage of a literal free lunch.
Sadly this is entirely depend on local geography and therefore very limited. My country for example has reached its peak when it comes to hydro decades ago. It's not a solution.
These are two different things: some countries do exploit existing hydro resources to the fullest, but that doesn't mean that gravity batteries have been exploited to the fullest, or that gravity batteries would be very limited. What the actual limits to gravity batteries are - now that's a good question.
Gravity batteries that do not already use existing bodies of water are just limited in the same way as most things: they are too expensive and therefore currently pointless.
Can I infer from this that the current limits on storage capacity are entirely in the amount of energy that can be stored (or the ability to ramp up storage quickly) and not at all in the efficiency of storage?
Jenny Chase seemed to suggest that the reason we don't just run a desalination plant with any energy spikes is the cost of the desalination plant?
Storage efficiency is somewhat of an issue, as round trip efficiencies tend to be like 60-80% depending on the method. But yeah capacity is the bigger problem.
Desalination plants are expensive enough that in order to amortize the capital cost, you want to run them with a high capacity factor. The same is true for plants that produce hydrogen by electrolysis. Also data centers, aluminum plants, etc. Using electricity in productive ways tends not to be cheap. That makes it tricky to increase demand during production spikes.
How expensive are desalination plants? Google's bot says that the capital cost is 500 days of operation. That sounds very low to me, allowing a low capacity power and timing of electricity prices. Do you disagree with the factual claim of 500 days of operation or do you think that is a high capital cost?
Well, let's see if we can find a good example. Ras Al-Khair Power and Desalination Plant, completed in 2014, is the world's largest hybrid water desalination plant, according to wikipedia (hybrid apparently means it uses both reverse osmosis and multi-stage flash distillation; I'm not sure why). But it also includes a combined cycle natural gas power plant, so we have to separate that out.
EIA gives a combined cycle gas cost of $614/kW for 2015, for just construction.
or from another source, $950/kW for an H-Class 1100 MW Multi-Shaft Combined Cycle - Multi-shaft configuration rated 1,083MW and 59.4% efficiency, $958 million total (950 $/kW installed) and 12.20 $/kW fixed O&M cost.
Going with the middle figure of $950/kW, we can compare to Ras Al-Khair. Its total cost was $7.1 billion. The 2400 MWe of gas generation, we would expect to cost about $2.3 billion. So the desalination part is about $4.8 billion, give or take maybe 500 million.
The throughput is 1,036,000 m^3/day, so capital cost is roughly $4600/(m^3/day). You know, plus or minus a factor of 2.
If we take $1.50 as a reasonable midpoint and use simple division, that gives a payback time of about 3000 days (~8 years). In practice, you would have to subtract off any fixed operating costs as well as the cost of electricity, and then compare to the risk free interest rate. Even if electricity is cheap, at first glance it doesn't look like a great investment unless you can operate at a pretty high capacity factor *and* sell the water at a pretty high price.
As for actually discarding electricity (which basically never happens for actually curtailing excess energy), turning it into heat is the only real way. That usually involves turning on what are basically big space heaters called Load Banks: https://en.wikipedia.org/wiki/Load_bank
Actually discarding excess electricity isn't usually part of normal operations though, as it's far easier to increase demand from consumers who are price-sensitive, or reduce supply directly at the plant.
As the original post writes, a lot is to do with how power plants run, such that they can't just turn on and off whenever they want without incurring cost, so it's cheaper to pay end consumers to just use electricity. Also, outside of Texas most energy markets will have a capacity market, paying generators to keep plants online in case of spikes in demand.
Most grids will have some capacity for using up excess electricity, but not massive amounts; the aim is for generators and retailers to balance supply and demand through accurate forecasting.
Electricity prices go negative for two main reasons:
- certain generators are paid whether the power is needed or not (eg the German feed-in tariff system means that solar and wind will get paid if generated). This is basically solvable by changing policy that was designed for lower generation numbers.
- it's worth keeping some generators online for later. As I recall this was a problem in California for some years, because the gas generators were needed in the evening, so had to keep spinning in the day to be ready - and hence generated some power, paying negative prices. I'm pretty sure this is now alleviated by batteries.
The only way to “discard” unused electricity is to put it in a battery, or run some powerful appliance or something. If the grid has too much electrical power running compared to how much is consumed, it messes up the voltage, which can damage lots of devices.
That’s fine if you always have excess electricity - but even places with excesses sometimes usually have the opposite issue. What you need is something you can turn on and off fast enough to balance generation and consumption. Gas power plants have been the common solution on the generation side. Batteries are a good option on both sides. Resistors seem like a backup option, but hopefully there’s something better.
Note that aluminum refining by electrolysis is essentially half a battery, like the recharge phase of a rechargeable battery. What I don't know is whether it can be done flexibly...
The issue is that the German government committed to pay a certain amount for (parts of) solar energy regardless of market prices. So this part of supply does not react to price signals.
On the consumption side, there will soon be ways to take up the surplus energy. Several companies have noticed the price fluctuations and are now building huge battery farms (in the order of GWh), to buy up the electricity when it's cheap (or in the extreme, negative), and to sell it when electricity is expensive. Once they are ready (next year or so), the negative electricity prices will disappear.
Fun fact: some of the battery farms are built right next to the nuclear power plants that Germany switched off, because those places have the connection to the power grid already in place.
Have you looked up the recent trends in Australia, especially South Australia? Solar there is something like 70% of total annual production. They've been working with Tesla to install lots of Powerwall units as a distributed virtual power plant to smooth out demand. The long-term official plan is to overbuild solar to 500% of electricity demand, and consume the excess for charging EVs, electrifying various industries, and manufacturing synthetic fuels and chemicals.
Edit to add: Phase change materials for HVAC are another way we can profitably time-shift demand to whenever electricity is cheapest (either times of peak renewables output or overnight) or operate the system at a lower, more efficient heating or cooling rate without loss of comfort. We're really only at the very beginning of implementing ways of matching demand to supply, because historically we didn't need to.
I think you’re vastly overestimating the solar percentage in Australia. Solar is currently at 16% of total production according to a quick google search.
Eventually they probably will be capable of overbuilding to that extent, but that involves 5x the cost a pure kWh comparison wouldn’t give you. That, plus the extremely expensive batteries to have multiple days worth of storage for the entire grid, and the total cost of it is going to be significant.
The problem isn’t impossible to solve, but it is significantly more difficult and expensive than “solar is now cheaper than coal” seems to imply.
You're right, it's 75% wind+solar and not just solar. Solar PV is 24% as of 2022. My mistake, sorry. The fastest growth is in solar, though. And yes, it *is* much more complicated and difficult and expensive than just saying "solar is now cheaper than coal." Solar started being cheaper than coal in Chile in 2015 (approximately), and only now are all the other bottlenecks becoming feasible to overcome in optimum markets.
The rest of my comment is pointing that there are a lot of efficiency-improving, money-saving things people and companies can do today, and are starting to do, that look like energy storage (dispatchable opportunistic demand?) to the grid if you manage them properly, long before you get to building grid-scale stationary energy storage systems. We're not dumb enough to just blindly pay for 5x as much electricity as we need and build week-long battery energy storage, but yes, that would be the implausible upper bound. It remains to be seen which markets are clever enough to actually be smart about it and find the lower bound.
I can’t find this 75% number. If it were true it would indicate Australia has had to contend with this problem and done so successfully. The Australian government says solar is 16% of total generation and wind 12%.
Wind doesn’t have the same problems as solar, but it has that are equally difficult its own due to the unpredictable nature of it when it makes up a significant portion of electricity generation.
It looks like the majority of their renewable generation is from wind: https://www.statista.com/statistics/985429/australia-energy-generation-capacity-in-sa-by-source/ which doesn't have the same issues as solar in that the generation is more randomly distributed. It seems they are also maintaining full natural gas capacity, even if they rarely use it. Current electricity cost per kWh seems to be ~$0.34, or even as high as $0.45, depending on the source.
This tracks with the discussion though. Wind doesn't have the same limitations as solar, but neither the rapidly decreasing costs. It only works in windy areas (and wind is pretty consistent over the entire year), but still requires almost full capacity from non-renewable sources (in this case natural gas). The end result is some of the most expensive electricity in the world.
Personally, I'm cautiously optimistic, but highly doubtful the SA model can be replicated anywhere besides high wind, low population density areas of the world. The immense electricity price tag isn't exactly tantalizing either.
A lot of people like Nuclear because it requires no additional non-renewable capacity, is completely throttleable and has low marginal cost of production, leading to very low electricity costs depending on the financing structure during construction.
South Australia does have 75%, but they have higher prices and rely on the larger grid in a similar way as Scotland
The Net Zero Australia report is favourable enough to wind and solar that they think there's nearly no case for nuclear, yet they also project that energy costs would rise from 6% to 8% of GDP (not certain, but very distinctly recall it being two percentage points)
A real energy revolution from wind and solar would have to involve batteries doing incredible things, which they could based on solar
Yes, the state has interconnects to other states that can export up to ~20% of peak power output currently, and yes they're trying to increase that further in the future, because why wouldn't they? You solve the easy problems first. Heck, there's talks of building multi-GW undersea cables from Darwin all the way to Singapore, because that's slowly becoming a viable way to decarbonize Singapore. Long distance interconnects increase maximum penetration of intermittent resources locally *and* overall by reducing output variation due to weather, time of day, and other factors.
As for the GDP prediction: How much of that is increased prices, and how much is increased demand? Australia has a lot of heavy industry it wants to electrify, and then of course you have to add in EV growth too. And while I have my doubts about a lot of specific projects and timelines, they're genuinely trying to increase green hydrogen, ammonia, methanol, and other chemicals production for export, which would eat up quite a lot of electricity production (and, assuming they're responsive to real time energy prices, would help minimize required battery storage).
As another example: see how the Port of Melbourne is trying to become a green methanol bunkering hub for methanol made in Tasmania (to start with). There are shipping companies ordering dual-fueled and fuel-agnostic ships, and this is an incremental step towards using one problem (high renewables penetration) to reduce another (shipping emissions).
I'm not saying everyone can do exactly what they're doing. I'm saying there's a path, and under optimal conditions the path is finally becoming traversable, and the number of obstacles that get in the way is steadily going down. Making the path viable in more places still requires tech advancement and an enormous amount of hard work, but nothing so revolutionary we can't see a way to it by normal engineering and investment processes.
I seriously doubt multiple days worth of batteries are necessary. Single-digit hours of peak capacity is enough ninety-some percent of the time, and seasonal variation or the occasional week of randomly bad weather can be planned around in other ways.
You’d definitely want to look more into it then. Unless you’re comfortable with accepting rolling blackouts, you either need other capacity (like natural gas in the case of the Australia example) or a very large amount of batteries to be confident you won’t be running out of energy. You need enough capacity to provide from the afternoon, to the upcoming morning, which is generally ~12 hours. Much more in the winter for northern hemisphere countries since the days are shorter and provide much less energy.
Alternate generating capacity is part of what I mean by "planned around in other ways," yeah. Synthetic hydrocarbons are just another kind of battery. Other part is retooling energy-intensive stuff to make hay while the sun shines.
Four hours of batteries is enough to shift peak supply at noon to peak demand in the evening, which seems like a good start, and twelve hours is still a lot closer to "single-digit hours" than it is to "multiple days." Doesn't have to be perfect, or all paid up front - arbitrageurs will gladly spend their own money to build more batteries so long as daily price cycles let them rapidly recoup the investment, while battery and solar-panel manufacturers will keep finding ways to make more, better, cheaper, so long as the arbitrageurs keep buying. No need to subsidize all that, it'll keep spinning self-funded until short-run variability is as smooth as it can reasonably get and overall demand for electricity is saturated.
Synthetic hydrocarbons can function as a kind of battery, but they are very inefficient, they only give back about 30% of the energy you put into them. They only make sense for energy intensive applications that can't use anything else.
I think you're underestimating the range of tools available to mitigate the problem. My own take is that we will need "days" of "energy storage" but that this looks very different from what either you or JamesLeng are describing.
First, you obviously understand that the difficulty here depends a lot on latitude and local climate. I assume remote, small, and near-polar communities will probably need generator backups at a near-1:1 rate, but they are so small it doesn't matter climate change-wise if they do keep burning fossil fuels, and it doesn't matter energy-demand-wise if they need synthetic hydrocarbons or hydrogen or whatever. So let's assume we're talking about a typical middle-latitude northern-hemisphere city with a mix of sunny, rainy, and (in winter) snowy weather.
In 2070, after all currently existing assets have been replaced, what does its grid look like?
Well, first, it'll be a mix of all the available energy resources. Let's ignore that for now, though, and assume it's all wind and solar, even though no utility in that kind of climate is stupid enough to do that. Let's say the worst case is you need about four days' worth of energy storage, average load.
You'll probably have 4-8 hrs of lithium (or by then, possibly sodium) ion, which will get you through a typical night and day. Then you might have flow batteries for another 4-20 hrs if anyone manages to scale them the way that the fundamentals suggest should be possible. If not, maybe more lithium, but probably not more than 12 hrs total. Most likely near the low end of both ranges, since solar peaks during the day and wind at night, wind and solar are somewhat anticorrelated, and PV is cheap enough that you can overbuild enough for average capacity to be sufficient even in winter. This will get you through 95% of the year or more, assuming you do enough overbuilding of cheap solar panels to deal with a typical winter day.
So what's next? Well, realistically any actual grid will have some proportion of hydro, geothermal, waste-to-energy (gasifiers are 4-5x more efficient and much cleaner than traditional boilers attached to incinerators), and if we're lucky nuclear. In many regions this will be enough to stretch the above storage to the point that you only need more a couple of times per year, if at all.
Elsewhere, we need to consider that "average load" now includes a whole lot of electric vehicles. I know people hate the idea of vehicle to grid because they assume it'll be done badly, which, fair enough. But a lot of people are going to be willing to get paid to partially discharge (or just not fully charge) their cars a few times a year, as long as you warn and/or politely ask them the day before. An average EV stores the equivalent of ~2-4 days' worth of power use by the average house, and many households will own 2 or more. So if 50% of people let the grid borrow 50% of their EV capacity a couple of times a year, that gets you an extra ~1-2 days of distributed storage.
If that's still not enough, then consider that heat pumps and heat storage and phase change materials can load-shift HVAC and hot water demands for residential, commercial, and industrial (including process heat) users. Companies are already building these in some places because even today it pencils out economically.
And if that's still not secure enough, consider that we're now talking about problems that will occur once a year or less. At that point, you can in fact just pay people (especially commercial and industrial customers) to use less power, and this can be done at much bigger scale than we do it today - especially as we electrify more and more of our industrial process energy needs.
And if even that isn't enough for your standards, you're more than welcome to buy a generator you'll use once every couple of years. Or buy very cheap multi-day batteries that, like Form Energy's, charge slowly over months to be ready for the rare occasions they're needed.
It’s a nice thought experiment based on vibes, but that isn’t really a useful critique. “We’ll probably have these technologies and they’ll probably be enough” doesn’t cross the gap from theory to reality.
You need to be able to have a functioning grid for (at minimum) 3 cloudy, low wind days in the winter. That means either having a very large amount of full batteries on standby containing days worth of full grid power (possible, but extremely expensive), or a throttle able energy source like fossil fuels, nuclear or hydro.
The suggestions you make will certainly help, but they’re the sort of small improvements on the grid that won’t make a meaningful total difference in creating a renewables based robust energy grid.
We must also keep in mind that in places where solar barely makes up double digit percentages of generation capacity, we are seeing negative prices at peak solar. I wonder how the economics of a new solar plant will work out when they literally can’t sell energy at peak production for more than a tenth of a cent.
At the end of the day, 2070 is too far to make any meaningful predictions or even have a useful discussion about. I’m talking about near term problems with solar that are effecting the world today, not the hypothetical technologies in 2070. Energy prices are going negative, and current examples of renewables-based grids require significant base load that’s usually taken up by fossil fuels plants on standby. Only countries fortunate enough to have ample hydro are able to cross that bridge to majority-renewable electric grids, and that’s not a repeatable thing. It’s great people are thinking up potential solutions, but as of right now they are just ideas, and we aren’t close to solving the baseload capacity problem with non-throttleable renewable energy.
Batteries could alter that very significantly or entirely. While I get the impression that physically they don't have the same low floor as solar panels due to just involving more material, according to ChatGPT batteries and panels seem pretty similar in weight per capacity (8 hours of batteries vs solar that can likely produce for 8 hours)
(*) What they would get paid in a spot market vs the average price due to high supply when they are producing. As mentioned in another comment the challenge with just building things to soak up the supply is the capacity factor of high capex electrolyzers or the like. Energy costs might be 40%, if you make that free you gain 40% but if you run 50% of the time your capital cost is doubled. Basically energy isn't a single factor unlock, although it could grease the wheels
Another major consideration is energy production throughout the year. At least in Germany (and presumably even more so the farther north you get) solar production in the winter is almost nothing. Thus, you still need to maintain a completely functional grid without the existence of solar, batteries or not (unless we’re planning on building 10-20x capacity for total demand).
In essence, we have to pay for throttleable energy sources that are independent of the weather that can satisfy demand no matter what. Those plants have very high fixed costs, so growing solar to a significant portion of the market just means the cost for electricity outside of solar production will be significantly higher. Arguments for solar over nuclear usually don’t take into account what would actually need to be built for solar and wind to replace fossil fuels.
Not anti-solar or anti-renewable btw, I just don’t see how we can actually have a renewable grid with current trajectories. As far as decarbonization goes it’s great, but certainly won’t get us anywhere near net zero.
Moved nowhere since 2009. Expensive (the Mediterranean is deep), plus being dependent on countries like Algeria for electricity doesn't sound too attractive. Spain could work, but the political/cultural differences and mutual distrust between Europeans and Arabs are too strong.
On the number of pages in the Federal Register -- that can be a very loose proxy for the regulatory burden, but here's something I wrote in 2003 about how loose that can be: https://stuartbuck.blogspot.com/2003/12/regulation_31.html
To quote:
Many final rules do not create brand new regulations in areas where no regulation existed before. Many, if not most, final rules simply replace a pre-existing rule by modifying it in light of changed circumstances or new research. To take an area with which I'm deeply familiar, the FCC issued an new telecommunications rule in August called the Triennial Review. It was massive. But all that the Triennial Review did was tweak telecom regulations that already existed, and which themselves already took up several hundred pages (including accompanying explanations) in the Federal Register from previous years. In fact, the effect of the Triennial Review was to loosen, if only slightly, those pre-existing regulations.
Imagine this hypothetical: The Federal Fruit Commission issues a Notice of Proposed Rulemaking in 2004 asking for comments on whether it should ban the sale of green bananas. This takes up 50 pages in the Federal Register. In 2005, the FFC issues a final rule banning the sale of bananas that are more than 75% green. The rule itself is two lines long, but is accompanied by 100 pages of explanation. In 2006, the FFC has realized that the grocery industry hates the rule. It issues another NPRM of 50 pages asking whether it should change the rule to apply to 50% green bananas. In 2007, the FFC issues another two line rule, with 100 pages of explanation, changing the previous rule to read "50%." In 2008, the FFC realizes that people are still unhappy, and that the rule doesn't really accomplish any public good. It issues another 50-page NPRM asking if the rule should be eliminated. Then, in 2009, the FFC finally gets rid of the green banana rule, along with 100 pages of explanation.
If you looked at pages in the Federal Register, you might say, "Oh my goodness, the FFC has produced 450 pages in the past 6 years just on green bananas. Regulation is out of control!" But nowhere are there "450 pages of rules" that apply to green bananas. In fact, at that point, the green banana rule would no longer exist, and the regulatory burden would be zero. Of course, the 450-page number might be useful if you want to approximate the societal effort spent on dealing with federal agencies -- i.e., monitoring rulemakings, writing comments, etc. But "450 pages" would not represent the actual regulatory burden on anything.
If that example seems fanciful, there's a famous administrative law case in which the FDA spent literally 9 years debating back and forth over whether "peanut butter" should be 87.5% or 90% peanuts. (!) See Corn Products Co. v. Department of Health, Education & Welfare, 427 F.2d 511, 513 (3d Cir. 1970). These hearings produced 7,736 pages of transcript -- all over the peanut content of "peanut butter." For all the Federal Register pages expended in that proceeding, there was only one very simple rule that kept being modified.
> Of course, the 450-page number might be useful if you want to approximate the societal effort spent on dealing with federal agencies -- i.e., monitoring rulemakings, writing comments, etc. But "450 pages" would not represent the actual regulatory burden on anything.
Keeping abreast of all that revision churn, constantly needing to re-do the cost-benefit analysis of disrupting established workflow for compliance with the new rule vs. evading or corrupting the inspector vs. fighting a legal battle to get the rule changed again, IS regulatory burden.
About supersonic flights: this one is probably a technical problem rather than regulations. Aerodynamics are so different once you go supersonic that you're basically building a whole different aircraft, and one that's way more expensive to fly (see how expensive fighter jets are per pound compared to passenger planes). And most good usecases for supersonic flights are transatlantic or transpacific anyway, so the rule about sonicbooms over land probably wouldn't completely block it.
(See also the construction physics post on the history of supersonic passenger flight attempts).
Does anyone have a good summary of why concorde stopped? It went between London and Washington, D.C. or New York City.
From wikipedia;
>They cited low passenger numbers following the 25 July 2000 crash, the slump in air travel following the September 11 attacks, and rising maintenance costs: Airbus, the company that acquired Aérospatiale in 2000, had made a decision in 2003 to no longer supply replacement parts for the air
Seems like it was for a reason other than "makes peoples windows rattle". Maybe the 2000s were a perfect storm against supersonic flights anyway.
"the Concorde used a series of pumps to move fuel between different tanks to adjust the location of its center of gravity (the fuel was also designed to act as a heat sink, keeping the aircraft cool at supersonic speeds"
Quick google (not high confidence) suggests that there are something like 50-100 flights between LA and NY every day.
If its' true that supersonic flight cuts that trip from 6 hours to 2 hours, then you only need 1/3rd as many planes in the air to make those trips. Is the cost increase for supersonic flight more than 3x the cost (per seat I guess) of existing planes?
You don’t quite get to cut down to 1/3. Planes need an hour or two on the ground between flights to deplane, clean, and board. Cutting from 7 to 3 is closer to reality. (And given that the supersonic speeds only apply to the cruising part of the flight, and not the takeoff or landing, it might be more like cutting 1 hour ground plus 1 hour takeoff landing plus 5 hours air to 1 hour ground plus 1 hour takeoff landing plus 1.6 hours air.)
European ULCCs (particularly Ryanair) have managed to get it down to around half an hour on average. If a supersonic aircraft has similar capacity to Concorde then the deplaning and boarding could be done in a similarly expedited timeframe as Ryanairs 737s. If they are built to 777 capacities then you of course end up with longer turnaround times.
The Concorde cost about 3-4 times as much per passenger seat to operate as the 747 and about ten times as much as the 737, so yeah, it still doesn't pencil out unless passengers are paying a large speed premium.
(This is just counting operating costs, not development costs. Development can make planes much cheaper - as demonstrated by the cost gap between the 747 and more modern 737 variants - but is much more expensive for supersonic jets, so that adds another cost barrier for them).
(to clarify, one reason the 747 is so much more expensive to operate than the 737 is its longer range; the Concorde's range is closer to that of the 737, so in practice that 3-4x factor is still underselling it)
Energy scales with the square of velocity, so if you are going 3x as fast you 9x the energy, so 9x the fuel to reach cruising speed. But drag scales with the cube of velocity, so you need 27x the energy to keep yourself at that speed. You can reduce that a bit with more aerodynamic design and by flying higher, like Concorde did, but fundamentally supersonic flight is going to need a lot more fuel. That's where the extra cost will come in.
This is a little chicken-and-egg, no? It's less cost-efficient to do supersonic passenger flight, at least in part, because we haven't devoted billions and billions of R&D dollars into it in the way that we have for subsonic passenger flight.
Also, the fighter jet comparison isn't useful here unless you can show two things:
1. Military passenger aircraft have similar cost per flight hour per pound to civilian passenger travel
2. Subsonic fighter jets didn't cost more to fly per pound than subsonic passenger aircraft
I'm pretty sure 2 isn't true but I don't have any data I can point to, and I'm certain that 1 isn't true.
Yeah, to be clear fighter jets are a directional gesture, they're more expensive to fly for multiple reasons.
We can use the Concorde as a better comparison baseline (it had a *lot* spent on development costs). It's nowhere near fighter jets costs but it's still about 10x the cost per passenger mile as a 737.
In my opinion the market for scheduled supersonic flights is a bit limited because it competes with both subsonic first class and subsonic private jets. Supersonic which still requires getting to the airport 2.5 hours early does not save much time over a private jet which requires getting to the airport 10 minutes early, and there is yet more wasted time in having to fly on the airline's schedule. And you can have meetings and get work done on a private jet, so your flight time is productive. And far more comfortable.
Compared to first or business class, you'll have more departures to choose from subsonic, and a much more comfortable work environment.
That leaves private supersonic as a potential market, which would be viable I think except that politically it would be a non-starter in a lot of places.
The other problem with Concorde was short range -- it did the trans-Atlantic hop which only takes five hours subsonic, not long enough to be annoying (especially in first class).
A supersonic jet which could do a 14-hour trans-Pacific flight in six hours, that would be more tempting. But not necessarily physically viable, let alone economically.
The SpaceX concept videos showing Starship doing point-to-point travel, anywhere to anywhere within an hour... that might actually be a more viable future than trying to push through the air at Mach 2.
Yes, and especially with things like Netjets and JSX democratizing and making the private flying experience cheaper, it'd be really hard for supersonic to compete.
I really want to understand the psychology and motivations behind people who block everything when there is no obvious financial interest.
There is a whole ecosystem of activists and regulators, who block houses because of their effect on pond life or solar farms because newts don't like them. They must be self aware enough to know those are ridiculous and it is not as if the newt lobby is funding lawyers and activists or influencing officials.
Most people do most things for interests that aren’t immediately financial. If someone is doing something that you see as a threat to your way of life, you’ll want to oppose it, even at some cost. Doesn’t matter whether that’s killing an animal, or putting a shadow from an apartment building on your vegetable garden, or teaching your kids that they can change their gender, or trying to replace your government.
I wasn't only talking about financial interests. I can understand situations where there is some tangible benefit to a human or charismatic mega fauna but there has to be some limits to circles of concerns.
Because some people believe pond life has value. You don't, that's cool. But they do. And it's profoundly unhelpful, this constant reversion to "these people want things for no reason."
No one is a nihilist. They just don't share your values.
They can be wrong because of facts they don't know or have failed to consider, or they can be wrong because their values are wrong. I guess you're saying their values are wrong, and that is why they are choosing a pond over a solar farm. I prefer the hypothesis that their values are good, but they don't know the facts, or they have failed to consider them.
Some people actually do value the environment. And some people are cynical nimbys who declare a laundromat or parking lot "historic" so it can't be built on.
Because it's easy enough to think that there's no harm done and that the houses or solar farms can simply be built somewhere that *doesn't* have newts.
Also, if you're the chosen and known Newt Crusader, you basically have to fight every newt battle there is, no matter how trifling, because there's a little voice at the back of your head saying that if you start going "Well, maybe the newts can take it on the chin this time a little bit", then the screw-the-newts people will go "Look, even the Newt Crusader agrees that the newts can suck it! Kill the bastards!" ("Newts" can, indeed, be a great deal of things here - such as abortions, guns, AI models, countries within a military alliance etc.)
My sense is that mostly people object to things over aesthetics, and the pond life or newt stuff is just a tool to block the thing, not the actual reason for wanting to block it.
I have also often encountered the mindset of "if it doesn't benefit me, then I am against it" — basically, the attitude that one's political power to object can be used to extract some concessions of whatever kind.
What if you can get giant airships to autonomously hunt down other airships, to steal its cargo or something if you want a purpose more than the plentiful lulz? Robotic sky piracy sounds pretty awesome.
If those giant airships are moving hundreds of tonnes of cargo faster than ships and cheaper than airplanes, are carbon neutral, and are being moored, loaded, and unloaded robotically, that seems truly wondrous.
Though I'd want a manned one for my own personal use also, of course. One that's nuclear powered and can dispatch a submarine.
I love the concept but I don't think the economics pencil out. They're only a little faster than ships at significantly higher cost, and it seems like ships generally could go about that much faster if they needed to but it isn't worth the extra cost. They can reach places ships can't, but there they have to compete with both trucks and trains. There are probably a few places where you could make it work because the transport infrastructure is bad or nonexistent.
And the associated fine dining, of course! Slow food as it takes at its fastest two days to travel between Europe and America, but who needs to be in a hurry when travel is this luxurious?
"Mary Day Winn of the New York Herald Tribune wrote: ‘ The real glory of Zeppelin travel … is its freedom from seasickness. It is the smoothest form of motion I have ever known, just a continuous floating, with no rolling, no dipping, and almost no change of levels. The sound of the engines can be heard only faintly – a low, steady murmur barely entering consciousness except when it slows up. There is no vibration.’"
The obvious tourism market for Airships to me always seemed to be air cruises over land. Instead of being slower point to point travel (where it is always going to lose out to the convenience of fixed wing jet aircraft) have an air cruise over the Grand Canyon stopping off at major tourism points like a cruise ship and so on. There is clearly a demand for cruise ships but not for ocean liners, so why not do the cruise ship market but in places with no ocean! If I had access to the huge amounts of capital that would be required that would be my business venture.
I think that would be a great idea, you could catch the 'luxury travel experience' market, and it would be 'green' travel by comparison with airplanes and helicopters.
The main problem is the safety factor; the Zeppelins had a great safety record right up until the moment they didn't, and since the Hindenburg was destroyed in minutes, that isn't appealing on grounds of safety considerations.
If the problem of "how do we ensure the giant flammable disaster waiting to happen doesn't happen" could be solved, then Land Cruises on the luxury blimp could definitely be a market! As the "Tasting History" video explains, the low altitude the blimps travelled at meant that passengers could almost reach out and touch the sights, and the smooth motion meant no turbulence or similar problems.
Imagine being able to cruise along *inside* the Grand Canyon like a drone, for instance:
I was under the impression that the safety issue was solved with helium although the greater required size of helium airships may make them uneconomical I suppose.
I thought so too, but as you say, nobody seems to have picked up on the untapped market for airship cruises. I do think the spectre of the Hindenburg disaster hangs over the entire area, the same with nuclear power and the bad reputation it has due to several incidents at power plants.
But lo! Out of curiosity I looked it up online, and at least one brave pioneer is doing this very thing!
Prices are a bit eyewatering, but if you're sufficiently flush with cash, why not? And this is Swedish, so I'm sure American entrepreneurship could work out a way to cut down prices to the middle-class budget (less luxury but still something special). Say $5,000 per person but you get everything included (meals, drinks, etc.) for a seven-day land cruise. Sounds good to me! Something you'd save up for, or that the entire family would chip in to give Mom and Dad (or Grandma and Grandpa) that holiday trip of a lifetime.
"The demand for our North Pole Expedition has exceeded our best expectations and we are now only accepting requests for interest on a waiting list mode for our Adventurer departures. However, we still have limited availability in our Pioneer programme which also secures your place within the first departures while also giving you the opportunity to invest in us. Cabins for our North Pole Expedition start at 2.000.000 SEK (~$200,000 USD) and departure dates are subject to the manufacturers’ calendar.
What is included and excluded from the price?
Included: roundtrip transfer in a snowmobile between your hotel in Longyearbyen and the airship, 2 nights onboard the airship in your private cabin, services of a professional expedition leader and crew on board, meals, beverages (alcoholic and non-alcoholic), activities and guided excursions and related equipment.
Not included: flights to/from Longyearbyen, accommodation in Svalbard, airport arrival and departure taxes, passport and visa fees, travel insurance, medical evacuation insurance, personal items, optional activities outside the main North Pole Expedition program (pre and post programs, additional ground services in and around Svalbard), gratuities"
Wow you weren't kidding about it being eye watering! 5000 seems like a good target, more than an ocean cruise (while its still a novel technology) but not so much so that its out of reach for the middle class.
A brief stroll through the history of rigid airships, almost all of them not named "Hindenburg", will show that the safety issue is mostly that the things are too big and flimsy to survive major storms, and can't fly high enough to pass safely over them. I am not aware of this problem having been solved.
If you just want to use it for tourism, that might not be a dealbreaker.
The rarity of extractable helium is a pretty tricky thing to actually know because of the US Strategic Helium Reserve sell off vastly suppressing the price and making it uneconomical to extract and thus we just release huge amounts from things like natural gas wells and don't even attempt to exploit it. Its fairly similar to the articles about all of the worlds Rare Earth Metals being in China when the reality is there just wasn't really a business case to survey deposits until recently.
The use of helium in dirigibles would be an absolute drop in the ocean compared to modern uses of the gas though. Its like when people complain about aviations carbon emissions, despite seeming like it would be a massive issue the inherently small market makes it miniscule.
Once the bugs are worked out, and revenue from superior door-to-door air freight service has recouped development costs, just swap out the cargo bay full of intermodal shipping containers for some modular luxury apartments.
"revenue from superior door-to-door air freight service has recouped development costs"
They're talking about doing that in Canada - it does have a problem with "large areas of the country are inaccessible except by aircraft, everything has to be flown in, a head of lettuce thus costs $15 (apparently Canada had a lettuce shortage crisis due to having to import lettuce from California which was hit by drought and bad crops, resulting in shortages)".
"In the last few weeks, Dean has considered himself lucky if he gets a quarter of the cases he normally needs, and actually, that’s fine, because at these prices — $6.99 for a head of iceberg, $14.99 for a bag of three romaine hearts — nobody’s buying it anyway.
Article content
“The price points are insane,” said Dean, who runs Mike Dean Local Grocer Ltd., named after his dad. “I won’t put a margin on it because we’re already looking ridiculous.”
The reason for the shortage and the ridiculous prices, at least this time around, is Impatiens Necrotic Spot Virus (INSV), a bug-borne disease that has wiped out a massive chunk of the lettuce harvest this fall in California’s Salinas Valley. The region, known as the world’s salad bowl, grows more than half of all lettuce produced in the United States. This season, however, lettuce producers reported losing as much as a third of their crop, according to the Grower-Shipper Association of Central California."
If lighter-than-air craft can bring down transportation costs, they'd be very happy about that.
Transport is the main challenge in bringing goods to remote areas. Goods are either transported by trucks over winter roads to inland communities, by seasonal summer voyages to coastal settlements, or by air freight, which is the only year-round cargo service. All three of these options are costly, not always reliable and can be limiting based on weight and weather.
Due to climate change, these forms of transportation have been drastically affected. In the last decade, the availability and reliability of transport has diminished, and northern communities are suffering. The earth’s temperature continues to rise, which shortens the winter roads season, and creates unpredictable sea and weather conditions.
This is only going to get worse as the years go on. The remote communities are desperately needing another solution for the transportation of goods."
You might be on to something here JamesLeng! It could be well worth your while to look for investment opportunities!
(And if the airships don't work out, you can always fall back on lettuce farming for the Canadian market 😁)
I don’t quite buy the story of “perpetually losing for 50 years”. Why didn’t the highly optimistic, pro-growth, anti-government-regulation attitudes of the Ronald Reagan administration manage to put a dent in any of this during the 1980s? Reagan had a pretty astounding popular mandate and hated big government regulation.
Stuart Buck, above, made a compelling argument that the number of pages in the Federal Register may be an exceptionally poor proxy for "amount of regulation."
It looks like you’re right about this — Reagan didn’t do all that much deregulating, even though he talked about it, it was kind of a neglected part of “Reaganomics”.
I always feel like Russia will not be able to go solar - we're too far north, and have too many cloudy days where even the most efficient solar panels will produce far below their top capacity. On the other hand, Russia is pretty much the only large country in such position, so if everyone else goes solar, we may continue to burn fossil fuels and use nuclear power (as well as use hydro power, and maybe geothermal in some rare places) for a long time without much negative effect. Economically, of course, solar revolution in the rest of the world might be a disaster for Russia, but on the other hand, oil is needed for more things than energy production, and with cheap energy, the demand for it might not fall that much.
Another thing of note: I often hear about energy prices going negative in Germany, but I also read that metallurgical and chemical companies (BASF, primarily) are closing plants in Germany, in part, because of cost of electricity (after loss of cheap Russian gas). Can anyone explain to me how this two facts can co-exist?
BASF needs consistent energy prices. If energy prices are negative 10% of the time and 50% higher than they were in 2019 the other 90% (or some other distribution with higher mean and higher variance), they lose out.
They even lose out a little if the mean stays constant and the variance goes up since it makes their expenses less predictable (though I doubt that's the main effect here).
Basically power prices in Germany are negative *sometimes* (ie on a sunny or windy day) but they're high on average. This impacts the economics of running a high-capex factory like BASF's, which needs to be run more or less constantly.
However it's also bad for the economics of wind and (particularly) solar, because the negative prices happen when they generate.
Germany built a lot of renewable energy infrastructure when it cost much more, I think their electricity price is triple the fossil grid. A high renewable system also needs a lot of transmission and backup
Germany is locked in to high prices for some amount of time while the grid effect means that it would take radically cheap batteries to unlock much cheaper whole system costs
They are actually coupled. About 20 years ago, Germany committed to pay very high prices for electricity produced by solar panels. The deal was risk-free and high-gain for investors, so all companies in the world stopped whatever they were doing and built solar panels for Germany. instead. (Ok, I *am* exaggerating here, but only slightly. It completely spiraled out of control. I think the cost for German electricity consumers has been something like 200 billions.
The effect is twofold:
- The solar companies will be paid for the electricity from their 15-20 year old panels, regardless of market prices. Essentially, the German government pays them directly. This means they never have incentives to switch off production. This sometimes leads to negative market prices for an hour or so, because even then they have no incentive to switch off production.
- Afterwards, the German government collects the money from all consumers of electricity. So companies like BASF pay twofold: first, the market price. Second, the insane amounts of money that the German government pays for 15-20 year old solar panels. The second part is what makes electricity much more expensive in Germany than in other countries.
Fortunately, the subsidies were only guaranteed for 20 years and will run out in the next 5 years or so. It's not even a big share of all installed panels in Germany, so it was a *really* stupid policy 20 years back. (And it did not look like a good idea even back then.) But on the other hand, at that time Germany was pretty much unilaterally responsible for a world-wide explosion of production of solar panels, all of which were installed in Germany. Without that, mass production of solar energy and the rapid decline of costs may never had started. So good for the world, I guess.
I remember reading a much more comprehensive list of ideas, but unfortunately can't dig it up.
I also recall a convincing blog post on the limits of reducing the cost of solar energy. It argued we're near the end of the exponential because the cost of mass-producing modules will stop being the limiting factor (other costs like land, installation, power electronics, maintaining the grid, etc. will dominate).
Yeah, but this is ridiculous. Casey Handmer is a dreamer, not an engineer. Terraform Industries doesn't have a tech to produce liquid fuel directly from solar, otherwise he wouldn't need to write with so many unsourced charts.
The reason that the cost of solar power doesn't go to zero - or to $10/MWh - is indeed that the cost of solar panels is now a fairly small part of the cost of building a solar plant. In markets that don't have import tariffs on China (ie not the US or India), a solar module currently costs about 10 US cents per Watt. The total capex of even a large solar plant, however, is in the range of 40-60 US cents per Watt - because this has to include mounting systems, racking systems, the inverter, at least some grid connection costs even if you get lucky and get a perfect site right next to existing transmission, and then land and arrangement costs.
If racking/mounting is a big part of the cost, could you just wrap the modules in a more rugged coating at the factory and then have the install just be dropping them on the ground? Since the modules themselves are getting so cheap, slightly lower efficiency per-module doesn't seem very important.
There's a company called Erthos trying to do this. I'm not a fan.
The design lifetime of solar modules is over 25 years, so you do want them to be installed on a structure that will fully resist wind and allow air cooling on the backside, access to the junction boxes and cabling, rain cleaning, and vegetation management. Depending on the specifics of your environment, it's likely that just plunking on the ground will end up with them blown away, buried, or covered in plants within a few years unless you're constantly working to prevent that.
There are some designs of system for flat roofs where the modules basically come on plastic trays you can with rocks to keep them in place - but the roof then needs to support that.
I know the comparisons are going around comparing PV panels with roofing material (but comparing high volume bulk purchased PV with consumer retail roofing).
But given the mechanical needs of a PV panels (resist weather and wind and heat, while staying sealed), my mental model is we shouldn’t expect panel costs to drop below wholesale bulk roofing material costs. Even if the cells were free.
And they need to be mounted about as well as those same roofing materials or they will blow away…
The elephant in the room is China. The current emphasis on progress seems like a precursor to the 1990s competitive mode, where Silicon Valley chip companies were "paranoid"¹ about Japan taking over, thus ensuring that the Internet boom happened in Silicon Valley and not in New England or somewhere in Europe or even in Asia. Progress studies is a mislabeling of what is ultimately competitive studies.
You can't defeat the God of Straight Lines, but every half-century, you'll find that a declining Empire will have an efficiency decade with some Thatcherian soul-searching and belt-tightening. For a time, some important numbers will reverse. Never in a million years would I have believed the U.S. would return to being the top oil producer and reduce its dependence on OPEC, for example.
YIMBYs are winning? As far as I can tell, YIMBYs have turned California into Austin (or is it the other way around?) by expanding low-density housing, thus making it more attractive to yuppies, which inevitably raises housing prices, therefore bringing more homelessness.
[1]: "Only the Paranoid Survive" was the famous mantra of the former CEO of Intel, Andy Grove
"Self-driving taxis have a big advantage over self-driving self-owned-cars: they can operate 24-7 and never have to park."
Really? So where you live, demand for taxis is the same for each time of day? No rush hours in the morning and evening? Or do you plan on having the idle taxis driving around aimlessly, clogging the streets?
I think the point is that they don't have to park right next to the places people want to go, so you can put in some concentrated parking facilities and not need to have mandatory parking next to your house *and* next to your job *and* next to the shopping mall. This is similar to how buses or school buses (or taxis!) don't really contribute the problem of finding a parking spot during rush hour.
There may be some advantage to that, but you need to factor in the additional traffic by the taxis coming and going to their centralized parking (im addition to driving around empty much of the time to pick up the next passenger). Also, they're essentially still cars, so they take up just as much real estate per passenger as cars, while buses take up much less space.
Edited to add: I don't doubt that robotaxis would reduce the demand on parking space compared to individually owned vehicles to some extent, because fewer of them would be needed - how strong that effect is depends on how strongly pronounced rush hours are. I don't think a massive rollout of robotaxis would reduce overall traffic - quite the opposite.
1) There is NO world where driving automation doesn't make congested streets MORE congested. Some hype artists like to claim otherwise but it's just not credible: so long as we don't have Star-Trek-style transporters, physics and geometry mean that roads are finite and more activity will congest them. And automated driving makes using roads easier at the margin.
2) We should be excited about the application of driving automation to traditional transit. If every bus was automated, AND it had its own right-of-way, you could run buses at subway frequencies if you wanted, and affordably too (modulo your city's definition of 'affordable').
If the fleet of robotaxis involves a dispatch server keeping track of where they all are at all times, and all the other cars which any of them can see, using that data to plan routes and calculate appropriate passenger fares, it would be fairly trivial to implement precisely calibrated congestion pricing.
1.) Sure there is. The world in which a large part of traffic problems are variances in human driving habits.
In Dallas, back in the ... 80s? 90s? It has been a while. Anyways, in Dallas at some point in relatively recent history, highway traffic had a pattern that was baffling analysts: Every, let's say it is Wednesday, at 2:00 PM, there was a massive backup on a key highway. (The exact details escape my memory and aren't important.)
They studied the problem for a few months, sent up helicopters to watch the traffic. And, after extensive expensive analysis of the problem, they figured out what was going on: An older driver had a weekly hair appointment, got on the highway going SIGNIFICANTLY slower than all the other drivers, and created a traffic problem.
To solve the problem, they sent a driver every week to take the driver to the hair appointment. And their mysterious traffic problem vanished.
This is not a small thing; traffic is full of variations on this.
Traffic is also full of coordination problems that automation can provide a path to resolving. A car does something dumb, everybody brakes, and now you have a wave pattern embedded in the moving traffic which is wasting fuel by making people brake for a hazard that hasn't been there in half an hour.
Huge amounts of traffic management is basically making sure that the traffic is sitting in a place that doesn't cause downstream problems - preventing gridlock. If the cars themselves can be coordinated to avoid gridlock, this problem goes away.
A couple of jobs back, there was a massive traffic snarl in my daily commute caused by an on-ramp that went for a mile before turning into an off-ramp; if you got there early, people used the exit lane to pass slower-moving traffic. After ten minutes or so, this turns into "people use the exit lane to pass the traffic brought to a stop by people using the exit lane to pass the traffic".
Nevermind basic driver inattention - a driver who misses their turn and has to do two u-turns is an additional load to the system.
There are many worlds where automation can easily make congested streets less congested, and we live in one.
1) You seem to be describing a world where ALL cars are automated, so there are no human drivers to occasionally pump the brakes when they shouldn’t. Sure, in THAT world, there’s potential for improvements, but we don’t live in it now… and I’m not certain that we ever will. If we do ever get there, no one driving today will live to see it, so it’s academic.
2) You also seem to be describing a world where making driving easier through automation changes nothing else. But of course that’s not so: make it easier on the margins, and people will do more of it. It’s supply and demand… or more precisely induced demand: https://en.wikipedia.org/wiki/Induced_demand
I’m no fan of congestion either, but it’s a necessary by-product of automotive trips. We can mitigate it, but never eliminate it, nor even signficantly reduce it (at least not in prosperous areas).
1.) I was responding to a specific person making specific claims. If you want to argue about -entirely different- claims, we can have that conversation. But even if you want to talk about -just these- claims, note that nothing I said requires every driver be automated, as limited coordination is still superior to no coordination.
2.) Question: Are there currently, right now, empty roads with no drivers? Why? Shouldn't they be filled by induced demand?
Obviously not; demand is not, in fact, perfectly elastic, consuming every inch of available road. Indeed, I spent ten years in a city which had lost a substantial part of its population base, which, combined with infrastructure built for a much larger population, made driving incredibly pleasant there. (Excepting areas of new construction, which were built with the newer, lower population in mind - which resulted in traffic concentrating in newly-constructed areas.)
Induced demand applies when there is a greater demand for transportation than there is available supply. Sometimes you can't reduce congestion, because every improvement results in more drivers - but this is not, in fact, a universal phenomenon, and even where it does arise, the demand is not, in fact, infinite.
More than this, however, even if an improvement merely results in more people being transported - this is still an improvement! It's a huge improvement! People are getting places they wanted to be, which they previously could not get to!
You can't compare buses to cars because buses aren't a realistic substitute for cars.
Very few people choose to take buses if they have the option of simply driving there, because buses inevitably go from somewhere that isn't exactly where you are to somewhere that isn't exactly where you want to be at a time that isn't exactly when you want to go, in the company of people you wouldn't choose to hang out with.
A self-driving taxi *is* a reasonable substitute for your own car so it's a fair comparison. There would be some need to construct additional infrastructure to cope with the huge influx of self-driving cars into the city centre in peak hour, but it's a lot cheaper than building trains.
Elena is right: dedicated facilities for robotaxi maintenance, storage, and charging are a plausible outcome (they won't ACTUALLY go 24/7 though the business incentive will certainly push as close to that as possible).
Street parking is also possible, since the robotaxis will obey whatever the local parking ordinances are... but street parking is antithetical to what makes cities great, so this would be something to discourage.
At some point—which won't be in my lifetime, but a man can dream—the proportion of automated vehicles as a percentage of total road users will be so high that robotaxis might simply park on freeway lanes between, say, 0200h and 0500h; just like street parking, but not in a place where they discommode anyone.
In the distant future, when automated vehicles are
As other people say, you can always have them drive to a parking lot way outside of town. But there could also be more creative solutions - maybe you can swap out the passenger chassis for a freight chassis + seat for one deliveryperson, and Amazon could use them to deliver packages overnight.
Actual deliveryperson seems inefficient. Could probably be replaced with a robot arm for shifting cargo within or immediately adjacent to the vehicle, plus recharging cradles for a few quadcopters capable of carrying, say, a 30lb box thirty yards, then ringing a doorbell.
We can do better. Once you've got the robot arm, just have it throw the packages in the general vicinity of the destination. You often wouldn't even have to slow down!
For some applications that'd make sense, but a lot of parcels are going to be fragile, or have significantly nonuniform density, high drag relative to their mass, or other inconvenient ballistic properties, and a lot of destinations will have awnings, pedestrians, or other obstacles. Also, an arm rated for throwing (alongside other cargo manipulation, so it can't be a simple trebuchet) would be a significant engineering challenge, whereas picking things up and setting them down more carefully can be adapted incrementally from warehouse bots.
Acting as a mobile base for quadcopters already covers most of the "without slowing down" benefit. Truck might dispatch a flock of short-range drones into a residential subdivision while cruising past on the main road, each drone with a package for a separate house, and drive on to complete other deliveries before returning to pick up the drones minutes, hours, or days later, or allowing them to find their own way to some other collection point - made easier by the fact that they only had to cover a short distance with the heavy payload.
Breaking the things you're being paid to deliver, and the windows and bones of people you're delivering them to, is not a viable business model even in absurdly favorable regulatory contexts.
I think the tfp plateau from 1970 onwards is just homeownership in the classic yimby/Georgist sense that promoting speculation of limited assets leads to innovation slowdown. But environmental and labour regulation also had some effects. I think we could get the productivity back with sensible regulation if we had a lvt/yimby/densification boom
TFP measures GDP growth vs. inputs and is inflation adjusted. I think this calls into question how accurately we define inflation. As an example think about software like video games. Their price has increased hugely over the last but their complexity has increased orders of magnitude. One could argue that inflation should be considered negative for many types of software. What about the reliability and efficiency of automobiles? The decision of what to consider inflation and what to consider as actual growth seems critical to the measurement of TFP.
This objection was discussed at the conference, but I think answered by a graph showing that manufacturing productivity has been stagnant since 2005; that seems harder to fudge.
I am not sure what this means though. I would not be too surprised if the manufacturing productivity of product X had stagnated. If the product is profitable and the market isn't growing then there may not be tremendous incentive to improve productivity. On the other hand, if the measurement is on a continuously changing mix of products then it seems that there is ample opportunity for errors. In my own, semiconductor field, production is now transitioning to 12" wafers (4" was the norm when I started in 1991). The transition from 8" to 12" alone results in a 30% increase in output (die) for the same cost (productivity). I just feel that there is plenty of room to be skeptical about these types of measurements, both in there accuracy and in their applicability.
It is a 125% area increase but 30% cost savings. Masking time is unchanged. Silicon, dopant, etchant and photo resist usage and process times are all unchanged. The exact cause of the cost savings isn't entirely clear. I guess, the building itself, as well as, implantation and diffusion steps yield 125% more die per (each). However, quite a lot of equipment must be updated to 12".
> As an example think about software like video games. Their price has increased hugely over the last but their complexity has increased orders of magnitude.
Has their price increased? My sense is that the nominal price of games has stayed pretty flat over my lifetime, which means the real price is dramatically lower after accounting for inflation despite quality and complexity both going up.
I suspect that a lot of the quality of life impacts of tech/software/the internet/etc are dramatically underestimated by these kinds of statistics since they're generally cheap or free.
I am guessing that my gaming career started prior to yours. I think high end games were $29, back in the 80's despite coming on 12 floppy disks. You are absolutely correct though that this means decreased price in inflation adjusted terms.
Regarding the “Public Transit” section: This is what everybody who tells you to “get tough” on homelessness means. Getting rid of the people who suck makes things suck significantly less. It is very easy to set up a simple test that eliminates the people who suck. You can just charge people money. Homeless drug addicts who can’t hold down a job and constantly commit crimes don’t have any money.
In my dream, cities replace all existing taxes with a flat per-day fee for existing within city limits. You could call it a "congestion fee" except it would apply to people rather than cars.
San Francisco has a $15.9 billion budget, and roughly 1 million people in the city on any given day, so the fee would need to be $43 a day. Residents wouldn't mind, because it would replace all their existing taxes. Tourists and business travellers wouldn't mind, since it's a small fraction of the cost of a hotel room. People who commute from outside city limits *would* mind, but you could probably have some sort of reciprocal arrangement for people who live in neighbouring cities and pay their fees there. The only people who would mind would be the people who currently live in the city without paying taxes.
Just coming here to be an annoying crank and remind everyone that 1971 is the year the US fully ditched the gold standard and started our modern freewheeling monetary policy experiments.
>The United States could produce all its power by covering 2% of its land with solar panels - for comparison, we use 20% of our land for agriculture, so this would look like Nevada specializing in farming electricity somewhat less intensely than Iowa specializes in farming corn.
I think this actually understates the case. The US uses ~3TW of primary energy. Fully electrifying everything would probably increase that somewhat, maybe to 5TW or even 10TW depend on whether we manage to be smart about it. 2% of US land area covered in solar panels would be 200,000 km2. At current panel efficiencies and typical capacity factors, that would give us 50TW, 10x what we need. Beyond that, there is growing (but much slower than I'd like) awareness of agrivoltaics enabling dual land use. This seems like it should be a tradeoff, but in many case the result is higher yield and lower water consumption for crops, plus electricity production to increase farmers' profits. For some crops, research is showing the extra shade improves nutritional quality. This gets even better if we manage to get the kind of semitransparent super cheap perovskite cells as standalone modules instead of how people are starting to use them now, which is as the top layer in tandem modules, because then they can go directly above crops instead of between rows without impact availability of the frequencies plants can actually use.
And I see someone mentioned the value of desalination as a use for cheap energy. I'd add wastewater recycling and indoor farming as other clear use cases. "We're losing cropland to climate change" becomes a less severe threat if you can say "We can drop a ten story building anywhere on the planet and power the HVAC and grow lights and water recycling and water supply with nearly-free electricity, with a year-round growing season and 2-10x the yield per floor area of a typical outdoor farm."
I agree with your point, but its even stronger than you think: electrifying everything would *decrease* primary energy. Primary energy includes fuel. Electric vehicles are much more efficient than combustion ones, and heat pumps are 5x more efficient in primary energy use than nat gas furnaces. We can likely *decrease* primary energy use by electrifying, not that we necessarily should.
Oh, for sure. Primary energy consumption will decrease for a whole lot of things when we electrify. In other cases it'll be nearly one to one. In the remainder it might be a rather significant increase, especially in some chemicals, fuels, and industrial processes, at least initially.
I suspect in the short to medium term the latter will balance out or outweigh the former, and in the long run we'll Jevons Paradox ourselves into much higher total energy consumption by the time we fully decarbonize. For example, most aviation (which is kind of the worst case scenario application) is still going to be kerosene based for at least the next 20-30 years, there's not enough biomass sustainably available to make SAF so we'll need a lot of electrofuels, and that alone could easily eat up the equivalent of 60% of current total world electricity consumption to eliminate 2.5% of current world CO2 emissions.
>Self-driving taxis have a big advantage over self-driving self-owned-cars: they can operate 24-7 and never have to park.
But demand varies over the day, right? Like, either you have the number of cars needed to meet demand at 3am and almost no one can get one at peak hours, or they are all wasting fuel and creating traffic to drive in circles all night, or they park somewhere?
I would guess the answers to that are either 'they can go park outside the city where space isn't at a premium' or 'they can be converted to other use at night such as package delivery'. Those seem plausible to me intuitively, but I wish I knew what the experts actually-thought-about answers were to this question.
As the expert Scott spoke to, I can say that both answers you offer are plausible but the former is more likely. Regarding the latter: even today people have experimented, but it turns out a vehicle optimized for moving passengers is NOT optimized for moving goods, and to try to do both is to do neither well.
Regarding the former: 'where space isn't at a premium' might be IN the city; design a parking garage for robots to navigate rather than humans and you can be MUCH more efficient. And such garages will be needed in any case for charging (they'll all be electric, of course) and for maintenance: human nature being what it is, robotaxis will likely need to be cleaned and groomed a few times a day.
Perfect, thank you, that is exactly the level of analysis I was looking for :). Much more efficient parking that also does charging/maintenance makes a lot of sense.
What happened in 1971? Looking at the graphs, the key date could actually be a couple of years later, in 1973. That's when the price of oil suddenly tripled and stayed high: https://ourworldindata.org/grapher/crude-oil-prices. I guess that had a major, enduring effect on the economy.
Yes, that graph certainly looks like the inflation-adjusted price of oil fell back to nearly the pre-1973 level. The graph I'd used for inflation-adjusted oil prices looks meaningfully different, though. It's from the BP Statistical Review of Energy, so I hope it's reliable, but I hadn't linked to it above because it requires downloading the pdf, but here it is: https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/energy-economics/statistical-review/bp-stats-review-2022-full-report.pdf. It's on page 26. The inflation-adjusted price stayed over 3x the pre-1973 price for around a decade, then fell to about double the pre-1973 price in the mid 1990s, which interestingly is when productivity started to rise a bit faster.
"Once you convert a problem to 'let’s manufacture billions of identical copies of this small object in a factory', our natural talent at doing this kicks in, factories compete with each other on cost, and you get a Moore’s Law like growth curve."
More canonically this is called Wright's Law, or "learning curves", whereby each doubling of the total produced units of something tends to produce a similar ratio reduction in the cost. Solar panels go down about 20% in price every time we double the total number of produced units:
Solar panels are currently producing about 4% of US electricity, which can give a ballpark estimate for how much cheaper they will get. Assuming constant electricity demand, we'd be saturated with cheap electricity in five doublings (8, 16, 32, 64, 128) at which point solar panels would cost (0.8^5) or about a third of what they cost now, which is still really great in terms of electricity cost but not 1/50th of what it is now as alluded to in the post.
(If we did dramatically increase our energy consumption as the costs drop - which wouldn't be very surprising - we might squeeze out a couple more doublings of total capacity and therefore get down to, like, a sixth of the current cost. Much further than that in a short period of time would surprise me.)
Would this mean that solar would never get past 1/3 current cost, or just that we would have to double amount produced by waiting for time to pass rather than by scaling up?
Also, is this a totally different thing from Moore's Law, or does that also work by doubling the number of chips produced?
The best mathematical formulation for the solar module learning curve is that every time cumulative production of solar modules by our species doubles, the cost comes down a certain percentage. Sources usually put this between 20% and 30%; my calculations put it at the upper range of this (and currently some really weird stuff is going on with solar module prices don't ask).
This is a way to express that solar modules will never stop getting cheaper, but it's harder to squeeze out costs from a $0.10 per Watt module than from a $5 per Watt module, so the rate slows down over additional experience.
Moore's Law is a specific case of the experience curve which assumes that the rate of cumulative production doubling is really rapid, to give a time relationship. It's even more empirical than a straightforward experience curve.
In any case the big issue with reducing solar total costs is now the non-module components, which are mostly not simple manufactured things.
I think you are conflating number of installed panels with yearly production. Your "five doublings" would apply only if all of the 4% had been installed at the current "productivity."
> this would look like Nevada specializing in farming electricity somewhat less intensely than Iowa specializes in farming corn.
This is probably not actually what it would look like; there are losses in transmission, so it's better to have generation within a few hundred miles of the consumers. Of course this is offset by northern states having less sun, and some places having more cloud cover, but the difference between solar irradiance in the southwest vs. the north is only about a factor of 1.5, so it doesn't take much in transmission losses (not to mention the cost of building and maintaining the long-distance infrastructure) before it's better to make solar farms local.
> less likely to end with major cities being converted to radioactive ash
Nuclear reactors can't explode like a nuclear bomb. A bad meltdown looks like everyone in the city getting radiation poisoning, not a fireball or any buildings getting destroyed.
I'll edit it for clarity, but another problem with nuclear reactors is that terrorists can steal the nuclear material (although this may be more true in the minds of regulators than reality)
That's true. Also potentially more concerningly, rogue countries can use a civilian nuclear program as source material and plausible deniability towards building a bomb.
Its why we have to ensure renewables are as cheap as possible, the developed world is always going to be antsy about exporting dual use technologies (which nuclear infrastructure is) to unstable countries.
Even more remarkably, a bad meltdown looks like three gigawatt scale meltdowns at Fukushima, where it is conservatively estimated that 50 people would have died from cancers with no evacuation and 2000 did due to the evacuation
It is now safer in health terms to live in the exclusion zone than live in the relatively clean Tokyo city air!
For similar reasons the Chernobyl evacuation zone ceased to make sense after just a few years as the most dangerous isotopes all have a short half-life, by definition.
There are some isotopes which don't put out much energy, thus are relatively persistent, but end up concentrated by biological processes into vulnerable organs or tissues and thereby cause disproportionate harm. Strontium-90 substituting for calcium, and Iodine-210 in the thyroid, I think are the most infamous ones? Even those only last so long, though, and the fewer types are left, easier it gets to filter.
>Cesium is uniformly distributed throughout the whole body similar to potassium, and it does not accumulate in any one particular part of the body like iodine
For what it's worth, people *are* currently working on battery-powered airplanes. Most likely, trans-oceanic flights will never be electric, but city-to-city shuttles easily could be. The gravimetric energy density of the most cutting-edge cells (silicon or lithium metal anode) is good enough that it's feasible, at least theoretically for small planes.
That depends what you mean by city-to-city. If you're talking 300+ mile trips, it is very very unlikely to be possible to get any service or safety level approaching what we have today. You'd need 8x battery energy density from what we have currently just to get a very limited capability aircraft which would never sell, and by the time we have that kind of energy density in batteries we will likely have hydrogen or sustainable aviation fuel alternatives.
There are some projects out there but nothing that isn't basically just fraud, the most egregious example being the Eviation Alice.
The other main advantage of hydrocarbon fuels in aircraft is that the aircrafts weight decreases considerably as the flight progresses. This is most commonly seen when aircraft that have to return to their departure airport after a moderate mechanical issue have to stay in a holding pattern to reduce their landing weight. This makes air travel cheaper as you can load an aircraft far beyond its maximum landing weight with the expectation that the weight will be lower when it reaches its destination. This is a pretty big barrier for battery powered jetliners.
Just to add: one way to assess whether a project is viable is by looking at how much the airframe without batteries weighs, and how much weight they are planning on taking off and landing with. So, an airframe like the ERJ175 weighs something like 50,000 lbs, and max takeoff weight is 85,000, and max landing weight is 75,000 lbs. So structurally they need a 50,000 lbs airframe to sustain landing at 75,000 lbs, for a 1.5 to 1 max landing weight to airframe weight. If they could land with more weight, they would either raise the max landing weight (to make more money) or lighten the structure (to make more money).
Most battery-electric airplane companies won't publish their MTOW/MLW or battery weight numbers. Eviation did have theirs published at one point, with a ratio of empty weight to MLW of 3.67:1, compared to the 1.5 to 1 for the ERJ above that is just obviously not possible.
That would add a lot of weight on takeoff with parachutes etc, and the battery packs would have to be very strong to withstand the impact, you’d need frequent inspections, and you’d need to retrieve the batteries after every flight which would be an enormous cost.
A better option would be an energy beam to power the aircraft from the ground so that your battery is just for reserve or cruise.
I have about 90% confidence level that Li-metal anode secondary cells will never be made safe enough for commercial use, much less aviation. The dendritic growth problem is an extremely tough nut to crack.
So I've spent my professional life studying solar, and I'm essentially a solar enthusiast, but it does sound like the ones at this event were not being intellectually honest. Solar cannot cost $1/MWh, even if the panels were free and of the maximum theoretical efficiency, because you need to install them properly, wire them up properly, and maintain them; this isn't a huge cost but it does become meaningful when the cost you are discussing is under about $100/MWh (currently the price of solar modules is only about 20% of total capex, depending hugely on size of your solar project etc). Solar is cheap, but it isn't that cheap. Even reported prices close to $10/MWh are from the Middle East and are pretty much fudged to make the bidders look good.
However there are plenty of places where if you could pay a solar plant a guaranteed $35/MWh, someone would build you a solar plant. Then the reason why you don't immediately build something to use it - like a desalination plant or electrolyzer - becomes the capacity factor. These are expensive kit, and solar capacity factors are at most 35% (I mean, even in theory you could only run them in the daytime, ie 50%. Yes you can add a battery but that's now even more kit!). This is at least theoretically possible but the economics aren't the slam-dunk you might think, because of capex.
What mostly we (ie the real solar industry) is banking on is making demand more flexible. Batteries, but also charging electric vehicles in the day, deep-cooling industrial fridges, maybe shifting datacenter loads to parts of the planet that are currently enjoying renewables.
Terraform Industries is not a serious company; if it really had the tech it claimed (making liquid fuel from sunshine), it would be an instant win, but it can't. Synthesizing hydrocarbons from air and water vapour is just... not a thing at the moment.
I'm pretty sure you don't want details so: a residential rooftop solar plant costs roughly three times what a utility-scale solar plant costs, in most countries. This is mainly because of sales and marketing costs to individual customers and also because you normally need scaffolding, people need to travel out to do the work individually, the inverter needs to be smaller, etc.
A commercial solar plant (ie like 1MW on a Wal-Mart) is between the two on cost. Some economies of scale but not *as* cheap. Except in China because the state organises bulk purchasing schemes.
Regarding demand, one generally overlooked sink that many people have is hot water tanks! At least in the UK most hot water tanks can be heated either from a natural gas boiler or an electric immersion heater, and the water stays hot for a long time. You can't convert it back to electricity, but it's a useful overflow mechanism that most people already have and is easy to set up.
There are many companies around the world who are serious about making green hydrogen, ie from running electrolyzers on renewable energy. (The biggest volume electrolyzers are, of course, in China). It's do-able but expensive and honestly there's just not that much demand for green hydrogen, but they're ramping up major capacity.
There's also interest in making this hydrogen into ammonia to make it easier to store and transport.
There is very little interest from serious players in making hydrocarbons. One issue is that there really isn't much carbon dioxide in air, and if you start with concentrated CO2, you haven't started "from air". It's actually relatively hard (energy-intensive, capex-intensive) to get high-concentration CO2 though it's definitely not impossible (that's why CCS has become CCUS, ie carbon capture, utilization and storage). But most discussions of Sustainable Aircraft Fuel, SAF, are talking about biofuels which are something different.
This is why I am dubious that a US startup which publishes very dodgy solar numbers has simultaneously cracked low-capex electrolyzers AND making hydrocarbons from air plus hydrogen.
Stick a couple of wires in a cup of salt water. You just made a low capex electrolyzer. And making hydrocarbons from carbon dioxide plus hydrogen is a trivial chemical process.
None of that means that this approach will be competitive! But to act as if it's some impossibility is rather silly.
I would not discount the existing criticisms of solar, namely that they take up space. In theory we can cover our roofs in solar panels. In practice it means converting the lovely farm or pasture outside of town into industrial infrastructure -- it's easier for a single company to just buy an old cornfield and cover it with 10,000 roofs worth of solar panels than to convince 10,000 individual homeowners to individually get permission for, purchase, install, and maintain rooftop solar. There is a *really strong* negative sentiment in rural places towards solar power now, specifically because developers are buying up family farms and paving them with concrete and silicon. There's the feeling that a farm benefits the local community, while the solar plants are just using the land to power an Amazon datacenter so tech workers in cities can get rich. People value land and its usage at more than just the pure economic value of it.
Maybe I'm biased because I'm currently reading *The Origins of Woke*, and it's front and center in my mind [1], but it seems obvious to me that Civil Rights law could have contributed to the Great Stagnation. When hiring based on merit became heavily restricted/turned into a dystopian nightmare in the latter half of the 20th century, it would make sense that our collective competence/growth/progress would suffer as a result. It obviously wasn't the only cause, but I'm not sure why Scott didn't mention it at all.
[1] My *Two Arms and a Head* book review horrified Richard Hanania, so I figured that reading his work was the least I could do to make it up to him.
Another possibility, related to civil rights law and crime, is a decline in agglomeration. Before 1970, the NYC metro area's population had always increased, but it went from increasing by 11.2% from 1960-1970 to decreasing by 4.1% from 1970 to 1980.
Even per Hanania, it seems unlikely that that would directly contribute to much stagnation. Sure it's a constraint, but Hanannia notes that markets are incentivized to optimize within that constraint. That could, take the form, for example, of relegating diversity hires to lower skilled roles, while keeping higher skilled roles more meritocratic.
Indeed, as Cremiuex discusses in this thread (https://x.com/cremieuxrecueil/status/1848828338456113528), the dollar value of an additional point of IQ has increased in later American cohorts, as high compensation jobs and low compensation jobs have become increasingly sorted between high IQ and low IQ individuals.
That's still extra friction, transaction costs. Time and mental energy and communication bandwidth the hiring manager spends on tactical maneuvers related to the prospect of being sued is, one way or another, coming at the expense of their actual value-adding job. Farmers can try to "optimize within the constraint" of siege warfare, but it'll hurt the yield.
The Great Stagnation is a West-wide phenomenon, though. Sure, other countries have laws resembling the CRA, but the issue is still generally not as front and center as in American politics.
>Over-regulation was the enemy at many presentations, but this wasn’t a libertarian conference. Everyone agreed that safety, quality, the environment, etc, were important and should be regulated for
Obviously everyone's entitled to their own opinion, but given that the post itself acknowledges that overregulation of nuclear cost literally millions of lives, which is orders of magnitudes more than the sorts of archetypal environmental disasters that such regulations are trying to prevent, and given that the post suggests that regulation could be (part of) the reason the US isn't twice as rich, which would equate to a cost of tens of trillions of dollars, maybe it's worth rethinking the paradigm of "just have the right amount of regulation," which has cost trillions of dollars and millions of lives, and give libertarianism a second thought.
"I asked some of the YIMBY leaders there what they were doing - did they have blackmail material on Governor Newsom? They said politicians had finally realized that housing prices were in crisis, started groping around for solutions, and they’d been there at the right time and worked really hard to get the message out."
This reminded me of Milton Friedman, who would say that politicians/regulators take whatever ideas are in vogue. That's why you need to keep your ideas in the public discourse all the time. You never know when the current ideas will become obsolete and people will start searching for new ones.
I'm attempting to coin a term for the progress studies lexicon: Q̶u̶a̶t̶t̶r̶a̶ Qattara depression techonologies.
These are techs (like solar and EVs were, or what airships could be) that require subsidies to break out of the local optima and reach a new, better local optmia. I think our system is sometimes too much of a greedy algorithm and gets stuck in local optmia too easily. I'd love to get progress studies folks thoughts on this.
Coining a term is rarely accomplished through brute force. The term itself is clunky, as you noted, and uh, misspelled in this post. Less link spam please!
I think the chart showing California's energy use on a typical spring day is interesting. I am surprised to see that batteries are a significant factor, even for a relatively short period of the day.
However, it is important to note that residential electricity rates in California are the second highest in the nation (Hawaii has the highest), and are triple those in Louisiana, which has the lowest rates. California also has a lot of desert, which is ideal for solar. Further, I wonder what the chart would look like on the day of highest demand.
"The objections failed because Progress Studies is the same type of field as Gender Studies"
Ouch! Damning with faint praise!
Though by the later descriptions, the get-together seems to have been at least fun and at worst, no worse than these kind of talking-shops generally are.
Love the idea of Progress Studies and the summary of this conference! It's a great antidote to my worries about the election and gives me at least some reason to be optimistic, even if Trump were somehow to return.
But not sure how to take the couple mentions of "Ralph Nader, Jane Jacobs, Rachel Carson, hippies, protectionists, and all those people", which seemed to be negative. Jane Jacobs is invaluable. Rachel Carson was important. Nader, in his early days, did some good and important things as well.
Scott isn't implying that embracing Progress Studies means rejecting their legacies is he? Wouldn't the Progress Studies we want incorporate their insights rather than reject them?
Someone/something can have had a positive initial impact, but lead to an overcorrection in the other direction in the long run. The answer probably depends on which of those things you're referring to as their legacy.
That's certainly fair. I think, for example, Ross Barkan's recent Times article (https://www.nytimes.com/2024/09/30/opinion/the-power-broker-robert-moses.html) attempting to rehabilitate some of Robert Moses' legacy is fair, although I think its framing as a flaw in The Power Broker was wrong (and I wouldn't accept it as flaw in Jacobs reasoning either.
The Power Broker definitely acknowledges the positives of Moses' legacy as well as its flaws, and I still believe the core of Jacobs' argument about how cities works is spot on as well. Just as I also think that Carson's cry for us to be conscious of the environmental consequences of our actions and Nader's argument that we shouldn't simply trust companies or the market to ensure product safety shouldn't be discarded either.
Having said that, I fully accept the argument that groups with an interest in stopping progress are abusing mechanisms like environmental reviews and community reviews to simply stop development. If Scott is just saying we need the pendulum to swing back to prevent these abuses (versus getting rid of these mechanisms altogether), I'm fully on board with that.
This is wrong in so many ways. (but it does make me wish I'd gone to this conference)
>Is it theoretically possible that - since there aren’t existing fusion-related regulations - we’d have a fresh chance to fight the same battle and maybe win this time?
This battle has already been fought over the last couple years. The NRC voted to regulate fusion power plants like any other facility that uses radioactive materials (e.g. hospitals), rather than apply all the regulations for fission reactors. Then Congress passed a law saying the same thing.
The real problem with fusion is not regulation, it's that it still needs a good 10-15 years of tech development (which might or might not work out well) before we can even start the process of scaling it up to a significant chunk of the grid. Also we don't know how much it will cost.
>Between 2010 and 2019, the nuclear/solar cost comparison fell from $96 / $378 to $155 vs. $68 (yes, nuclear got more expensive).
Nuclear did not increase in price by 50% over 10 years. Very few nuclear plants were being completed during this time, so the data is noisy. They're not a fungible commodity like solar panels; this is like saying the cost of space stations has gone up because the ISS cost more than Skylab.
>Solar used to be limited by timing: however efficient it may be while the sun shines, it doesn’t help at night. Over the past few years, this limitation has disappeared: batteries are getting cheaper almost as quickly as solar itself.
This limitation definitely has not disappeared. Battery storage is cheaper than it used to be, but still about $300-400/kWh and only expected to fall about another 50% by 2050 [https://www.nrel.gov/docs/fy23osti/85332.pdf, see figure ES-2]. My guess is the OWID figure doesn't include packaging, installation, facilities, finance costs, or the power and control equipment to connect to the grid.
At $200/kWh and electricity revenue of 5c/kWh, even if you assume the input energy is free, you would have to cycle the battery about 4000 times to pay for itself. That means you're cycling it just about every day; so storing daytime energy to use in the evening is great, but anything longer term is out.
>If these trends continue, solar power could reach $10/megawatt-hour in the next few years, and maybe even $1/megawatt hour a few years after that. This would make it 10-100x cheaper than coal, and end almost all of our energy-related problems.
$1/MWh is laughable. Transmission costs alone would be far more than that. Even at $10/MWh, further decreases in generation cost aren't worth much because other costs become dominant.
>In fact, nuclear has been increasing ever since the ‘60s, when it cost about $20 (in current dollars). This is proof of concept that nuclear can be much cheaper than it is now.
Eh, kind of. Those old prices are pretty fake because the manufacturers sold plants below cost as loss leaders to increase their market share.
>If we did everything right - got really good regulations, innovated hard, switched to the most promising type of nuclear reactor - we could not just get the cost back down to $20, but go even lower - as low as $1 per MWh within a decade or two.
No, again $1/MWh is not real and it will not be real in the foreseeable future.
Costs for systems on sale today are already below the most optimistic projections for 2050 from that NREL analysis. How confident are we that storage isn't going to get a lot cheaper than the NREL prediction?
But to answer your question, it could get a lot cheaper than the NREL projection without getting cheap enough to really change the use case. Even a 10x decrease will need hundreds of cycles in its lifetime to justify itself, so it won't (can't) be adequate for unusually long periods of high demand or low supply.
We need the TIC cost to get to ~$50/kWh for batteries to be actually cheaper than NG at North American prices for diurnal arbitrage (assuming an average of 0.5 cycles per day).
For true seasonal storage (1/yr) to make sense, we are needing more like $1/kWh…
In my following of fusion news I find that companies are definitely sensitive to regulation. It seems to be a key parameter to limit the on site inventory of tritium, due to the potential emergency planning zone and risk of regulation. I believe the NRC ruling is that pilot plants will be regulated under particle accelerator and medical application rules while fleet power plants are likely to have some kind of fusion specific regulation
Helion has mentioned that one reason they might have two different kinds of machines is that you could have the energy producing D-He3 machines near populations with virtually nil tritium inventory. The D-D machines producing tritium and He3 might end up on an island somewhere. At scale that might not pose much cost for them but I feel like it's a real issue for the D-T reactors
It does look really promising for regulation, for example kind of a slam dunk in Britain who are being aggressive with fusion and have formalized a favourable regulation setup already. The US will likely follow so it's probably not a high risk. With tritium's mobility in the environment it does seem plausible to me that ALARA type regulations could impose a significant cost (despite its extreme harmlessness in actual fact)
>I believe the NRC ruling is that pilot plants will be regulated under particle accelerator and medical application rules while fleet power plants are likely to have some kind of fusion specific regulation
The NRC did consider a hybrid approach that would use the byproduct materials rules for "near term" fusion reactors, and left open the possibility that future fusion designs would be more dangerous somehow and need more regulation. But the actual decision, and the language in the Advance Act, make the byproduct approach permanent and ensure the courts can't overrule it. See here: https://heatmap.news/climate/advance-act-fusion
The YIMBY website I was on in 2010 was really into Jane Jacobs, so it's interesting that they're seen as opposites now. I think the synthesis is that Jacobs had some good ideas about how to build dense cities that people actually like to be in, but the anti-urban sprawl stance is a luxury we can't afford with today's housing situation. Any new housing is better than no building.
Early YIMBY generally focused more on "Let's build good cities rather than bad cities", but they underestimated the strength of build-nothing NIMBYS and the severity of the housing crisis. It doesn't matter how attractive your city is if people can't find anywhere to live. Hence California's population drops as house prices keep going up.
> Hence California's population drops as house prices keep going up.
I see this as due to decreasing household size, specifically baby boomers whose kids have grown up and moved out. Boomers keep their large houses instead of downsizing, while the kids can't afford to stick around and move out of state.
>(this means the claims that “AI companies are going to leave California!” or “Elon Musk is only supporting this because his AI company isn’t in California!” were even more deceptive than I thought - leaving California would make no difference!)
As one of the people that made this case, I wanted to offer that I sincerely believe it. It was not intended to be deceptive. Consider: Elon Musk hates CA and is moving his companies away (though not xAI yet), AI companies are actively avoiding regions due to AI regulations i.e. the EU, AI offers a level of control and power that pork does not, and Elon is doing extreme things to catch up in AI (e.g. securing lots of H100s, building an AI datacenter in Memphis in record time). When it comes to directing the flow of power, I am extremely wary of how power brokers behave.
I do think alternative explanations, including your own, are reasonable. I recognize it is perfectly consistent for Elon to be afraid of AI, try to control AI, and want more regulation of AI. Admittedly, fitting all these assumptions together is a strange puzzle: AI is powerful, yet people obsessed with power would not deceive to get more of it. AI is powerful, but companies wouldn't change their operations to get more powerful AI. People are anything but perfectly rational in these circumstances, so I am open to lots of explanations. I regret you believing the arguments I made are in bad faith.
Very suspicious that the day computer time starts on is also when NEPA began. We must just be in a counterfactual simulation of how things would've gone if progress slowed down. (/s)
>Nuclear is in an even worse situation, as you can tell from the recent $96 → $155 price rise mentioned above. This is partly because - as Matt Yglesias writes - nuclear plants are held to an “As Low As Reasonably Achievable” safety standard. Suppose an company invents a nuclear plant which is twice as cheap and twice as safe as existing models. They might like to try selling it for a bit below the cost of existing models, then pocket the profits. But the regulator will say that because it’s cheap, they have more money to spend on safety, and demand safety improvements until it costs just as much as existing models (even though the new model was already twice as safe).
No, this is not what that means. At some point I guess I have to write an actual article about this, but here I'll just point out that people who are in the process of building nuclear plants complain about plenty of regulations that make their lives difficult, and ALARA is not generally one of them. Also it applies to how a plant is operated, not how it's designed or built.
>The solution would be to enshrine into law some specific safety standard for nuclear - maybe 1000x safer than coal power.
A standard like this has existed since the 1980s. It says living near a nuclear power plant shouldn't increase a person's accidental death rate or cancer death rate by more than 0.1%. This has actually contributed to decreasing operating costs quite a bit for existing nuclear plants. Nobody has really built any new ones in the US except for Vogtle (which was a disaster for many other reasons) so the jury's still out on what impact it has there.
I have been reading and listening around in the Nuke space for a couple of years now and Scott's article/your comment strikes me as exemplary of recurring discussions one sees there.
I want to highlight this here because I think these debates are indicative of an ideological rift within the nuclear community that goes much deeper than mere disagreements about regulatory details or the economics of small modular vs large reactors and is not always easily discerned from the outside where people might think of nuclear essentially as one blob on the political map.
Basically - as I perceive it - in the nuclear community you have two camps. You have a nuclear start-up adjacent group who roughly believes that overregulation and lack of learning are nuclear's main problem, They aim to address this via new (small modular) reactor designs that can be factory-built as much as possible on the economical side and lobbying for deregulation on the political side. If one wants to read such a perspective I recommend Jack Devanney's Substack who argues - for example - for underwriter certification for nuclear power.
Importantly, these people are libertarian-leaning and fundamentally believe that nuclear is held back by regulation but if unshackled could be extremely competitive on the free market - perhaps even leading to a "to cheap to meter" energy abundance situation.
The other camp is more aligned with the traditional nuclear industrial-scientific-regulatory complex. While they agree that there is a problem with regulation in general, they think that this has much improved since the 70s/80s. They are also very skeptical that significant learning can ever be achieved in the nuclear industry and view SMRs (Edit: more precisely, advanced nuclear/Gen 4) mostly as warmed-up versions of tried and failed designs from the 50s and 60s. Instead the only way to make nuclear cheaper is building bigger reactors (>1000 MW) to achieve economies of scale. According to them the main problem with nuclear currently is a lack of strategic industrical policy. Nuclear know-how and supply chains have withered to a degree that we have simply forgotten how to build new nuclear. This perspective can for example be found on the Decouple podcast which I highly recommend. In Germany, the main proponent of this perspective is the nuclear historian Anna Veronika Wendland who also comes recommended.
Crucially, these people don't believe that given changes in regulation in a free market nuclear would necessarily succeed. They believe that cheap fossil fuels, e.g. the shale revolution, are as much to blame for nuclear's demise than regulatory issues.
But they still think that nuclear ought to be our primary choice of energy generation. They think nuclear is "cheap" but in ways that the free market fails to appreciate. These market failures include:
- financial markets and privatised utilities are incapable of funding very high CapEx (10-20 billion), very long term (60-100 years) projects such as new nuclear plants;
- electricity markets insufficiently reward clean, stable, always available electricity;
- markets on their own will not ensure that the supply chains necessary to keep the industry alive are maintained (the problem being that a build-out of 1-10 large nuclear plants is enough for many countries and ought to be done in series to be cost-effective, but after that workers leave the industry and know-how is lost).
These are nuclear socialists (in the social democrat tradition) who want a government-driven nuclear industrial policy including state-funded build-out programs such as the French Messmer plan and national champions such as Russia's Rosatom.
I think it is important to keep track of these fundamentally different perspectives or else one gets a skewed picture of what nuclear advocacy is all about. Scott clearly only got to hear the libertarian perspective at the Progress Studies conference.
Now who is right? Personally, my view is that the nuclear socialists tend to score a lot of tactical victories against the nuclear libertarians in debates. They tend to come from "inside" the industry and therefore are just much more knowledgeable about individual issues like ALARA or whatever. The libertarians are also just biassed against government and tend to take every absurd regulatory story they hear at face value. For example, check out the episodes on "What went wrong at Vogtle" with James Krellenstein where he goes into (excruciating) detail to show how the NRC actually performed rather well according to its mandate and did everything it could to help get the plant built.
This mirrors Scott's semi-regular anti-FDA posts on here which regularly involve him getting the facts wrong on some supposedly terrible thing the FDA did.
But I still think that strategically, the nuclear libertarians are right. Whatever the details, the core problem really *is* that nuclear is overregulated. To be precise: I think the parallel universe in which we continued to build nuclear power plants as we did before the mid 70s, achieved ~50-100% nuclear electricity generation globally but had ~10-20 Chernobyls would be much preferable to our current trajectory. Moreover, I don't think that is a likely scenario. Much more likely nuclear would have gotten much safer without getting extremely expensive through the knock-on effects of scaling and *energy abundance*. In an energy-abundant world our economy and technology level would have been so radically transformed that solving nuclear safety would likely have been a trivial problem.
I refuse to believe that we are now forever doomed to the current path. It ought to be possible to remove the insane anti-nuclear bias introduced by our peculiar nuclear history (imagine if the first contact we as a species had made with nuclear would not have been the bomb but a power plant) from our public psychology. So pro-nuclear politics ought to consist in doing the work of educating people about the real benefits and risks of the technology, making the argument for lighter regulation and thereby enabling nuclear to compete on the free market. If the winning approach will be small or large reactors, who cares?
Thank you, I think this is a good breakdown of different perspectives within the nuclear world. Clearly I fall a bit more toward the second camp than you do, but rather than open up that debate in general, let me point out one motivating factor that I think is pretty commonly felt in the industry (and has been from the start). Radiation is scary, partly for contingent historical reasons like the bomb, partly because it can give rise to particularly gruesome imagery (the radium girls, Louis Slotin's nine days as a zombie, Eben Byers' jaw falling off), and partly because it creates a threat of cancer specifically. I have heard that people go to much greater expense to avoid cancer than other ways of dying (can't find a source right now). I think public fear is a legitimate reason to have stringent safety standards, even if that fear might not be entirely rational. From the outside, to people like Devanney, this probably looks like regulatory capture or self-justifying bureaucracy. There may be a grain of truth to both of those. Public opinion has improved lately, and maybe if it improves enough then the world will be willing to tolerate another Chernobyl or two in exchange for cheap electricity. But a number of countries over the years have turned against nuclear power catastrophically, most recently Germany. Requiring proof that reactors are very safe was always meant to be a defense against that, though obviously it's not a perfect one.
Something I feel could point a finger to ALARA (or not) is Germany's experience in running nuclear plants for $20/MWh vs the US $30/MWh multi unit cost
I posed this to ChatGPT and it jumps out at me that it mentions a leaner workforce due to regulatory environment. I've heard it mentioned that the US pads the plant workforce as a means for spreading out annual worker doses, increasing the number of people who can sub into a higher dose task throughout the year
Yeah, the idea of having more workers in order to reduce dose does make sense as a consequence of ALARA. I would be surprised if it accounts for a majority of the difference in operating costs between countries, but I don't actually know.
I would be very interested in a write up on what ALARA is because I have also hear more or less this exact explanation from a lot of different sources that all seem relatively reputable.
If you read this it's clear that ALARA is mainly relevant to trying to limit the dose to radiation workers. In particular, I'd like to call out the following consideration for determining what is reasonable:
"The economics of improvements in relation to the benefits to the public health and safety—will the cost of those tools be offset by actual harm reduction?"
That is, in principle, ALARA corresponds to cost-benefit analysis (i.e. maximizing expected utility). In practice, usually nobody does these calculations. People use a qualitative approach to consider options for reducing dose and whether they are reasonable under the circumstances. Whether the company or the plant is profitable is not a consideration. But if, for example, there were a technology that would make it extremely cheap to reduce all plant workers' doses from 4 rem per year to 2 rem per year, the nuclear plant operator would be expected to implement that, because the benefit outweighs the cost. All of this seems perfectly sensible.
On the other hand, the expectation presumably is that this cost-benefit calculation is based on the Linear No Threshold model of radiation harm, which is not realistic. But I want to emphasize that the lack of realism is mainly for very low doses and dose rates. The NRC uses a cost-benefit ratio of about $5000 per person-rem. Workers at a nuclear plant are allowed a dose of up to 5 rem annually, and they may receive it all at once (for example, when doing maintenance on a highly radioactive piece of equipment). There is fairly strong evidence that a rapid 5 rem dose will raise cancer risk by an amount comparable to that $5000/rem value. In practice, most radiation workers never get near that. If the industry were spending tens of millions to reduce ten thousand workers' doses from 1 rem to 0.5 rem, then that would be justified under LNT and maybe not justified under a more realistic model. I think there may be some of this going on, but the amounts we're talking about are not a big fraction of overall operating costs.
Recently, there was a bit of a kerfuffle about the possibility of applying ALARA to design/construction of advanced reactors if they applied for a license under the NRC's proposed new Part 53 framework. I think the intent was good (that designing the plant with e.g. extra shielding to reduce worker dose would make it unnecessary to take special measures during operation). But nuclear advocacy groups objected, and the NRC ultimately agreed with them and clarified that ALARA would remain an operational principle, not a requirement that applies to design or construction.
In conclusion, I really think ALARA, while not perfect, just isn't that big a deal, and there are lots of other regulations that would be better targets for the ire of nuclear enthusiasts.
I think there are some fundamentals where I'm more optimistic about a radically low nuclear system cost
I was going to say a canonical wind and solar grid requires transmission to triple, but a quick search seems to indicate tripling that would increase system costs by 25%, not severe. This does still create a regulatory vulnerability which involves high voltage lines flying over states that neither sell into or buy from them due to the point to point nature of DC lines
Transmission plus distribution does add to 50% of delivered electicity costs meaning that a high capacity factor onsite power source could deliver radically lower costs. Solar could do this with batteries while nuclear has ~done it with less regulation. I think pretty equal weight should be given to a "$10/MWh" nuclear or solar future
I believe the solution with nuclear would rely on a system that fully balances regulator incentives for safety and economy. This exists in the commercial shipping insurance system where the insurers define a regulatory and inspection system that you have to meet to get insurance. Too strict and the insurer loses business, too loose and they risk losses. This would also require a dose rate based rather than linear radiation harm model, while the market system would itself erase ALARA, as the reactor operator would not buy insurance that came with it attached
Under this proposal, what happens if, say, an Iranian insurer offers cushy rates and minimal questions asked, conditional on handing over the occasional ingot of highly-enriched "waste" ?
Just physically, nuclear inventories are closely monitored. It's technically possible for small distributed amounts, probably from several plants over a large time frame, under the inventory precision to add up to a critical mass (if reprocessed, Iran went the easier route of building an enrichment plant). However it's not practical to my understanding
More generally, if it was legal to sell your insurer nuclear waste, it would also be legal to sell it in general, which is not part of the proposal! Nuclear plants do have insurance providers without a regulatory aspect but they cannot give spent fuel to them
Nuclear plants (in the US at least) have very limited ability to get insurance. Liability insurance is only available up to about half a billion dollars, which is peanuts compared to what they would owe in a Fukushima-type accident. (They also "self-insure" by agreeing to each pay some of the costs if any of them has an accident.) But ultimately the government covers the bulk of the risk. It seems like insurance-based safety rules only work if private companies are willing to provide that insurance.
Yeah, Jack Devanney likes to say things like that. But even he admits that we'd need wholescale tort reform to make it work in this country, so good luck with that. His implied policy for Japan would have involved spending many millions on radiation monitors and dosimeters prior to the accident (i.e. at every nuclear plant) and then refusing to compensate the people who lived nearby because their doses weren't that high. All indirect damage (like loss of business for local farms or fisheries, let alone psychological effects) he would prefer to ignore. I'm still not sure what he thinks should have happened with the cleanup costs, which are nearly as high as the liability. Maybe he thinks they should have left it to rot, let radioactive material dribble out forever, and given tiny amounts of compensation to people who decided to live nearby anyway. I think that sort of response would be a perfect way to convince voters to make nuclear power illegal, as they have done in many places in the past.
The SNT model, by the way, is not "still conservative." It's guesswork, and fairly optimistic guesswork imo. He assumes a 1-day repair period, and he claims we don't see chronic harm below 20 mSv/day (which is about 1000 times the natural dose rate even in high-radiation places like Denver). He's right that LNT is unrealistic and highly conservative, but he doesn't seem to think it's possible to go too far the other way. The society of health physicists has argued in the past (though it's not their current stance) that a reasonable dose response model would calculate cancer risks linearly but only for annual doses above 50 mSv or lifetime doses above 100 mSv. That would make a huge difference (compared to LNT) in number of expected cancer fatalities, but it would still suggest that the people living closest to Fukushima benefited from temporarily relocating, and that many deserve some compensation.
[Health Physics Society, “Radiation Risk in Perspective: Position Statement of the Health Physics Society” (PS010-2), adopted January 1996, revised July 2010, McLean, Va.]
The compensation scheme pays everyone according to their dose above some cutoff. His concept involves some form of local approval for the plant to go into operation in return for both building it and to agree with the compensation rates. Likely on the lines of paying a special tax rate or funding schools
I believe the compensation also pays non-radiation costs like business disruption. Based on SNT there would be no motivation for large scale replacement of topsoil and the like as the radiation compensations would be far less expensive. The Fukushima exclusion zone is about three times safer than Tokyo when comparing radiation rate to air quality
If his model is accurate the benefits would outweigh the expected potential costs very easily for the local people. One potential controversy is that it pays a flat rate for dose with no special compensation for developing cancer, which is not provably related and unlikely. This would again be worked out in the contract with the local community in order to accept that
"It feels like the United States, after a fifty-year binge on over-regulation, has woken up, wiped the vomit off its chin, noticed it’s lost half its net worth, and started to consider doing something else. I am equally confused why it took so long and why it’s happening now."
Wild-eyed, half-baked hypothesis: maybe the people who remember the real environmental degradation that took place prior to 1970 (e.g., the Cuyahoga River catching fire, acid mine drainage turning streams bright orange throughout Appalachia, dead zones in the Chesapeake Bay, etc) became overly enamored of regulation as just the price we pay for clean waterways. But now that the US has to a great degree shed its dirty industries anyway and the people who remember the world pre-1970 are dying off and being replaced by those who don't, the mood is shifting.
The water in our local harbour used to turn blue when the leather factory released its wastewater into it. Nowadays the factory is gone and the water is (relatively) clean.
People who weren't around during the 70s before the big push on environmental regulations really don't know the other half of the story. Manufacturing facilities in the western world don't release pollutants into the air and water today, but that's not out of the goodness of their hearts or civic-mindedness, it's down to regulations and legislation being applied.
Re. "Why is solar improving so quickly?": Two overlooked points. One, solar has been heavily subsidized by governments, at every level: research, development, production, and retail. Two, solar has low costs now only because coal and nuclear plants can kick in at night or on cloudy days. On a cloudy day, a solar panel's output doesn't drop linearly with the sunlight; new solar panels that I've tested have all dropped to zero output even at noon when the sky is cloudy. There is a minimum of solar power they need to overcome the resistance in the panel to produce any energy at all.
I recently calculated that, if we replace fossil fuels with electricity, then for the US energy grid to store enough solar power for one cloudy day across America, we'd need enough lead-acid batteries to hold 80,293 gigawatt hours of power. A lead-acid battery stores about 35 watt-hours per kg. So we'd need about 2,294 billion kg of lead for batteries. There are only about 95 billion kg of lead left on Earth that we know of. The situation with lithium is much worse.
We can, however, store energy kinetically. For instance, build dams to create lakes, and use electricity to pump water up into the lake, and then let it out to make hydroelectric power. Hydroelectric is about 90% efficient, and big electric water pumps are over 90% efficient, so we should get > 80% efficiency out from the output of the solar panels.
Pumped hydro is great in places that already have dams with big reservoirs. But we pretty much dammed up all the best places many decades ago, plus the environmental concerns around dams are more acute now than they were back then, so building lots more of them in the US is implausible.
I did see a cool idea a while back for storing gravitational energy by cutting free a giant cylinder of rock in the ground and then pumping high pressure water underneath it to make it rise up above grade. The advantage is that the cost should scale with the surface area, but energy capacity scales with the mass, so the square-cube law sort of works in our favor (though not as much as with dams). Downside is lots of expensive digging, among other things. Also I'm not sure how you structurally stabilize the rock once it's towering up in the air.
carbon nanotube nanostructures have the potential to use mechanical stress to store energy at similar density to lithium-ion batteries, using much more common atoms. But that technology is way beyond our current tech level.
Accelerator driven nuclear could easily be throttled up and down as the inverse of solar supply, but in that scenario it'd probably cheaper to run the nuclear 24/7 and dispense with the solar.
Lithium batteries use on the order of 1kg of lithium per 10kwh of capacity, so 80TWh would use on the order of 8 million metric tons of lithium, which is only ~40x current annual lithium production. As with all natural resources, proven reserves keep expanding and the amount available depends a lot on on whether the price makes it worthwhile to explore for more.
Utility-scale battery storage is already down to ~$120/kwh, and these batteries can cycle 3000x, so ignoring discount rates that's 4 cents per kwh. The commodity lithium is <10% of the cost of the batteries. At CNY 70,000 per tonne of lithium carbonate, that's about $105k per tonne of lithium, or $10.50 for the lithium in 1kwh of capacity. So as the manufacturing gets more efficient, there's room for battery prices to decrease at the same time as lithium prices go up and lithium supply increases because of the increased lithium prices. This happened in late 2022 when lithium was expensive, and then a lot of new supply came on line and lithium prices crashed by almost 90%. See the history of failed predictions around "Peak oil" for an analogy to supposed constraints on lithium supply.
Just tried to reproduce your lead conclusions, don't think its right:
USA uses more like 12,000 gWh in a day, not 80,293.
A lead-acid battery stores about 35 watt-hours per kg of battery. Lead is only 30% of this weight so its actually more like 110 wh/kg lead.
So we would actually need about 100B kg in your scenario.
Meanwhile, while there's 95 billion kg of lead reserves, it's lead resources that are more akin to "lead left on Earth that we know of" and that is 2000 Billion kg.
> If we had cheap clean limitless power, what would we use it for? Someone asked this question in a presentation, and the cheap clean limitless power advocates didn’t have a canned answer ready to go. They eventually proposed things like improved public transit, supersonic flight, carbon capture, AI compute, geoengineering to prevent hurricanes, and city-wide air filters (really).
Seriously? No one mentioned the obvious? The thing that would immeasurably improve quality of life for whole civilizations, if only it weren't so energy-costly?
De. Sal. I. Na. Tion.
If you can make water purification cheap and abundant, at least a dozen major problems just vanish in a puff of vapor.
Maybe someone mentioned this and I forgot, I took pretty bad notes.
There's also the immediate use of removing carbon from the atmosphere to reverse global warming.
Scott mentions it in passing for producing carbon-neutral hydrocarbon fuel, but I'm talking about net removal of carbon.
Also, we need to bury those new hydrocarbons in case post-collapse civilizations need fossil fuels like we did!
If you want to use solar power to remove CO2 from the air and fix it in a more solid form, you don't need high technology or cheap power to do that. Just plant a tree! Or a few thousand trees.
On a related note, the biggest contributors to the "urban heat island" effect are asphalt (turning cities almost literally into Black Bodies) and the lack of vegetation. Adding a few shade trees can reduce the scorching temperatures of a hot summer sun by more than 40 degrees F.
I think trees are much less efficient than purpose-built carbon-removal machines.
Perhaps. But they're also much less expensive, and provide ancillary benefits such as shade, fruit, and wood.
"Efficiency" necessarily takes expense into account. I think this is one of those situations where trying to get everything at once is the wrong strategy. Imagine if we damaged our ability to fight global warming because in order to get one billion pieces of fruit. We should let private industry create as much wood and fruit as there's demand for (which they're currently doing, in orchards and timber farms), but also pay for carbon removal at some socially meaningful rate (which would incentivize private industry to capture carbon directly, and then the industry could choose between trees and machines, and if I'm right and machines are more efficient even when counting the extra profit from fruit, they would choose the machines).
Perhaps. I guess some of it is just my own pro-fruit and pro-shade biases peeking through.
When my wife and I bought a house a while back, one of the first things we did, before the previous owners had even moved out, was go to a local nursery and buy a young fruit tree. We waited until they were gone to actually take delivery and plant it in the yard, of course, but that's how important it is to me.
This only works long term if you then cut down and bury the trees deep under the earth, otherwise the carbon is released when the tree dies and decays. Of course, cutting down, transporting, and digging mines to bury the trees is going to take a lot of work with heavy machinery that probably runs on fossil fuels, and will require a massive amount of land dedicated to the task since trees grow slowly relative to our rate of emissions. If your high tech carbon machine can make something like graphite bricks, that would be much easier to deal with logistically, and possibly even useful for some purpose, and is relatively stable so even if it ends up in a landfill won't just turn back into atmospheric carbon. I'll also note we already farm trees for wood and paper, most of than ends up in a landfill and decays at some point or gets burned, re-releasing all the carbon.
That all depends on just how long of a long term you're talking about. Building long-duration objects like houses and furniture out of wood does a pretty good job of carbon-sequestration for a decent length of time.
And trees lower surrounding temperatures. Nothing like photosynthesis to absorb energy! Just walk from barefoot from a grass lawn over to an astroturf lawn, and you'll feel the difference! The grass lawn remains coolish in the afternoon heat, while the astroturf will burn your feet.
Yeah, that's what I mean by the benefits of shade. (See above, re: urban heat islands)
Photosynthesis is REALLY inefficient. What you are feeling is mostly evaporative cooling. Ie plant sweat.
Actually, how good and how expensive are purpose-built carbon-removal machines right now? (As Bob Frank says, there are other factors to take into account, but I'd like to get the right order of magnitude.) Compare them to, I don't know, a particularly efficient tree (or bush or weed or what have you).
Here's an article citing ~$1000/ton for the climaworks plant: https://edition.cnn.com/2024/05/08/climate/direct-air-capture-plant-iceland-climate-intl/index.html
To get back to the pre industrial revolution levels of CO2 we'd need to remove 1e15 tons of CO2 from the atmosphere, so that's about a quadrillion dollars. To be fair your number is probably for mass of carbon rather than mass of carbon dioxide so let's call it 300 trillion dollars.
It sounds like an untenable solution, in that you can throw an unfeasibly large slice of global GDP at the problem and still fail to make a meaningful dent in it.
Grinding up CO2-absorbing rock, and dumping it in some warm part of the ocean to speed things along, could be done with current tech at a cost two orders of magnitude lower: https://worksinprogress.co/issue/olivine-weathering/
They might be eventually, but they are not cheaper yet. We're going to need both, for a lot of reasons (industrial point source capture, higher output per acre, ability to site capture where we need the carbon for utilization, and on the other side of course we still need wood and fruit).
Reforestation is an politically easier sell, though.
Unfortunately, a lot of what gets sold as “reforestation” is actually trying to force a forest to exist on land that people didn’t find useful, and which can’t support trees without lots and lots of inputs and attention.
Certainly when it goes by the name of “tree planting”, which has a nice ribbon cutting, and then no one pays attention when the trees all die.
I do volunteer work with a wildlife trust (I know its not very effective altruism) and its sad how true this is. Just this week I was at a session where we were mostly undoing the harm done by a well meaning landowner trying to reforest land poorly. Boring charities like that which just do tree planting and maintenance as part of their regular operations rather than "adopt a tree bring your kids to plant a forest!!" seem to be the only ones who actually do tree planting effectively. My trust manages about a 95% survival rate with the only maintenance after the planting being removing the tree guards.
99% of that problem is planting the wrong trees in the wrong soil. Most EU for example has a fetish-like fixation on growing vast spans of pine, on soils that do not support pine at all, and for variety, force poplar, linden and decorative apple-trees in the middle of a city where they instantly get sick and die.
Meanwhile, the ecosystem of continental Europe is mostly geared towards various types of oak.
It's not that they're less efficient. It's that there isn't room for the needed amount of trees. (We can't even keep the ones we still have.)
I've got no idea whether they're less efficient, or even how you would measure efficiency in that context. Personally I find just having trees around is a net benefit. So by that measure trees are extremely efficient.
Who is we? The U.S. is reforesting rapidly
We is humanity. And I'm not sure the US is reforesting at all. I see lots of forests of dead and dying pine trees. Canada isn't the only country that's experiencing a increase in forest fires. (OTOH, it's been about a decade since I rode along 5 and saw all the dead and dying pine(?) trees. (I assume they're pine, because they're some sort of conifer, pine trees are common in the region, and I'd just been reading about the pine beetle infestation. But the exact species is irrelevant to the point. The area was still called "forest", but about half the trees looked dead or dying.)
I once calculated that the entire world’s net carbon emissions could be neutralized by planting paulownia trees in an area the size of Ukraine and burying all the resultant wood. This would be considerably less expensive than the dozens of trillions people want to spend on other methods, and not require hypothetical future tech.
It would be even simpler to plant the correct species of trees in the correct environment and let it do its thing. Sure, paulownia is the most efficient carbon grabber, but its not the easiest to grow on any type of soil. The beat solution is just to replant the logged areas with the types of trees that used to grow there before humans cut them down.
Direct air capture costs $500/ton now, and the EPA SSC at a 5% discount rate is $12/ton. It doesn’t make any sense to use discount rates lower than nominal gdp growth, because 1. any public good that increases productivity by a fixed amount forever would have infinite value. 2. Other investments can easily beat the rate of nominal gdp growth.
perhaps, but an average tree absorbs approximately 22 kilograms of carbon dioxide from the atmosphere each year, so a copse of trees can make up for one car. If you add all the other economic, social, psychological and environmental benefits of trees, there is no excuse why we are not spamming saplings and seeds at every available surface.
I am not a botanist/ecologist, is it true that there is tons of land where trees would naturally flourish without further human intervention if seeds were planted, and the problem is just that no tree seeds have ever gotten there through normal ecological processes?
That seems like a stunning L for evolution and ecological equilibriums, but maybe it's trivially true?
Also - since bamboo grows so fast, is it a more efficient way to capture carbon than slower trees/plants?
I think for many purposes, bamboo counts as a “tree”, because it is a tall woody plant. Basically every plant family separately re-invented the “tree” body plan, and bamboo is the grass family’s version.
A lot of tree planting initiatives fail for the reason you mention - they figure “trees are good, we aren’t using land, let’s plant some trees here” and don’t stop to think about why there aren’t currently trees on this land if we aren’t using it.
There is a more relevant plan though of finding land where humans cut down forest to turn it into agriculture or coal mines or whatever, and get people to stop using it for agriculture or coal mines and let the trees come back. Planting seedlings sometimes accelerates the process a bit compared to just letting the trees seed it (particularly if there are apex trees that used to live in the region but now only exist far away).
The number one region where replanting is needed are urban areas. Trees wont naturally replant on carefully maintained asphalt no matter how long you wait.
Yeah urban trees are great. They also have the advantage that they are often actively maintained, so they can survive even if the local climate wouldn’t support them. They aren’t that significant for carbon reduction though, since there’s so little urban land and the possible tree density is limited. They are good for human benefits though.
Hmmm, just had a thought: bamboo is carbon negative whereas concrete is carbon positive. So have a building material consisting of mostly bamboo with concrete pouted in between it.
Convenient compilation of prior research on related subjects: https://projectrho.com/public_html/rocket/stellarcolony.php#bamboo
I think you've just reinvented wattle-and-daub.
There's an enormous amount of land that used to be old-growth forest and has been cleared by humans over the millenia. IDK how suited it necessarily is for trees *now* - in a lot of cases we've probably made the soil less fertile - but in principle it's no knock on evolution that trees aren't growing in fields we deliberately cleared of trees.
With that said, land is ultimately finite, and there's an awful lot of fossil fuels in the Earth's crust. Also, we're using some of that land for stuff.
Another area that is ripe for replanting are man made cattle ranches. I don't know what happens to feed grass growth under a canopy of trees, but there's probably some mutually exclusive dynamics going on (else, why clear the land in the first place?).
Generally a closed forest canopy doesn't allow enough sunlight for grasses to grow. And if the forest floor is carpeted thickly enough in herbaceous material like grass or ferns, it impedes the regeneration of tree seedlings.
the main issue is the water table, not the soil itself. As long as the land is the right level of moist, you can replant the trees that used to grow there since the last Ice Age, and they will beat the odds eventually.
The main problem with reforestation is the fetish-like desire to plant the kind of trees we want, not the kind that can survive there, and drying/watering the land to achieve that. Europe takes the cake here, we destroyed countless square miles of perfectly fine, soggy-soiled oak forests and now try and fail to replant them with pines. We buried miles of alder and willow-rich swamps in sandy soils to force more pine there, which promptly rots at the roots from too much groundwater.
Bamboo is considered an invasive pest by many people, because it grows so quickly in places where they don't want it.
There is a fair bit of land that was covered in forests. Then humans chopped all the forest down and didn't replant it.
A lot of that land is covered in farms, but not all of it.
There's factors like soil ph, soil nitrogen content, slope, drainage, and sunlight. Different trees have different preferences, and might do poorly in a non-optimal environment. I can imagine a situation where a change in the local geography means that trees suited for environment A can't get across a stretch of environment B to an uncolonized patch of environment A.
That said, birds and deer and other similar animals often "carry" seeds around, maybe even depositing them with a bit of fertilizer. But then, animals like deer also eat young trees. So the animal factor works both ways.
The problem with trees is that they take a lot of space. There are limits to how much more of the planet we can permanently cover in forests.
Trees are actually generally not an effective carbon fixation method in the present environment. They used to be, in the carboniferous period, before microbes evolved which could decompose wood. But now, dead trees generally release their carbon back into the atmosphere over the course of their decomposition. Reforestation is carbon-negative, but forested land is carbon-neutral since it's in an equilibrium where it's releasing as much carbon through decomposition as it's taking in through photosynthesis. Although, you can sequester the carbon longer term if you cut the tree down and treat it in a way that makes it resistant to decomposition.
But over a long time frame, carbon sequestration is not usually a meaningful benefit of planting trees.
But surely the trees are not being planted in existing forests?
The point is that trees only sequester the carbon temporarily before releasing it back into the atmosphere. Forests are carbon neutral because they're in an equilibrium between trees sequestering carbon and releasing it. If we want to sequester carbon, trees don't offer a very long-term solution, unless we cut them down and store the wood under conditions that offer long-term protection from decay. Whatever other benefits trees provide, other methods come out far ahead in terms of effective carbon sequestration.
But surely, new forests are not carbon neutral compared to a past where there was no forest?
Are you assuming that any tree you plant dies without reproducing?
You have to bury the trees and make sure they don't decompose. But you're probably going to have to bury whatever hydrocarbons you get from carbon capture anyway, so I'm not sure how big a difference it will make in the end.
Or actually use the wood and then replant the trees; repeat. Are my bookcases decomposing faster than wood would in a mine?
What are your grandkids going to do with your bookcases when they inherit them, given that they probably won't care to own any books?
That’s about as likely as their not riding bikes, or having pills for dinner.
Throw them away, letting them get buried in landfills. This study says that it mostly doesn't get broken down. I think. I'm not good at reading studies. https://www.sciencedirect.com/science/article/abs/pii/S0956053X1730942X
People did the calculations on that, and basically, trees don't remove enough carbon.
I ran some rough numbers. And the amount of forest we need to return CO2 to per-industrial levels worked out to be 2x earths land area. (Rough numbers.)
wonder, if there is a cheap, environmentally friendly, and technologically durable alternative to asphalt that would be white (or light gray/beige/pink/light blue whatever) instead of near black? Or maybe its just a matter of adding some white agent to existing asphalt to make it lighter in colour?
The first thing that comes to mind is concrete. It's light gray in color, abundantly available, and just as good as asphalt at most of the things that asphalt does. Unfortunately, it has well-known problems with cracking under thermal variance that asphalt does not have, making asphalt the clear winner in terms of durability. I've read about a few different groups working on solutions to this, but nothing that's commercially viable quite yet AFAIK.
"Carbon capture" is literally in the list that Scott wrote and the person you're responding to quoted.
I mean, the correct answer to "What do we do with clean, limitless power?" is "Literally everything". Everything that isn't a physical plot of land becomes cheaper and/or more accessible.
If the question is narrowed to "What do we do with clean, limitless power that we don't already do a lot of?", then space flight, routine terrestrial flight, desalination, and massive reshaping of the Earth's crust are good options.
Fair enough. But a lot of those would require other advances in addition to cheap energy. Desalination really wouldn't. The only major other thing that would be required is to ship the purified water around in bulk, and we already know how to efficiently ship liquids in bulk; we do it with petroleum all the time.
Remember that we can apply purification technologies to wastewater as well as salt water.
1/3 or so of water today is used for thermoelectric power generation. That goes away in a nearly-all-solar world. Residential consumption usually goes through wastewater treatment facilities, and with current technologies implemented at scale we could get 80-90+%recovery (a lot of these systems can also be implemented in a home if you want!). Farming is the more difficult case. Can't really recapture transpiration water unless you're farming indoors... which gets a lot easier with super cheap energy. But even without that, a mostly-solar-power world can cut fresh water demand by nearly 2/3 long before you start having to worry about shipping water from coasts to farmland.
This is true, but that's mostly cutting freshwater demand from places where freshwater is already abundant enough for farming. But places like Australia, huge swathes of the Middle East, etc. would become vastly more habitable with access to cheap desalinization, much more easily than we could divert water to there from other more arable locations around the world.
100% agreed
I'm imagining a gigantic desalination plant on the Mediterranean, with the outflow diverted to the Qattara Depression.
Could do halfway-indoors agriculture with inflatable tents, like an enormous transparent air mattress: https://caseyhandmer.wordpress.com/2019/11/28/domes-are-very-over-rated/ Natural shape doubles as a funnel directing both rainwater from above, and any condensation on the underside, toward some sort of combined irrigation / HVAC machinery at the anchor points.
We know how to ship petroleum for a price of several dollars per barrel. That’s not going to be a viable price for water unless there’s a very special place that has a reason to exist even with really expensive water.
If the power to run the pumps on your pipeline is cheap, that cost comes down pretty significantly.
Does it? I would have thought most of the cost was actually building and maintaining the pipelines (and more importantly, tanker ships), not the energy for pumping. In any case, no one has ever transported oil at the scale that water would be transported here. The entire world’s oil consumption is only a bit more than New York city’s water consumption.
Well, I don't think you'd be shipping much desalinated water by tanker ships, you'd produce it at the nearest coastline.
In terms of flow rates, the trans-Alaska pipeline can ship 340,000 cubic metres per day, which is about 4 cubic metres per second. The flow rate of the Thames River (to pick a medium-sized river that most people will be kinda-familiar with) is about 100 m^3/s so you'd need 25 trans-Alaska pipelines to create one Thames. The trans-Alaska pipeline is only 48 inches in diameter (smaller than most TVs) so it's not hard to imagine scaling this up by a factor of 25 at less than 25 times the cost.
Desalination costs are about 1/3 power. We'd need other advances if we want to get more than 50% extra water for the same cost.
How much of that is because the existing desal plants were optimized for energy being expensive? A setup designed for intermittent operation, taking advantage of electricity prices down around "we'll pay you to take it," might be able to get away with far lower capital costs relative to throughput capacity.
I have no idea. And looking for a better source it's a third to half the cost. You're probably right that we could get it cheaper than that.
Possibly useful, though I haven't double-checked it myself: https://caseyhandmer.wordpress.com/2024/10/26/we-can-terraform-the-american-west/
> The only major other thing that would be required is to ship the purified water around in bulk, and we already know how to efficiently ship liquids in bulk; we do it with petroleum all the time.
Quantity scales with water are quite a lot bigger than for petroleum,
Humans already move large amounts of water around with dams, canals, pipes ect.
What do you mean by massive reshaping of Earth's crust?
It currently costs a lot to move enormous piles of dirt from one place to another, especially if you want to do it in a precise way. (e.g., not exploding a mountaintop). This is mostly because dirt, rock, and clay are heavy, which means a lot of force is required to move it. With boundless energy, you can generate the force more easily and cheaply, meaning we'd probably do a lot more customizing of medium-to-large swaths of land. Things like building local mountains, valleys, and the like. Like how in Sim City 2000 you could start with a "blank" city and then just create the rivers, oceans, and mountains you wanted. Applications include aesthetic appeal, preservation activities, and for large scale industrial efforts.
I'm sceptical that sufficient cheap energy is the main limitation towards doing things like building mountains and valleys. The engineering challenges regarding f.ex. geotechnical, environmental, hydrogeology, hydrologic issues etc. etc. would be daunting.
The real answer is "create jobs." Everything else is detail.
We could use the limitless clean energy to rotate wheels in one direction, and then employ a lot of people to rotate them back. Problem solved.
But those jobs won't create yet more jobs. A functioning economy consists of multiple nested recursive systems, replicating themselves outward.
Yes, I know you were being facetious.
Right. It's sort of like asking "what would you do with more money?" I dunno...a bunch of different stuff, most likely.
Sufficient abundant energy and automated labor solve every other scarcity challenge. This is just totally obvious.
True, but the words "and automated labor" are doing a whole lot of work there. (No pun intended.)
It doesn't solve the scarcity of land in fashionable cities, which is the only scarcity problem that the Western middle class cares about.
>which is the only scarcity problem that the Western middle class cares about.
<mildSnark>
Well, there is also the limited number of admissions to prestigious universities. :-)
( (near)-zero sum from "prestigious" )
</mildSnark>
Cheap energy means cheap steel, allowing more and taller skyscrapers. Cheap lighting, potable water, ventilation, imported manufactured goods, so the spaces below remain pleasantly habitable. So, yes, to some extent it could.
The greater obstacle is land speculation, solution to which is a land value tax.
The current limiting factor on tall skyscrapers isn't the cost of the steel, it's the planning permission.
And the difficulty of that planning permission is deeply entangled with incentives from land speculation, which I already agreed is the greater obstacle.
The cost of building is actually significant. Look at skyscraper costs in UAE for example. It's billions of dollars.
And most of that cost isn't steel. There is a lot of labor. And a fair bit on carpeting and plumbing and some on equipment.
Ok. Everything is connected. If steel gets much cheaper, then sky scrapers get a bit cheaper, and so slightly taller. But this isn't a big effect.
abundance of free energy would do wonders solving global poverty, improve literacy and education worldwide, thus making Western Middle Class mostly obsolete, and thus, irrelevant. WMC are a bunch of specialists that we only need until we can automate them away or have developing world's lower middle class do their jobs remotely.
I thought that even with free electricity, there is a big problem of what to do with the leftover brine. It's corrosive, and toxic to ecosystems.
Well, just for starters, a lot of it is directly useful. You can extract salt, other mineral resources, and even small amounts of gold from seawater.
The gold isn't worthwhile, but the uranium might be. It seems right on the edge of practicality. (But what' you'd use it for if electricity were really cheap isn't that clear.)
Uranium would still be critical when optimizing for watts-per-kilogram, mostly meaning space travel. Gold might be worth setting aside a channel for if you're going to be thoroughly sifting and sorting all the rest of the sludge regardless.
Like Bob Frank said, it can be used. The chemical industry will guzzle it up in a lot of cases, certainly for all of the desalination plants that are geographically close to chloralkali plants that need brine as a raw material, and potentially for the ones that are further away as well depending on demand.
a pool a brine is essentially a giant battery - we could do interesting things with it.
But realistically, the amount of brine that would result from this is negligible compared to the size of the ocean. Just tow it a 100 miles off shore and dump it into the Pacific. All the post-salination brine we could ever produce would be like a single grain of salt in a bathtub when thrown into the ocean.
That would be my immediate answer too. There is hardly a more fundamental issue than access to potable water. Even in regions like Central Europe, which have never been in danger of permanent water scarcity, falling groundwater tables have been causing concern in recent years. Climate change and changing rainfall patterns will only compound that problem.
Where would Central Europe get water to desalinate from though?
Well the point was more that water is or can be an issue just about anywhere. If it really came to a need for desalination for Central Europe though, it would be solvable in principle, for example through desalination on the coast and transporting the fresh water through pipelines.
https://en.wikipedia.org/wiki/Pipeline#Water
Bring back aqueducts! The future is retro!
Hours since I last thought about the Roman Empire: 0
From what I can find, energy is about a third of the cost of desalination. Getting rid of that cost wouldn't be that big a deal.
Huh. Where did you see that? I'll admit I'm no expert in the field, but what I've heard from people who are is that energy costs are the largest factor by far. If you've got solid data that says otherwise I'd love to see it.
Google. Looks like my first result was this: https://www.sustainabilitybynumbers.com/p/how-much-energy-does-desalinisation
Googling around more I found this better-looking source: https://www.sustainabilitybynumbers.com/p/how-much-energy-does-desalinisation which said it's a third to a half. Do you have any sources at least that good saying otherwise? Or say the same thing. The important thing is to get accurate information.
What are the dozen major problems that desalinization would cause to vanish?
We could have home appliances that never break. Right now, our dishwashers, refrigerators, etc. are designed to be as energy efficient as possible to the detriment of being long-lasting. *This is an excuse I've heard, planned obsolescence being unreasonably financially successful is probably playing a role here too.*
The biggest problem with dishwashers isn't energy; it's that they can't *scrub.* It's the classic "we have to wash the dishes before putting them in the dishwasher" complaint: if you put dishes (or especially cookware) in with non-tiny amounts of food stuck to them, it's more likely to end up baked on in the drying phase than cleaned off.
Adding more power won't fix that. More likely would be the outcome (admittedly exaggerated for comedic effect) memorably demonstrated in the pilot episode of Home Improvement, when Tim soups up the dishwasher with More Power™ and it ends up exploding.
Desalination is already cheap and abundant.
Water Desalination is a big one, but you could also do some pretty insane recycling of materials with super-cheap energy.
Eventually you'd have to. At some point it has to be easier to recycle landfills to get at various metals than to dig up increasingly less profitable mines, right?
Seems to me the question of what we'd do with cheap or near/limitless power is similar to someone in 1970 or 1980 wondering what consumers would do with a personal computer.
I agree that fresh water is likely one of the commodities which will become much cheaper if the price of energy goes to zero.
An other example would be aluminum, where 34% of the costs are electricity costs, and another 37% of the costs are for the raw materials (alumina, etc), which would likely also drop if the prices for electricity and metals would drop.
For some plants, it might become economically feasible to grow them under optimal artificial light 24x7 (the way cannabis is sometimes grown in areas where it is illegal), thereby decreasing prices of raw food.
That is a very nice closing sentence
I'd be grateful if anyone could explain the costs of solar to me. On the one hand I see this 'Our World in Data data', which are "without subsidies", and which show that costs are already among the lowest (lower even than nuclear).
On the other hand, I keep seeing people post what we (in the UK) actually pay solar providers via "contracts for difference" (https://www.iea.org/policies/5731-contract-for-difference-cfd) and it is usually dramatically higher than any other source.
Prima facie, the figure with subsidies seem like it might be the more relevant, but it depends on to what extent the subsidies are actually necessary in order to continue receiving the energy from this source.
The purpose of CfDs is to drive investment into renewable generation by making sure they generate enough revenue, meaning the strike price will be set at a figure high enough for payback and profit generation. But in any case the UK is an awkward market to look at investment in generation assets because the planning system is so terrible.
Lazard recently published their 2024 analysis of LCOE, including a slide on costs with and without US tax subsidies. Here's the headline chart:
https://imgur.com/a/R8UJlWz
And here's the impact of subsidies on renewable energy:
https://imgur.com/a/gRtpIKn
The chart that Scott provides shows the "Levelized Cost of Energy". For solar it is the cost of electricity delivered locally by the solar panels when the sun is shining. It includes the amortized cost of producing and installing the panels. It is not the cost of so-called dispatchable power that is available on demand 24/7 at distant locations via "the grid."
For those who really want to get into the details of deploying carbon-free energy, I strongly recommend reading the Department of Energy's Advanced Nuclear "Liftoff Report." https://liftoff.energy.gov/wp-content/uploads/2024/09/LIFTOFF_DOE_AdvNuclear-vX6.pdf
It provides an excellent, detailed, technical analysis of both renewable and nuclear energy production, looking at everything from costs and cost-curves to tax policy to land efficiency. (My own aesthetic preference is for solar to be deployed on rooftops and in urban areas and not to cover hundreds or thousands of acres in solar panels for commercial generation. Nuclear is *much* more land-efficient on a energy-unit basis. See fig. 13 of the report.) There is an interesting discussion and critique of the Levelized Cost of Energy (LCOE) metric. There is also a counter-intuitive graph showing a cost-analysis of decarbonizing California's grid using renewables with storage only versus renewables with storage AND nuclear. (Fig. 6.) I know most people won't read the report, so here is a helpful and, I think, accurate summary:
"Nuclear has an essential role in the energy transition as a clean firm complement to renewables. Nuclear provides clean firm capacity; modeling shows including nuclear and other clean firm resources with variable renewables reduces the cost of decarbonization. Nuclear can help address the power needs coming from load growth, where much of the demand is disproportionately for 24/7 electricity, e.g., data centers. Nuclear does not “displace” or “compete with” renewables; decarbonization requires both nuclear
and renewables. Nuclear provides clean firm generation that enables the increased deployment of variable renewables like wind and solar."
https://caseyhandmer.wordpress.com/2019/06/21/is-nuclear-power-a-solution-to-climate-change/
The world has a lot of cheap low value land that is used as cattle ranch or is just sitting empty.
Nuclear uses less land, but a lot more of much rarer and more valuable resources. Namely engineering and regulatory competence.
Solar is rapidly getting cheaper. Nuclear takes a while to build, and longer to pay back.
Cost curves and economies of scale mean that once solar gets far enough ahead, nuclear becomes financially nonviable and can't catch up.
This is something of a battle of assumptions that will only be resolved by time. There are some reasons, however, to think that nuclear will play a critical role and that Big Tech's recent embrace of nuclear electricity generation is not an ignorant mistake.
First, aesthetics of solar farms aside aside (and the fact that one person's "cheap low value land" is another person's beautiful open space with various endangered species), transporting electricity long distances from massive solar farms on "cheap low value land" to urban and industrial centers requires large, expensive upgrades to the electricity transmission grid. Those projects are time-consuming, regulatory nightmares that rival the construction of nuclear plants before the federal government decided to promote nuclear energy. Check out the history and costs of the SunZia Transmission Project for an eye-opening example.
Second, as noted above, the Department of Energy itself has concluded that "decarbonizing the grid will be very difficult and expensive without 20-40% clean firm power. Firm power refers to power or power-producing capacity intended to be available at all times during the period covered by a guaranteed commitment to deliver, even under adverse conditions. With an increasing portion of the grid supported by renewables, the value of dispatchability provided by firm power increases. A variety of technologies including nuclear can help maintain grid stability via synchronous inertia, reactive power, and other benefits."
Third, many of the growing use cases for electricity are especially sensitive to even small fluctuations in power that make even solar-with-battery-backup infeasible. Although some Big Tech companies have claimed to use 100% renewable energy before their open embrace of nuclear, this is merely a financial claim; their data centers, especially those devoted to AI, are actually powered by the most stable "non-renewable" sources of electricity-generation (including taking power from "the grid," which is stabilized by utilities using firm power sources).
Casey Handmer's arguments are interesting, but the people at the Big Tech companies who are responsible for securing large amounts of electrical power for real world uses have looked closely at the same issues and the same straight-line graph forecasts and have decided to embrace nuclear electricity generation.
Is this true? RWE recently announced CfD (https://www.rwe.com/en/press/rwe-ag/2024-09-03-rwe-welcomes-success-for-onshore-and-solar-projects-in-latest-uk-cfd-auction/) in the UK for about 50£/Mwh. By comparison, average electricity costs in the manufacturing sector in the UK April to June this year (https://assets.publishing.service.gov.uk/media/66f3cb577e1625ee0d0c6f0b/Quarterly-Energy-Prices-September-2024.pdf) were 18.23 pence/KWh (or 182£/MWh).
EDIT: This isn't a perfect comparison since the nature of CfDs is you pay the generator even when you don't need the electricity, which means the actual price you're paying for the electricity you use is higher. You're also paying for a lot of electricity you don't use depending on the make up of the grid. See my last paragraph for how this can work like a subsidy if you build too many renewables without sufficient storage.
The nature of electricity markets in the UK (and most/all? of the developed world) means gas will almost always be the most expensive source of electricity when purchased since it has a high marginal cost (the price of gas) so generators won't produce until there is demand that can't be met by other generators. Sources like solar have negligible marginal costs of production and will outbid gas generators as long as they have capacity available. This means gas tends to set the price of electricity during high demand periods and tends to set that price much higher than when zero-marginal cost sources have sufficient capacity to meet all demand (in the case of excess capacity, you even get negative prices which have begun appearing in the UK with some frequency).
In theory, CfDs are not subsidies, they allow the government/grid operator to pay less during periods of high prices but pay more during periods of low prices. The UK saved lots of money due to CfDs when prices skyrocketed after Russian invasion of Ukraine (https://www.current-news.co.uk/wind-farms-to-pay-back-660m-under-cfd-scheme-amid-high-gas-prices/).
But CfDs look kinda like subsidies under current policy which is supporting renewable construction for climate reasons. At a certain point, you've used CfDs to support overbuilding the renewable system hugely (to deal with periods of limited sunshine/wind and high demand with limited storage capacity) and have lots of excess capacity/curtailment. CfDs will almost never save a grid operator money at that point while forcing up rates for customers in the long-term.
Levelized cost of electricity (LCOE) is basically what you have to pay someone per-MWh to get them to build a project. It's the best we have to compare the costs of generating electricity from gas, solar, wind etc where the costs come at different times in the project lifecycle: for solar and wind nearly all the cost is that of building the plant (capex) while for gas most of the cost is the fuel. The main inputs are capex, opex and the cost of capital.
Since LCOE is a theoretical calculation, it makes sense to publish "without subsidies".
Average global LCOEs aren't particularly illuminating because, for example, the same solar panel will generate twice as much electricity in southern California as in the UK. However, it is true they have come down everywhere, mainly because solar capex has fallen drastically. In markets like the UK where solar panel costs aren't affected by import tariffs, a Watt of solar panel costs about 10 US cents today, compared with over $5 in 2005.
However, even in the UK (which is not very sunny) the Contract for Difference prices in the last round of auction were slightly lower than for wind:
https://www.gov.uk/government/publications/contracts-for-difference-cfd-allocation-round-6-results/contracts-for-difference-cfd-allocation-round-6-results-accessible-webpage
(Personally I think this is not great, because the UK needs more power in the winter - it's a highly seasonal climate. But it does also have a lot of wind).
LCOE is particularly poor for judging the cost of renewable power because there are system level costs associated with it that do not exist for traditional generators. Mainly, solar and wind are not available on demand so additional backup capacity or storage is needed to support them (there are other minor costs as well, a more distributed grid which solar tends towards has higher interconnection costs for instance). You should assume the actual costs of solar/wind are anything from slightly higher than LCOE to much higher depending on the grid. At the moment, this still makes them the most economic option for additional power generation in many places (hence the solar boom). Though as more solar is added, more backup is needed and system costs increase.
Batteries do a lot to solve this problem and costs are coming down. I don't full understand why we haven't seen wider deployment of grid-scale batteries yet given the cost data I've seen and I suspect it's a lot of it is inertia. But batteries don't completely solve system costs. The current economics of batteries make storage for 1-2 weeks (think cloudy, no wind and cold in the winter for several days in a row) very expensive. Without batteries, an entire separate generator system needs to be maintained as a backup for these situations which is a huge system cost for renewables.
This seems like a solvable problem to me, although slightly more nuclear would help most grids. In the meantime, 60-70% renewable grids (excluding hydropower) seem very achievable in most places but getting much higher than that could be very expensive.
>Batteries do a lot to solve this problem and costs are coming down. I don't full understand why we haven't seen wider deployment of grid-scale batteries yet given the cost data I've seen and I suspect it's a lot of it is inertia. But batteries don't completely solve system costs. The current economics of batteries make storage for 1-2 weeks (think cloudy, no wind and cold in the winter for several days in a row) very expensive.
Any suggestions on where to look for the cost of solar + batteries, as a function of how long the battery backup is good for? I tried doing a BOTE calculation, and it looked like solar + 12 hours backup for night approximately doubled the cost, with a 10% interest rate for the capital cost of the batteries - but this was really crude. Are there good numbers available for this?
Cost =/= price.
Prices are determined by local social choices - policy, regulation, politics, etc.
Costs are a function of the global economy.
If prices are divorced from costs, you have market failure in your society. Take that up with your elected officials.
In Massachusetts I regularly got offers in the mail to switch power providers to 100% renewable sources. The quoted rate per kWh was often lower than the default National Grid rate. Solar, unsubsidized, as produced, does in fact have the lowest cost of any power in the world in most places. This is not the same as the price charged to residential customers of a utility.
Often the problems here in practice include the fact that grid operators need to provide dispatchable power on demand, and right now the cost of "firming" solar power supply is very high in many places. There are quite a lot of rabbit holes and quagmires of technology and economics and infrastructure and policy to go down to untangle that mess, and they vary a lot by location.
I also get those offers, and have been told that the gotcha is that the quoted lower price is subject to change at any time, whereas National Grid rates are guaranteed stable. OTOH, can confirm that unsubsidized solar is cheaper than National Grid, based on the solar panels on our house.
Well, in my experience from 2012-2021, those rates were not any less stable than National Grid's rates, and when they did change, it was very easy to switch providers a second time.
Many Thanks! Glad to hear real data instead of hearsay.
I do not know the law in UK well enough to explain it, but I have listened to enough people complain about the laws in the UK to know that the energy grid is subject to extremely weird and bad regulations, and nothing that happens there is anything like a natural price.
Levelized Cost of Energy is a highly imperfect metric that obscures as much as it reveals.
What is needed are comparisons of total network cost of making reliable 24/7/365 electricity ubiquitously available. LCoE does not enable such a comparison.
https://www.dbresearch.de/PROD/RPS_EN-PROD/PROD0000000000528292/Costs_of_electricity_generation%3A_System_costs_matt.pdf?undefined&realload=gE~0QhqwDhR~NOPIw6AsQMD42h5HELCepAe8l7PyxtrKtFJP69xPEZ6HsXVzj1Zf
Solved the Twin-Nuclei problem and then post Bretton-woods. I think Eric Weinstein coined the term EGO (Embedded Growth Obligations), but I think that sums it up pretty well.
> somebody told me that the San Francisco legs of the BART - the Bay Area’s light rail, infamous for being noisy, dirty, and violent - had become comparatively safe and clean over the past few months, after the city installed fare gates that actually worked and couldn’t trivially be jumped over. Apparently the people ruining the BART for everyone weren’t even paying the fare. I always would have guessed there was a correlation between bad behavior and nonpayment, but am surprised at exactly how high the correlation has turned out to be
Psst! Don't look now, but California is about 2 steps away from rediscovering the principle of Broken Windows Policing. And they think it's a new thing they just came up with!
Also, couldn't the *Grants Pass* Supreme Court ruling have had something to do with it? I've definitely noticed a serious improvement in Philly over this past summer.
Perhaps, but for the moment it doesn't seem to be having too much of an effect yet, even in Grant's Pass, for various reasons: https://www.nationalreview.com/news/grants-pass-fought-to-the-supreme-court-to-clear-public-camps-so-why-are-city-parks-still-filled-with-homeless/
(Disclaimer: this article is 3 months old. Conditions could have changed since then.)
Reading the article, it seems like Grants Pass is having continued problems due to restrictions specific to Oregon and the 9th Circuit. That's really frustrating.
Meanwhile, on the east coast, I noticed improvements within a month of the decision. The Philly ACX meetups are in the old town/tourist-y part of the city near Independence Hall. There used to be lots of homeless sleeping on the sidewalk and panhandlers all over certain intersections. Now the sidewalks are all clear. There's one spot coming off the Ben Franklin bridge that still has panhandlers, but it's usually just 1-2 guys who are only there during the day, instead of an entire tent encampment under the overpass.
Downtown Atlanta was similarly cleaned up when I visited over Labor Day weekend.
(Side note - there are no apostrophes in US place names. Places like Harpers Ferry, Pikes Peak, Grants Pass, etc. are all spelled without them. There's an act of Congress banning them.)
That's good to hear. I used to live in the general area, and have been to Philly many times. It's encouraging to hear they're improving things out there.
The ambiance is improving for some, but the homeless people are still there. I believe in progress deeply, and I fully believe the Grants Pass decision is the opposite of progress. It's like sweeping guts under the rug and calling it a day.
Unfortunately, sometimes that's all you can do when too many people who believe too hard in "progress" make actual solutions impossible. The first and most significant problem by far in dealing with "homelessness" is Progressives cynically redefining a problem that has nothing to do with housing as "homelessness" so that they don't have to face up to the consequences of their failed policies in the areas of drugs and mental illness.
“We all want progress. But progress means getting nearer to the place where you want to be. And if you have taken a wrong turning then to go forward does not get you any nearer. If you are on the wrong road progress means doing an about-turn and walking back to the right road and in that case the man who turns back soonest is the most progressive man. There is nothing progressive about being pig-headed and refusing to admit a mistake. And I think if you look at the present state of the world it's pretty plain that humanity has been making some big mistake. We're on the wrong road. And if that is so we must go back. Going back is the quickest way on.”
― C.S. Lewis
Which part of Downtown Atlanta did you go to? I haven't seen much difference, things are arguably a bit worse in some areas.
I was in the core part of downtown for DragonCon - the Hyatt, Marriott, and Hilton. Andrew Young Blvd and Peachtree Street. We stayed up in Midtown and took MARTA back and forth every day.
There were some wandering homeless/mentally ill guys without their stuff around MARTA, but way fewer panhandlers or campers on the sidewalks than in past years. I lived in Atlanta for college from 2013-17, so I'm calibrated to those years. It seemed about steady for the last ten years, and then dropped off a lot this time around. I wasn't accosted by a single panhandler this year.
But I was only there for a few days, so maybe it's different in other parts of the city/different times of year. Maybe DragonCon beefed up security outside on the streets this year?
I was very amused by all the ads on MARTA this year telling people to move to Philly. I saw those and was like, yup! Philly's great!
Aha. Thank you - I live in the East Atlanta area and don't go down to the Central Business district that much. East Atlanta has seen a marked increase (and so have interstate areas from what I've seen). Greater enforcement would explain the shift.
The government can pry my apostrophes from my cold, dead hands!
Won't someone please think of the greengrocer's?
Philly was not effected by the Grants Pass decision, so any improvement is unrelated. The ruling that the Supreme Court overruled in Grants Pass was for the 9th Circuit which covers the western US, and it never affected the east coast.
Frankly, that Grants Pass supreme court case was so disgusting that is still baffles me that some effective altruists like it. It solved nothing and only increases pain and suffering. The homeless people it affects are being offered no services, just a temporary jail stay or even just a violent eviction. It's effectively social murder. As a society we don't want to actually address homelessness because it would be expensive, but to make the Grants Pass decision make sense morally, you have to actually address homelessness.
I agree with most of this, but not the part about it solving nothing - I think that criminalising homelessness and driving all the homeless people to move to other towns probably will solve the problem of "we have lots of homeless people in this town, lowering quality of live for others".
Good objections to criminalising homelessness in order to force homeless people to move to other towns include 1) it's unethical, 2) it makes life worse for the inhabitants of those other towns, and 3) it doesn't scale, because if everyone does it you're back to square one except everything is worse.
But if your goal is purely "solve my town's homelessness problem here and now, without worrying about any of those things", I think it's probably a really effective way of achieving it.
One of the ways we address homelessness is by criminalizing homeless encampments. Most of these people are dug addicts and will slowly kill themselves unless something stops them first. Allowing these people to sleep where they want and die slowly in peace is not compassion, and making that not a live option for them is. The Grants Pass repeal will lower homelessness and help people by forcing them into the recovery and support programs we already provide, and that addicts aren’t interested in because they’d have to get clean. Well, get clean, sleep in a ditch, or go to jail: setting up a shack next to the playground at the park is no longer an option.
+1
I think the broken windows version of this would be attempting to arrest everyone who peed on a bart and making the punishment relatively harsh. I think preventing the peeing in the first place is an importantly different thing. I'd be interested in hearing if anyone involved in this decision even believes the broken window thing that doing this one small thing will have larger knock on effects.
In the original Broken Windows program in New York City, turnstile-jumping/fare-beating in the subways was one of the things they focused on cracking down on, with significant effects on downstream crime reduction.
Right, but my point is that arresting people who do it (presumably how NYC dealt with it) is very different than preventing it from happening.
Un-jumpable fare gates is equivalent to removing the fare-jumpers from the population of people riding BART. Arresting the fare-jumpers would be removing them from public society. So right now only BART gets the privilege of a social environment with a significantly-reduced jackass population.
It also lowers their mobility, which may help keep them out of nice areas.
>So right now only BART gets the privilege of a social environment with a significantly-reduced jackass population.
There has to be a thermodynamics joke about semi-permeable membranes, impermeable only to jackasses, in here somewhere... :-)
Semi-permeable membranes are very important in the study of transport phenomena.
The policy is literally (maxwell's) demonic.
I don't think it is equivalent. There's a set of people who will jump when they can, and that set is also more likely to commit other crimes. But they're capable of paying a fare, just like they're capable of using a taxi/uber if they actually need to.
I don't think "fare-jumper" is necessarily a persistent, intrinsic property of individuals. Lots of things work more by perceived value, or back-filling excuses for decisions made on some entirely different basis.
If you're able and willing to ride for free, that sets a tone, a social context, where everyone who DID pay the fare must ipso facto be a self-sabotaging coward. Why, then, play along with any of the rest of society's rules?
On the other hand, if jumping is inconvenient enough that it's easier to pay, riding the BART becomes a privilege you've willingly paid for, so the space itself must be worth something, and thus ought to be treated with respect.
That sort of context-sensitive morality is more extreme when drunk.
The real success was that NYC discovered if you arrest fare jumpers, a lot of them already have outstanding warrants (petty theft, assault, etc). So they'd get a big unit of cops together with a paddy wagon and would sweep through the subway, arresting every fare jumper they found. Then they'd process them right there at the paddy wagon (they brought all the necessary paperwork and equipment for processing with them) and everyone they had with an outstanding warrant would go the jail and the rest would get a ticket for fare-jumping and be released. It really helped clean up the streets.
Now imagine having an ICE van there as well.
Its that thing about how arresting a tiny percentage of people would end an absurdly high percentage of crime. People who flagrantly disregard the minor laws tend to be bad people who also flagrantly disregard the more important laws.
I don't think that's actually a good example of "broken windows" because it leaves no visible effect. Littering, on the other hand, does. Arresting jumpers is a good way to grab a lot of people who turn out to be carrying illegal items and have other charges against them though (similar to traffic stops).
I think it does have a visual impact, fare jumpers are noticeable when they literally jump the turnstile and in this case they are clearly causing problems on the mass transit and removing them from it ensures a better environment.
Fare jumpers are phenotypically different as a class on average from non fare jumpers. Based on just how someone looks you can come up with a decent idea if they fare jumped. Therefore removing fare jumpers changes the average phenotype of a BART rider.
It does indeed appear to have changed the composition of riders for the better.
I am shocked, shocked to hear that the drug-addicted, mentally ill homeless guys on the BART did not pay the fare.
I thought Giuliani-style broken windows policing is basically discredited?
https://en.wikipedia.org/wiki/Broken_windows_theory#New_York_City
Not really, no. What we see in the Wikipedia article is commonly known as "lying with statistics," particularly the last bit, where it says:
> A 2017 study found that when the New York Police Department (NYPD) stopped aggressively enforcing minor legal statutes in late 2014 and early 2015 that civilian complaints of three major crimes (burglary, felony assault, and grand larceny) decreased (slightly with large error bars) during and shortly after sharp reductions in proactive policing. There was no statistically significant effect on other major crimes such as murder, rape, robbery, or grand theft auto.
You have to be careful here or you might miss the trick. First, there's the big obvious one: the decrease in reports of crimes is "slight[] with large error bars." But then there's the much more subtle one: they're looking at the rate of *civilian complaints of* crimes, not the rate of crimes being committed. And keep in mind that this was happening in 2014/2015, at the height of Obama-era anti-police pressure tactics!
It turns out that, when people have little reason to believe that calling the police to report a crime will do them any good, they're far less likely to call the police and report the crime. (See also the broad divergence under the Biden/Harris administration between the FBI crime statistics, which are based on police reports, and National Crime Victimization Survey numbers, based on asking people if they've been the victim of a crime.)
The only crime statistics that can be taken at anything close to face value are homicide and auto theft (including smash-and-grab of stuff in automobiles). Murder, because basically nobody tries to hide the bodies, literally or statistically speaking. Auto theft, because even if you know the police aren't going to do anything you need the police report to file an insurance claim.
Everything else, reporting bias makes the error bars uselessly large. So if someone is counting "burglary, felony assault, and grand larceny", but not murder and not auto theft except as buried in the larceny numbers, yeah, suspect statistical skulduggery.
I think vibes beat statistics in this particular case. New York in the 1980s had a reputation as crime-ridden -- I remember a lot of Mad Magazine jokes about the impossibility of crossing Central Park without getting mugged. By the 2000s it was had a reputation of being remarkably safe by US big city standards.
The rest of the US followed the same pattern, so it's hard to tell if NYC-specific polices had any effect.
NYC cut its murder rate by a factor of 4 while the rest of the country cut it by a factor of 2. No, it is not at all hard to see the effect.
And then there was a called shot: Los Angeles hired Bratton and cut its murder rate by a factor of 2 when the rest of the country was stable.
Depends entirely on who you ask. As far as I can tell, there's still a substantial minority of people who believe in it, but they're almost exclusively conservatives who believe other things that make me distrust their judgement.
Fixing literal broken windows ("nuisance abatement") is the ONLY criminology intervention replicated in RCTs https://entitledtoanopinion.wordpress.com/2018/04/28/uncontrolled/
I would like the BART gate thing to be true, but I question it -- the gates are really new, only 8 stations on the entire BART system have them so far -- and one of those 8 is at the airport, which is probably not a major location for sketchy people to board. In San Francisco, only 3 BART stations have these new gates. One is Civic Center (which is less than half a mile from Powell, which doesn't have the gates) and another one is 24th and Mission (which is less than a mile from 16th and Mission, which only got the gates as of October 12th). It seems surprising that the gates could be having such a dramatic impact on the BART this early in the installation process.
Oh that you had returned to the Gender Studies comparison because in fact gender studies has been progress studies. Progress studies should include how gender has progressed, including (not to wade into anything) gender technologies.
There’s a lot of gender technologies! Everything from stiletto heels to mom jeans to the manbun to testosterone injectables to surgery is an invention that enables people to present their gender in new ways.
Alice Evans would be an example of someone who studies both gender and progress. https://www.ggd.world/
There's something missing here, perhaps because it is difficult to quantify: at what point do you have enough?
It is clear that, up to a certain point, economic growth in the US went hand-in-hand with progress at a more concrete and human level - becoming "richer, healthier, safer, and better-educated than our ancestors", as Scott put it. Scratch that - it's not as if it were just a coincidence - we can probably agree that the former was a prerequisite and the main engine of the latter.
Ditto in the developing world today, *up to a certain point*: you need a certain level of production if you want a middle-class society. No, it's not a sufficient condition, and yes, GDP growth can be very deceptive, particularly in extractivist economies (high GDP growth can mean that resources are being extracted at an accelerated rate by foreign company X, which got an amazing deal from subornable and/or naïve officials, and only has to pay a pittance in royalties). Nevertheless development remains a crucial part of, um, development.
But do most problems in the US really have as their necessary root a slowing-down in productivity? GDP per capita in the US is now 50% larger than it was in the mid-90s. (Note: yes, in constant dollars, obvs.) Are people really that better off? Life expectancy has barely budged. Are people much better educated than then? There are some positive trends (some cities are safer) but their relation to the rest is less than clear. Now, if a wizard were to replace the GDP per capita of our days with that from the 90s, *perhaps* the immediate result would be immense suffering, but not because we were miserable in the 90s; this is what 'growth addiction' would mean, I suppose.
There is some point (probably well before the early 70s) at which people in the US could sensibly talk about an "affluent society"; that's actually the title of a book by Galbraith from '58. That really meant a society were more than half of the population were not in any obvious material need, and their main problems were of other sorts. Jacobs/Nader (who did make us safer)/Carson (whose concerns were valid even if she was wrong on many details) can be sensibly seen as a product of those times. Whether they had any sort of macro impact (for good or for the dastardly forces of anti-progress evil) is much more doubtful, at least outside the very particular fields that they cultivated.
Looking at the big trends is useful and necessary. Looking only at productivity figures alone seems reductive and potentially self-deceptive.
(On this last note: it would be interesting to have Emmanuel Todd's books come up in the annual Book Review here. I would get into that myself, except I don't feel particularly qualified in any way - I am not a demographer, or an anthropologist, or a historian, or for that matter an economist. Of course his stuff is often rather controversial and has to be taken with a grain of salt. Still, it's a good example of big-picture analysis based on a close look at quantitative data that does not stop at the level of dollars.)
> GDP per capita in the US is now 50% larger than it was in the mid-90s. Are people really that better off?
Is that figure before or after accounting for inflation? Because if it's before, then we're significantly *worse* off: https://www.officialdata.org/1995-dollars-in-2024?amount=100
After accounting for inflation.
My guess (haven't looked into it) is that some of the GDP per capita gains have gone to the top 1%, some of the rest have been eaten by higher housing costs, and some of the rest have been eaten by stricter regulation (or, more optimistically, been spent on safety rather than consumption). I think of the remainder - maybe 20% of the total - some of it *has* gone to things like bigger houses, more cars, more vacations, eating out more, getting DoorDash, and other things that do make us happier.
If somebody were to ask "what are the main things that make Americans unhappy?", how plausible would the following answer sound: "their houses are too small, they have an insufficient number of cars and they don't eat out enough"?
(An outsider wouldn't answer that, and an American wouldn't either. I know, someone will argue "but the fact that people spend money on that shows that's what their priorities lie", but no, really.)
Vacations in the US remain very short relative to the rest of the developed world (only Canada can compete in vacation shortness) - how much has that changed in the last 30 or even 60 years? The only thing in that direction I can think of is that there may be more people in the jig economy, but that's often not because they want to be.
I agree rich Americans probably don't worry about these things. But poor people are much less happy than rich people, and it seems like most problems that poor people have but rich people don't are problems with money.
(I agree there's a sort of mystery here, which https://www.lesswrong.com/posts/fPvssZk3AoDzXwfwJ/universal-basic-income-and-poverty gets at better than anything else I've seen, but I still think that all else being equal more money has to be good for the poor)
Well, I'm neither rich nor poor (nor for that matter a US citizen, though I lived there for ten years), but my impression from living in non-rich areas of the US (mainly intercoastal; I've seen more than my share of square-shaped states) was that relatively few people in the US are short of *stuff*, as in, stuff you buy at the shop. Some have trouble getting to the end of the month (that's probably not very closely correlated with income level if you exclude the very top and the very bottom), but that's not quite the same thing.
Money does obviously help with access to many things that are in short supply - health care, a good education, housing. In fact what really helps there is having a lot of money. Is the root cause of that really that there's not enough money to go around? Or that not enough stuff is being produced?
I'm not sure what distinction you're drawing. Yes, there should be a good monetary policy where the Federal Reserve prints the right amount of dollars to denominate the amount of wealth, but the basic and harder problem is having enough goods that they're very cheap.
Let's get the money supply question out of the way (that's an irrelevant issue that human language has somehow invited - though of that as I was about to press Enter). What I dispute is the last bit: that the basic, hard problem, *in general* is having enough goods that they are very cheap. That may make sense at certain stages in development, but it seems countersensical in a developed economy, and even more so in the US in particular.
When it comes to relatively new technology, particularly when it solves a clearly demonstrated societal need, then I agree with you: obviously we need more and cheaper solar panels (keeping an eye out for their environmental cost, which is real, if much smaller than their benefits) and ditto for batteries (there the issue of environmental cost vs. benefit is less clear; if you disagree, I'd be thankful if you could convince me - I have actually just had solar panels installed and I am trying to decide whether I ought to get batteries). And yes, this is possible because we have such a thing as an industrial economy, mass-production, etc.
It's the claim that the slowing down in the rate of growth or qualitative change in the "world of atoms" was "one of the greatest historical tragedies" that has me unconvinced. At least it's not a claim that makes itself.
That's impossible with Veblen goods https://josephheath.substack.com/p/key-stages-in-the-decline-of-academic and when "good housing" is a matter of living in a "good neighborhood" with "good schools" (a function of the parents of the students), that's also a Veblen good.
Poor people and rich people are different sets of people. As an example, the upper class now work more hours, contrary to Eliezer's hypothetical of Anoxistan: https://www.visualcapitalist.com/cp/charted-actual-working-hours-of-different-income-levels/
Having been poor, in the absolute poverty sense of living in a pre-electrified semi-subsistence farming community where there was no municipal water and three towns shared one phone line and then later in the American sense of having way more money than 70% of humanity but less that most other americans, I much preferred the former (unfortunately, modernity has reached that part of Costa Rica, and now they have wage work, cellphones, depression and diabetes).
The thing that was really awful about being poor in the US was how precarious life was, that your shitty dodge you bought used because you couldn't afford a Toyota could blow a distributor cap (I had three rattling around in the trunk), you would miss work, you would lose your job, you would lose your house, and you would be on the street three weeks later.
Now that I have a LOT of money, I still live like I did when I was poor (other than owning my house) but with more big vacations. The extra material possessions and services (to me!) are hedonic treadmill net negative distractions from what actually makes me happy: eliminating stress, and work that isn't completely alienated in the Marxian sense, and novel experiences.
I know I am much less wigged out than my friends who have bigger houses and nicer cars by having a bullshit MBA type job, but have to worry about losing them and dealing with the existential ennui.
>"their houses are too small, they have an insufficient number of cars and they don't eat out enough"?
Replace 'their house is too small' with 'they can't afford a house and their rent eats 50% of their paycheck and their apartment sucks and the landlord never fixes anything,' and yes, this basically describes the gripes of a large number of my working-poor friends.
Right, but that's a different thing.
Different how? They want shelter of a particular quantity and quality, but aren't getting it, despite spending as much as they feasibly can.
"If somebody were to ask "what are the main things that make Americans unhappy?", how plausible would the following answer sound: "their houses are too small, they have an insufficient number of cars and they don't eat out enough"?"
Those are real problems for poor people in America who lack those things.
I'd say that a comparison to other developed countries makes it clear that those are rarely problems if the above is taken literally - or rather what Americans have can be a problem *because of the way much of America is set up*.
- I have not yet seen a truly small American house, but of course many Americans don't live in houses. (Note: for my first four years in [major European city], I lived in 23m^2, and that was considered normal for a single academic; I had lived in much less space while in college and grad school in the states.) Compared to just about any place in Europe, American houses are not small - they are, however, impermanent (not just compared to most places in Europe, but even to most non-slum housing stock in Latin America), though that's a preference that *may* make some sense given how cheap lumber is and how much people move. Much more to the point: there is a real problem, in that access to house ownership has become much more difficult, and being a renter sucks much more than in e.g. Germany (where, to boot, places are *usually* better built and insulated, though not always).
- It seems bizarre to think that Americans don't have enough cars (what is next - their cars are too small?). What *is* the case is that the working poor sometimes have a car in good condition and live in a place in poor condition, even if they own it - they depend entirely on cars to work, in that public transportation is pretty terrible outside a couple of major urban centers (indeed, it's less useful than the chaotic, polluting, unregulated 'public' transportation (really madmen with vans) that you find in many Latin American cities - no, I don't mean Mexico City, which has a good metro).
- Eating out often is not any sort of basic need - growing up, we'd eat out once a year, after we got and showed our report cards . More of a problem in the US: a lack of corner shops (you can't run out and across the street to get a clove of garlic), plenty of people get half-hour lunch breaks (they are lucky if there is a McDonald's nearby), etc. (In France, your place of work must either have a cafeteria or give you a 'ticket restaurant' valid at a restaurant. That, together with sufficient lunch breaks, hasn't been invented in the US yet.) Also a reason why eating out feels like a basic need: a lack of third spaces which are either free or close to free. Social life *is* a need.
There is a small percentage of Americans that do suffer from not having enough in a gross material sense - people who are poorly nourished, homeless, etc. It's shameful that anybody is in that condition. There's a much larger percentage of people who have very real problems - but they are not at the level of a gross lack, though it may look at first that way, looking from inside the US, without outside comparisons. Making more stuff available more cheaply at Costco is not going to solve their problems - and these problems haven't become less serious since the 90s, when the US was 2/3 as rich in the sense of GDP per capita; some seem to have become more serious. It's less obvious, then, that the source of these problems is that American affluence has increased more slowly since the 70s.
The American preference for larger houses is because Americans are richer. Rich people in Europe live in bigger houses too.
Lots of people would be helped by America getting richer. The family that moves out of the trailer park and into a house. The 19-year-old who can move out of momma's basement and buy a car.
"The American preference for larger houses is because Americans are richer."
Staggeringly wrong take.
More Americans than Europeans are (a) able to (b) interested in imitating the lives of the rich in some ways, at least in the sense that a stage setting imitates a house or a landscape. Hence - nothing follows.
Again, GDP per capita in the US has increased (correcting for inflation) 50% since the mid-90s and has doubled since the mid-80s (to give two points of reference at which many readers of this blog were already alive). Are there really many fewer people who live in trailer parks and worse? (I'm actually asking for statistics.) What I would believe more readily is that some trailers have got larger, maybe even 50% larger, but that doesn't solve things, does it.
And yes, Scott is almost certainly right that part of the issue is new wealth being sucked up by the very top, NIMBYism, etc. What is completely unclear to me is that the solution is for the country to become richer faster, in the sense of real GDP per capita.
There is probably a stage of development at which the most important thing is arguably for the country to develop quickly - the gains obtain thereby overwhelm everything else. Developed countries, and the US in particular, would seem to be well past that stage, almost by definition.
re American houses being impermanent. I was just thinking the other day that the old morality tale of the Three Little Pigs seemed to have gone over their heads. Climate change is increasing the number of big bad wolves...
Well, improving the energy efficiency of an existing 'permanent' house can be expensive (ask me); I suppose that's one good potential side of a place where people consider a 30-year-old house to be "old" - when they build, they could build to contemporary standards. It's more that having to build from scratch every 30 to 50 years (numbers I just made up) is not in itself eco-friendly.
Yes, without a doubt, houses could be either bigger or better furnished for ~99% of Americans. Look at the furnishings in the $10M California houses: any house that's not furnished as nicely as those houses could use an improvement. Even if you live on dirt cheap land in the middle of nowhere and only need ~1000 square feet/person, you'd still need to spend at least ~$1M on furnishings/renovations alone to get the "perfect house".
As one example: if your house doesn't have a heated pool where the water can stay at 85F even in the middle of winter, it could still use some improvement.
"Better" is subjective. It may be that to the owner of the $10M house, everything in it is necessary, or could even be improved. And the owner of a $50,000 1000 square foot house could have everything they want, and adding, for example, a pool would detract from the value for that person, either by making the walk around the pool longer to get somewhere (for an indoor pool) or cause a wildlife hazard (for an outdoor pool), or something else.
Keeping up with the Jones's is a fool's game. One needs to figure out what is best for oneself.
> "their houses are too small, they have an insufficient number of cars and they don't eat out enough"
I don't know that I would identify these as the main things, but I am already doing pretty well and yet I would still enjoy living in a bigger home, having a nicer car, and eating more luxurious food more often. Living, eating, and transportation are *huge* parts of everyone's lives, and making them better will increase your QoL for a very long time.
Would you be ok with 30% smaller house / apartment if it would have reduced your (and your family) daily commute by 2 hours (made-up numbers)?
If so, then won't it look like 'we need smaller houses to be happier'?
I mean, that's exactly the sort of tradeoff (with less made up numbers) that a lot of people make these days. For the same price, some people choose to live in smaller dwellings closer to the city centre, and others choose to live in big sprawling acreages on the edge of town.
There's no right or wrong answers to this tradeoff, just a matter of personal taste, and the important thing is to ensure that housing is being built to suit everyone's taste.
For me personally, I wouldn't accept a 30% smaller house closer to the city (but then again I don't commute).
There are many margins at which I would accept many tradeoffs! That's totally orthogonal to my point, which is that I think having bigger houses, more and better cars, and more and better food _would in fact_ make Americans happier. If you want to talk cost, that's a different conversation. I'm just objecting to parent comment's dismissal of the idea that lack of big houses, fast cars, and fancy food is keeping us unhappy. Those aren't the *only* things that keep people from sublime happiness. But they're good things to have, and I want everyone to have more of them.
*More* food? You are making my point.
(There are issues with how the poor eat in the US and with what they can eat, but a lack of calories is not one of them, not outside the very, very bottom.)
It may be that many people who are far from thin may want more food, in the sense that they are addicted to it. If their income starts being a limitation at some point, that is not a tragedy.
"how plausible would the following answer sound: "their houses are too small, they have an insufficient number of cars and they don't eat out enough"?"
Very plausible? At least for Americans who are not rich? I can recommend this very impressive article from residentcontrarian:
https://residentcontrarian.substack.com/p/on-the-experience-of-being-poor-ish
Two excerpt:
"I am always consistently shocked by how little people living at a decent-to-great income level fear their cars; it seems like they hardly lie awake at night thinking about their iffy alternator much, if at all."
"Do I spend 50% more on an apartment and stress my budget to death, or make my wife and kids in a place where I have 5 front doors within 15 feet of ours in the murder-iest part of town? "
I've addressed that point below. Neither of these would be a problem if the US had decent public transportation and cities were less murdericious. More cars and fewer front doors are not going to solve the real problems (even if having more money does help an individual).
It would be easy to build good public transportation and fund better policing if we had more money as a society.
Would it? Plenty of societies have built better public transportation systems while having a fraction of the US's GDP per capita. Ditto for murder rates (better policing is only part of the story, obviously) - take a look at https://en.wikipedia.org/wiki/List_of_countries_by_intentional_homicide_rate .
Given that, is it clear that things will be better once the US has twice the GDP it has now?
> If somebody were to ask "what are the main things that make Americans unhappy?", how plausible would the following answer sound: "their houses are too small, they have an insufficient number of cars and they don't eat out enough"?
If we look at revealed preferences, then yes. People who get extra money do spend it on improving their houses and cars (let's leave the eating out aside as more of a personal taste.)
I do think that people also have a lot more *stuff* than they did in the 1990s though. TVs are bigger, kids have more toys, everyone has another dozen appliances in their kitchen, we fly around a lot more, etc.
Separate point: no, that does not tell you what would make people unhappy or happy, if scaled up to the level of a society; in fact, it barely tells you something about what they prefer, and that only at the level of things they can afford.
(a) Yes, the US is car-dependent - most of the working poor need a reliable car.
(b) What people need is things that can in fact be bought (in the US, where they are for sale), but at a completely different price scale: health care, good housing that they either own or that remains affordable and is kept up to a high standard (not: bigger houses), not living in a neighborhood with crime rates five or ten times what is typical for a developed country, etc.
Of course advertising is very good at making people want what they don't need, but one would think most people have smarted up by now.
>I do think that people also have a lot more *stuff* than they did in the >1990s though. TVs are bigger, kids have more toys, everyone has another >dozen appliances in their kitchen, we fly around a lot more, etc.
And what good has that done?
Advertising doesn't work that way https://meltingasphalt.com/ads-dont-work-that-way/ It's impossible for people to "smarted up" so the signalling doesn't work.
Here in Denver, a safe neighborhood in Superior with large, new houses is cheaper than a safe neighborhood in Louisville with crappy old small houses and a vibrant little downtown with lots of families and kids bustling around.
I think people do prefer larger houses over smaller ones, but their main preferences are for a safe neighborhood with good neighbors, a good community, a reasonable commute and interesting things to walk to. Larger houses tend to get built in new developments for practical reasons, and those developments don't usually have any nearby shops to walk to, but "lack of shops and restaurants" is not a revealed preference. It has more to do with regulations.
s/"house is too small"/"apartment is too expensive and an unreasonable commuting distance from work", and you're getting a lot closer.
Those are both manifestations of housing costs, but they affect different populations differently. You've got one large group that's happy living in the houses they bought decades ago (raises hand), another large group that despairs of ever owning a house and are dissatisfied with their expensive, poorly-situated homes, and some smaller groups that like their apartments or dislike their houses.
Higher GDP can be converted to more vacations and the same amount of stuff, if a society wants. The fact that we all keep working the same level of hard to get MORE stuff instead is a revealed preference, no?
Or perhaps it's a deeper societal ill around coveting what others have ... but that's orthogonal to GDP IMO.
A few other things can also cause big divergence between measured GDP per capita and real utility/well-being.
First, and simplest, is hours worked. Increasing labor force participation, longer work weeks, later retirement; all of these will increase measured GDP per capita but at the cost of both leisure and unmeasured home production, so the welfare gains are not as high as you would think from the economic statistics. (But while we’re specifically comparing the US in the 90s to today, this isn’t a big part of the answer.)
Second, harder to measure but potentially unbounded in effect, is status externalities. The extent to which your house is “too small” is partly about real consumption, like how much space you actually physically need to move around in comfortably, but also partly about social comparison. If we double everyone’s house size, this won’t double the utility everyone gains from their house, partly because some of the utility from having a “big” house is specifically that other houses are smaller.
Third, and more speculative, are meretricious Needful Things. A dollar spent on Doritos counts the same as any other dollar in the economic statistics. The feeling of regret I get after finishing the entire bag of Doritos is not measured on economic statistics. To the extent that improving technology makes vendors better at selling us Dorito-like goods, this could also create a increasing gap between measured GDP per capita and actual well-being.
These are great points. A labor-hour-adjusted GDP per capita metric is relatively easy to obtain, but it’s significantly more difficult to try and adjust for status or for Dorito-like goods.
>Third, and more speculative, are meretricious Needful Things. A dollar spent on Doritos counts the same as any other dollar in the economic statistics. The feeling of regret I get after finishing the entire bag of Doritos is not measured on economic statistics. To the extent that improving technology makes vendors better at selling us Dorito-like goods, this could also create a increasing gap between measured GDP per capita and actual well-being.
That's fair. There are also analogs in the sorts of things degrowthers push:
An hour spent guilted into a community activity that one would actually rather avoid, for instance.
Economics recognizes decreasing and negative marginal utility. Sure, it can't be in the statistics, because economics also assumes people behave to increase their total utility which, as you illustrate, isn't always true.
Many Thanks! I also agree with one of your previous comments, that:
>Keeping up with the Jones's is a fool's game. One needs to figure out what is best for oneself.
But I don't have a good intuition (let alone data!) on _how much_ of GDP growth is in purchases of items with negative marginal utility to the purchaser, essentially mistakes at the level of individual households.
There is also the problem of what a society could _do_ with such information, even if they had it. In theory, a wise and benevolent regulator could try to discourage such purchases - except that wise and benevolent regulators are as available as the Tooth Fairy and Santa Claus. ( I'm writing from the USA, where we are currently trying to decide between Horrible and Worse for POTUS. )
Median or household income after taxes, rent etc - discretionary income is the real measure.
Certainly the case in the UK that GDP per capita gains have gone to higher housing costs. It's a sellers market of an essential so if you put more money into the system everybody ends up in the same house but they pay more for it. A big part of the rise in house prices has been driven by larger mortgages which take longer to pay off, and then impinges people's saving for retirement.
You might be interested in Easterlin's book on happiness. If your disagreement with Scott in the comments is over whether more wealth -> more happiness, his data suggests you're right that it doesn't.
Thanks for the recommendation - I'll take a look. I think what we are getting at (I'm not sure there will be a disagreement in the end) is something a bit finer than that. I think we can all agree from the start that
(a) it's good for a society to be competent at producing a wide range of things well, quickly and cheaply
(b) the life of a random individual poor person in a rich country would, on the whole and most of the time, be better if they had more money. (On the whole, the less money you have, the more difference a given amount of money makes.)
So the issue is not that (but I'll take a look at Easterlin).
The idea that more money doesn’t make people happier after a low threshold has been thoroughly, thoroughly debunked.
Affective happiness goes quite consistently up with the log of income in the US.
*Unhappiness* does seem to have a threshold where more money doesn’t help, which is one reason why people get confused.
The idea that more money doesn’t help with happiness is, to me, an irritating just-so story folks like to tell themselves to help them feel less jealous of the wealthy.
As it happens, my mother is one such person. She went from middle decile income to receiving an unexpected inheritance.
She is much, much happier. This didn’t surprise me, but it sure surprised her.
In the US, enough money will buy you out of many problems that non-rich people in some other developed countries generally don't have. That does not mean that increasing per capita GDP as rapidly as possible will make those problems go away.
>In the US, enough money will buy you out of many problems that non-rich people in some other developed countries generally don't have.
Could you give two or three examples of what you view the most important such problems are? It would help to make the discussion more concrete. Many Thanks!
Health and education
Something strange seems to have happened. In email I see:
>Health and education
Many Thanks!
Yes, those are indeed problems. In Scott's post about cost disease
https://web.archive.org/web/20200426191456/https://slatestarcodex.com/2017/02/09/considerations-on-cost-disease/
their costs grow much faster than CPI, and much faster than wages.
One common factor to the two: Neither is a straightforward cash-for-goods transaction, like those for groceries or consumer electronics.
Health care is usually financed by insurance, and education by (~90% federal https://educationdata.org/student-loan-debt-statistics) loans
It is certainly the case that, unlike much of the rest of the economy, we _don't_ want to do-what-we-have-been-doing-but-faster.
Education is something of a perfect storm:
- Employer credentialism drives many people to need college degrees for jobs that don't actually make use of them.
- Driving up the fraction of the population funneled through college dumbs down college teaching
- The federal student loans makes students at least less cost conscious about the colleges
- Higher education, not being as cost conscious as the rest of the economy, gets less efficient, most notably via administrative bloat
You should check out Easterlin's book if it's something you have strong views on (or Scott has a post on it on SSC) he has responses to all these objections.
The log(income)->Happiness data is correlational among individuals, Easterlin's claim is specifically that there's no causal link between higher wealth and happiness for the society as a whole, if that makes sense.
It makes sense to me that someone successfully working toward what they want increases their happiness, which would correlate with higher productivity for that individual.
Bad stuff certainly happens to rich individuals.
And there's no doubt in my mind that it's possible to have bad culture combined with relative prosperity, and trading some prosperity for better culture may well be worth it at some margins, for groups, not just for individuals. So I don't think I'm rejecting the thesis that it's possible.
I just don't think it happens in practice to large countries not involved in wars or truly giant natural disasters.
Said another way: holding culture constant, I don't see a case for more wealth doing anything but making people happier.
Regarding Easterlin... I found Betsey Stevens and Wolfers (many) responses to him quite convincing. His statistical analysis just seems shoddy.
"I found Betsey Stevens and Wolfers (many) responses to him quite convincing."
Hadn't seen that one. Giving it a quick look it doesn't seem convincing though.
The first part of the paper is just re-establishing the (income, happiness) correlations between individuals and across countries that are already a part of Easterlin's paradox.
Then they find small correlations with log(income growth) and happiness in a few individual countries that would be relevant evidence against Easterlin but just looks like random noise or a very small effect size at most.
I'm thinking about reviewing Easterlin's book for the next review contest so I'm looking for counter arguments. So far they've all had obvious weaknesses. Is there something more to the paper I'm missing?
"Are people much better educated than then?"
Yes, the statistics are clear on that. If it hasn't made people smarter or happier, that's a problem with higher education. It doesn't mean more wealth is not being expended.
There can be an increase in years of education without an increase in education in a more real sense, even by crude measures (in fact this easy: make it trivial to get a diploma). But no, even at the level of enrollment, the increase was much faster in the early 20th century. See e.g. https://algerie-infos-saoudi.over-blog.com/2017/09/la-stagnation-educative-explique-le-triomphe-des-inegalites-selon-emmanuel-todd.html (this is just an excerpt; I'll try to find a better source readable online).
Yes, we've greatly increased the number of people going to college at the expense of watering down college.
> at what point do you have enough?
Never! Or at least not anywhere we can see from here. I want everyone to get richer and richer at increasing rates and enjoy science-fiction levels of comfort and luxury. I want our current fantasies of a rich life to eventually seem like quaint anecdotes of harder times.
That seems like entirely the wrong approach, particularly for Americans. There are (at least) tens of millions of Americans that have enough material comforts and more than enough tchotchkes, yet are doing badly in many other ways.
Sure, unhappy families are unhappy in their own way. Poverty is only one thing that can make people unhappy, but since it is one and we can address it, let's do.
Just temporary poverty can cause a lot of damage. Unemployment leading to cash-flow problems causes lots of divorces. If people have more money in the bank, they can ride through more storms.
My point is that American poverty is real but that a lack of stuff is not what makes it real. Yes, a financial cushion is nice - and plenty of people in the US who are not poor or would not be considered poor elsewhere have no cushion; compare its savings rate of 4.8% to https://tradingeconomics.com/france/household-saving-rate-eurostat-data.html .
there are zero humans who have enough material comforts by my standards.
it seems you might be holding some sort of context in your head where you are advocating for prioritizing one side of some trade off, in which case you should probably say so explicitly. it is true that at some margins some interventions might be better than simply prioritizing 'everyone gets richer' for addressing some problems.
nonetheless, overall i very much want everyone to get unfathomably richer.
This, um, is so extreme that it makes my point. Actually, this makes me think that there must be a character in myth or fairy tales created to warn others of this particular kind of pitfall, sort of like King Midas, the utility monster, the sandman, etc.
Centuries ago even the wealthiest people were poor by our standards because things we value weren't available at any price back then. Centuries in the future, we will also be poor by the standards of such future people.
So?
There are ways in which a poor person in most developed countries is genuinely better off than a rich Victorian - a chance at getting cured of cancer, say. As for being able to have pineapples in a can, or Internet access: who cares, in comparison? The problem of Victorian squalor was of a different kind. Consider the poor in the Third World nowadays - their squalor does not go away when they get a TV.
No, it just sounds extreme if you don't a) read it carefully and b) take it literally, rather than as flag-waving.
It genuinely is always true that, *all else being equal*, making anyone better off is a good thing.
It's also true that all else often /isn't/ equal - as Alexander Corwin implies, there are usually tradeoffs, and the value of making someone who is already rich even richer, while technically positive, is often so small that practically any negative downside will outweigh it.
But "how rich is rich enough?" isn't really a useful way of thinking, because the answer is "the point at which the sign in the cost/benefit calculation flips depends entirely on what method of making them richer we're discussing them, and what the costs of it are", so there isn't some set level of wealth that answers that question.
Even if you clarify the question to "if you had a magic want that would give someone an extra cent and not change the world otherwise, so the only downsides are inflation and Spirit-Level-type inequality effects, at what point should you stop waving it at them?", I think the answer is that it will be sooner for someone who prioritises consumption than for someone who prioritises investment, and sooner for someone who prioritises investment than for someone who prioritises philanthropy, and for any of them it will depend on the state of the economy they're in.
And what would you propose as the alternative, to resolve those "many other ways" without requiring a net increase in available material resources?
My point is that's the real question that we should be discussing, or at least a real question! A net increase in 'available material resources' (available to whom?) would be nice, but the US has kept on getting wealthier since the early 70s, just more slowly (like every other developed country), and yet in some ways it has genuinely stagnated or got worse (e.g. access to housing).
>GDP per capita in the US is now 50% larger than it was in the mid-90s.
As a function of median income, housing prices in the US are 50% higher than they were in the mid-90s. Source: https://agglomerations.substack.com/p/how-the-next-president-can-solve
That's kinda why YIMBY has become such a big deal now. Regulation has artificially made living so expensive that the extra money doesn't go to extra things, like allowing the poorest people to live in decent housing.
There is no point at which it stops. However, the connection between wealth and happiness appears to be logarithmic rather than linear.
I don't think Nader actually did make us safer. Cars were getting safer over time even before him.
Volvo and Mercedes-Benz led the world in safety, Ralph Nader didn't have anything to do with them.
I'm just going to quote Zvi in full here: from https://thezvi.substack.com/p/economics-roundup-4
And I would add that we have a lot of new undeclared part time jobs. Last week I spent two hours and five phone calls trying to book a referred appointment for an ultrasound. This week I will be calling various offices about some weird medical billing that I don't understand and seems wrong.
"Paper claims that we are gaining 0.5% per year in terms of how much welfare we get from across a variety of categories from increased product specialization and variety. Households increasingly spend funds on specialized products that exactly fit their preferences, with the increased variety driving the divergence in consumption.
This is also evidence we are richer. Increased product variety requires people able to consume enough, and pay enough extra for quirky preferences, to justify greater product variety. This represents a real welfare gain. However, instead of making people feel less constrained and wealthier, it puts strain on budgets and competes with and potentially puts additional strain on raising families rather than making it cheaper to raise one.
I very much appreciate the product variety, but increasingly I think we need to consider three different measures of wealth:
The welfare value of the experience of the items in a typical consumption basket.
The combined welfare value including goods that remain unpriced.
The difficulty in purchasing the typical consumption basket, and what affordances that leaves for life goals especially retirement, marriage and children.
Or: The Iron Law of Wages proposes that real wages tend toward the minimum to sustain the life of the worker. So we can measure four things.
The minimum real wages required to sustain the life of the worker.
The welfare value of that minimum consumption basket.
The surplus available after that to the typical worker and what that buys them.
What else is available that is not priced.
When we either effectively mandate additional consumption, such as purchasing additional safety, health care, residence size, education or other product features, or our culture effectively demands such purchases, or the cheaper alternatives stop being available, what happens?
We do increase the welfare value of the minimum basket. We also raise the cost of that basket, which reduces everyone’s surplus.
What happens when things that people value, like community and friendship and the ability to raise children without being terrified of outside intervention, and opportunities to find a good life partner, are degraded?
Life gets worse without it showing up in the productivity statistics or in real wages.
The current crisis and confusion could be thought of as:
The value of the minimum consumption basket is going up a lot.
The cost of the minimum consumption basket is going up less than that.
Real wages are going up, but less than the cost of the basket, so the surplus available after purchasing the basket is also declining.
Key other goods and options are taken away, like those mentioned above.
Economists say ‘workers are better off,’ and in many ways they are.
People say ‘but I have little surplus and do not see how to meet my life goals and I have no hope and my life experience is getting worse.’"
Enough for what? If you are seeking to maximize total happiness then the correlation with GDP is very weak, if at all existent. e.g. Bhutan is frequently said to be one of the world's happiest countries, and it's GDP per capita is around $3500. That's about the point at which your population doesn't need to go short of calories or die from lack of access to shelter. Once you go beyond that point...does electricity increase long term happiness? Does education? Does indoor plumbing? I don't think any of these things is clear. Why is the 1990s American (or 2020s Western European) material standard of living the ideal to strive for, rather than the 2020s Bhutanese, or indeed the 1790s American?*
On the other hand, people generally prefer to be richer than poorer, and if you respect people's preferences, then being richer is good actually. Also, wealth is power. European economic stagnation in the last 30 years has translated into a loss of hard power. Maintain that long enough and you end up at the receiving end of an opium war or a bunch of conquistadores or something, which is definitely bad (at least for you, obviously it's very good for the conquistadores). Meanwhile, America's robust economic growth over the same period has helped it mostly maintain it's hard power. Of course, as long as Europe can shelter under America's security umbrella this isn't a problem, but that umbrella may not be extended indefinitely.
* Relative to 1790s America there is a big difference in life expectancy, and certainly it is better to be alive than dead. But most of the gains there just come from childhood vaccinations, antibiotics, and an understanding of the germ theory of disease. e.g. Cuba at GDP/capita of 7k can match US life expectancy, and the aforementioned Bhutan is only a few years off. You certainly don't need modern Western European (or 1990s American) levels of economic development to get modern-Western life expectancies.
“at what point do you have enough?”
I encourage you to resist thinking like this. You can ask “At what point do *I* have enough?”, but let other people answer for themselves.
And at any rate, the question shouldn’t be “What does any given person need?”, but rather “What would any given person choose, if given the choice?”
Oh, but the meta-question here really is: what makes sense for presumably intelligent people to think about for free?
Want to spend time thinking about how to
a) make bigger and better tchotchkes for people who already have plenty
rather than
b1) alleviating poverty at a world scale
b2) making energy cleaner and cheaper, and our use of it more efficient (PS: yes, of course I know that "energy conservancy" should be really "low-entropy energy source conservacy" or "watch your increases in entropy')
b3) discussing the pressing and real needs (health, education, clean environment) of people who have enough tchotckes (note: other people also have those problems)?
Be my guest! But I hope you are being paid for it, preferably by the people concerned :).
>After a few disappointing years, these are finally coming into their own. The expert I talked to said...
Scott, it was a pleasure to meet you at the conference, and to speak with you about driving automation. To be clear: no firm, yet, is explicitly proposing self-contained pods within automated vehicles, but I do think the only way we're going to get shared-use at scale is if SOMEONE does this.
If the light-touch approach that Waymo and Zoox (and to a certain extent Cruise) are using works, and we manage to avoid a suffocating blanket of regulation on the sector, then this model would be a promising avenue for a new entrant: even if 'incumbent' firms fail to offer a combination of an optimized-for-sharing vehicle, some upstart will: they could compete on price all by themselves, and more so if they can get a licensing deal with transit operators.
(BTW I write about self-driving and innovative mobility at my own Substack, CHANGING LANES, and would be delighted if readers of ACX joined my audience)
Would you like me to link your blog? I was going off the rule that we shouldn't name names of anyone we talked to, but if you're commenting publicly here then I might as well.
Please do! I think many people at the conference have good reason to prefer a cloak of anonymity, but I am trying to grow my audience, and wish as much (good) publicity as I can get
Another thing that might help with that is that governments may be more willing to tax robotaxi companies for congestion externalities than people driving their own cars (purely for vibe reasons), which would make the cost difference sharper.
Yes, that would help shared robotaxi vs. one-rider robotaxi... but it would hurt ALL robotaxi vs. their #1 competitor, the private car.
Such a move would be a really bad one and would make overall congestion worse. That said, it's QUITE plausible that a populist government would go that way.
The best outcome would be that as electric vehicles wax, and correspondingly the gas tax wanes in its ability to finance maintenance of roads and other transport infrastructure, that governments move to a global VKT (vehicle-kilometre travelled) charge. Cars today are pretty-much always connected to the Internet, which means collecting VKT data would be straightforward for a government that wished to do so. Such a charge could nudge in all sorts of ways: less for shared rides, more for solo; less during off-peak hours, more during peak hours; and so forth.
I'm imagining this would only meaningfully happen after robotaxis already have enough market share to be a significant cause of traffic (at which point their other benefits and momentum are enough to swing them over into being the primary mode of car use anyway).
Re VKT charge, you'd also want it to be location-based (downtown trips cause more congestion than rural ones), but that's an easy enough adjustment. If by the time these rules start coming up most people in cities use robotaxis instead of private cars anyway then maybe local governments would be able to pass a universal VKT tax without riling populist opposition.
I'm writing a book right now (https://www.amazon.ca/End-Driving-Transportation-Planning-Automated/dp/0443223920/ref=sr_1_1), forthcoming 2025, manuscript due to the publisher in six weeks!, that argues that, as much as we might wish otherwise:
1) the benefits of widespread robotaxi use would be significant, but
2) those benefits WON'T be enough to make them the primary mode of car use, not on their own, so
3) we should start arguing for the right mix of policy and regulation NOW to try and shift widespread-robotaxi to the default path.
I'll be explicating these arguments on my Substack over the next year if you're interested.
Location-sensitive VKT could be integrated with legislation defining GPS navigation as mandatory safety equipment, and avoid populist opposition by also integrating automated speeding tickets and similar traffic enforcement. Less paperwork all around, takes away a cheap pretext for police harassment, yet makes the more serious crimes (hit-and-run, carjacking, etc) vastly easier to investigate.
>Location-sensitive VKT could be integrated with legislation defining GPS navigation as mandatory safety equipment
Gaa. That's pretty much tracking the location of all the people, all the time. I suspect that, at least in the long run, this may well be unavoidable, but it pretty much would remove the last vestige of non-online privacy.
A car's license plate, and where it is when it's driving on public roads, already lack reasonable expectation of privacy. There'd be no need (nor ready method) to record who exactly is inside the car, so someone inclined to travel anonymously could likely still find plenty of ways to do so.
In fact, the transparent impossibility of thus-equipped cars being stolen to useful effect - or even negligently misplaced - might allow the formation of a robust subculture of peer-to-peer rental or borrowing, and thus a sort of VPN equivalent. Means investigators would probably need a judge to sign off on digging through relevant records when the vehicle's owner wasn't inclined to cooperate, which seems likely to be an all-around improvement in due-process protections compared to the many ways a traffic stop can go wrong.
Here's an argument for benefits we've already been getting from improved transparency: https://www.paulgraham.com/inequality.html
"Cars today are pretty-much always connected to the Internet, which means collecting VKT data would be straightforward"
This will, unless somehow otherwise accounted for, make older vehicles without the ability to track this data get subsidized. One possibility is substantial increases in annual vehicle registration fees, but that penalizes those who drive the least. Maybe a gas tax which is reduced based on VKT data?
The straightforward solution is get a (verifiable) odometre count at time T, another one at T+1 year, take the difference, and charge accordingly. This would be a remarkable pain in the neck, but that might be a feature to encourage non-connected vehicles off the road.
This approach works as a straight-up VKT charge but fails to do any nudging regarding congestion, but there's no straightforward way to do this absent a connected vehicle (or massive infrastructure investment).
>The straightforward solution is get a (verifiable) odometre count at time T, another one at T+1 year, take the difference, and charge accordingly.
Yes, that gives VKT without congestion pricing, but it also avoids a massive loss of privacy that continuous GPS data would cause.
True. Pick your poison!
In many countries, vehicles need to have a regular checkup to stay licensed (for example the TÜV in Germany), so for those taking the odometer value is easy.
Could you briefly summarize what mistake you think Tesla made vs. Waymo, etc.? I'd have guessed cameras vs. LIDAR, but I thought that cost was a major issue for LIDAR, and I don't really know much about the topic. Or maybe link to an article you've already written on the topic?
I wrote this back at the beginning of September. That was before the Cybercab launch but the overall picture was clear then: https://www.changinglanesnewsletter.com/p/tesla-isnt-going-to-succeed-in-robotaxis
And then I wrote this AFTER the Cybercab launch. It's a detailed analysis of the launch, which ultimately did not make me update my views: https://www.changinglanesnewsletter.com/p/ten-notes-on-the-tesla-cybercab-launch
Thanks!
Why do you think so? What have I missed?
Banned for low content, high temperature comment.
Another possibility would be using technologies already developed for dating apps to allow prospective robotaxi customers to efficiently select among their potential fellow passengers. Lots of people can tell at a glance who they would or wouldn't be able to get along with, while clever route-planning algorithms could calculate and display appropriate fare adjustments, allowing tricky externalities to be priced into a simple, elegant swipe right / swipe left decision.
This would lead to all sorts of interesting scenarios, like single women who spend all day shuttling back and forth between the two richest parts of town in the hopes of meeting a wealthy dude.
Some sort of opt-in will likely emerge. THIS kind had not occurred to me! I will file it away...
On the topic of "Humanity is very good at mass manufacturing things in factories" here's a fun anecdote from Neolithic Africa: we have evidence of humans mass producing bone tools (presumably to trade) as far back as the Paleolithic era. Elephant long bones were systematically broken to make blanks appropriate for shaping tools in what amounts to a Neolithic factory efficient enough to drive Paleoloxodons to extinction. You can learn more about it here if you're curious: https://doi.org/10.1371/journal.pone.0256090
Without re-reading SB1047 to check … if it had become law, and the hypothesised mass casualty incident had happened, and some of the people who got killed were in California, the liability provisions would apply even if the negligence that caused them to get killed were in, say, Texas.
New version of the AI alignment problem: the AI is only allowed to kil people if they aren’t Californians. This seems likely to be at least as hard as “don’t kill everyone”.
Alternatively, the AI is allowed to kill people, but if any one of them is a Californian, it has to kill ALL Californians. That might be slightly easier.
I think you mistakenly wrote MW/h instead of MWh in a few places.
Thanks, fixed.
This is all great until you talk about factory farming. Under some reasonable moral theories, we are therefore declining. But Tyler, for example, has pretty weird takes on why animals don’t matter (see his podcast on 80k for the full view). I would love to hear what others at the conference (Pinker, for example) would said about this.
What effect should you expect to have progress (let's say increasing wealth) have on factory farming? I would guess Americans are rich enough, and meat cheap enough, that we don't have the simple relationship that more money = more meat-eating. So shouldn't we expect that more wealth gives us more money to spend on moral luxuries like non-factory-farmed meat?
Interesting point.
I guess this goes into one of my issues with the progress movement in general (while I’m certainly on board with a bunch of things) — they don’t define progress in a way that can be measured! Progress for what? Progress in what sense?
Also, if you don’t think that morals are aligned with more money and you think something like the signaling theory of altruism is correct, the incentives can be wrong. Perhaps the future can be better. Maybe we will live in a future where we do solve this problem, and the incentives are better aligned - for instance, there is social stigma around eating meat and the alt protein industry does very well. This seems plausible to me. Under many moral theories, however, it still may not be worth the 23 million animals that are basically tortured to death each day.
I think it's mostly fair to think of them as talking about economic progress / GDP. I also think it's a fair first-order assumption - obviously not the whole story, but better than not making the assumption - that whatever problem you have, having more money is a good start to solving it, on the grounds that if people are only willing to devote a fixed fraction of their income to solving it, then the higher their income, the more money goes to the solution. If everyone was a billionaire, then the cost difference between factory-farmed meat and humanely-raised meat would be trivial, and people would give into animal rights activists just to shut them up.
The main exception is where wealth itself causes more problems. I think this might be true in the developing world where more money might translate into more meat-eating. I'll sort of plead the Pareto exchange defense here, but I recognize it doesn't work that way in the real world.
I hear that argument. Seems somewhat fair.
I still think EAs have a better proxy for thinking about good or progress, so I’m not really sure why we would want an only okay proxy if it takes away counterfactual resources from the better proxy -
On the other hand: perhaps multiple proxies are good because they’re a good way to hedge on values that other proxies would miss (as long as one of the proxies isn’t just the terminal values which could be the case with EA).
What is the good proxy that we have?
In theory, some mix of total utilitarianism and prioritarianism (I agree that things get tricky when you actually try to design a function that fits the intuitions that you want because of the impossibility theorems in population ethics, but the approximate idea should do much of the trick). In practice, while I’m not very well read on the topic, from what I do know it’s probably something like total QALYs.
I do think that these are clearly harder to measure than GDP, but if we want a better proxy, it’s probably this. I still stand by my point about maximizing for multiple proxies in case one misses something (either for systematic, moral theoretic, or practical reasons).
There are some things that will definitely be missed if you only optimize for GDP (e.g. animal welfare), and while it may even be a good proxy for more stuff in the future (the reason you described seems plausible and maybe likely), it still may not be a better total proxy because of all the damage done on the way (i.e. a billion factory farmed animals being killed every year).
Even if higher income does increase demand for meat/eggs/dairy, it's also true that non-factory-farmed meat tastes better and works better in many recipes (eggs and butter in particular). We should expect, in the limit of high enough income, that people will demand the higher-quality non-factory-farmed meat, eggs, and dairy products.
It’s true. I think this will largely be a cultural thing, though. Seems to be a large factor in things like what people eat (traditional foods often stay common in countries when there are better alternatives, for example -- though I admit that there are other factors at play here as well). I just hope the culture moves in the right direction (which I’m optimistic about!).
More general economic progress would probably also affect other countries, including ones that would eat more meat if they could afford it.
Maybe, but it's unclear that we would the increased wealth for a reduction in meat eating, or use the wealth for even more meat eating.
The factory part I think is irreversibly tied in with the actual wealth creation. The greatest source of increased wealth in physical terms, including food, is in finding more and more efficient ways of creating the physical products. Factory farming is a natural result of the process that makes things efficient. Other than mandating non-factory conditions (which seems counter to the Progress mindset and goals), increasing factory conditions seems inevitable.
Sufficiently advanced factory farming hits a tipping point where it abruptly becomes cruelty-free by default. Instead of raising, feeding, and slaughtering whole animals, better to build a machine which turns electricity and distilled chemicals directly into a pure, continuously-extruded tube of the most desirable types of muscle tissue.
It would. But factory farms are also part of the history of progress. We should pay attention to the fact that some past progress has been moral monstrosities, so that we are thinking about whether some current instances of progress might be similarly monstrous so that maybe we should direct progress differently. (Some might say that by definition, moral monstrosities aren’t progress. Still, studying the monstrosities is going to be relevant to distinguishing them from true progress in advance.)
Earlier, West Indies sugar plantations were a font of wealth created largely by improved ships and navigation and transatlantic trade. This was all impressive progress, but the sugar plantations themselves were basically hell on Earth and were morally about one hop above Nazi death camps.
Tyler has a number of weird beliefs, but discounting non-human animals is probably one of his more normal ones.
The way he does it is quite weird. I agree it’s a normal belief.
With the increase in technology, I'd personally expect us to decrease in meat-eating. The current alternatives are 1) worse-tasting in various ways 2) different-tasting, which can be a big problem in converting people who didn't grow up with it 3) more expensive and 4) politically polarized.
Still, once we have the technology to synthesize better meat in terms of taste & cost, I don't expect factory farming to linger for a long time. They'll latch on and try to make it harder, but I expect them to fade away.
Ex: Climate change. Oil companies have more power & state-backing, as well as the problem being harder to substitute. I can't just replace the oil in my car with electricity without a large investment, even if the electricity happened to be cheaper the car is not. This makes so the people who are most price-sensitive (buying used cars from 1990s, 2000s or early 2010s) can't just switch to electricity even if it was quickly cheaper.
Food is less like that. If there's a good cheaper meat substitute on the market, then that makes it far easier for people to individually switch over.
That, and the technology for synthesizing products in that manner likely being useful for a lot of other food, makes me optimistic about meat production.
I think there are good arguments for this (and I appreciate you bringing these up!), but like I said in response to Scott, the object level arguments miss what I’m saying. Technology in this sense entirely leads to progress always is clearly missing a lot. The goodness of technology and science is that they are contingent on other facts about the world. While it may be a proxy for values much of the time, it should also be viewed as a proxy that is quite removed from the terminal values we have.
One example of this might be that without technology we would never have been able to systematically kill animals - maybe the bad we do now outweighs the future good, even given that we stop at some point in the near future (20 million animals being slaughtered every day might be good reason to believe this).
Another thing here might be that giving humans more power might lead to them killing themselves with existential risks (see Bostrom’s Vulnerable World Hypothesis paper for a good form of this argument). While I certainly hope humans do not do this, the chances are not small enough to make it negligible or say that we should just rush technology and science without thoughts about what is it that we truly value.
Solar is amazing when it’s a small percentage of overall capacity. When you are actually trying to replace a modern power plant with it, like a coal power plant, it’s incredibly expensive.
Capacity of solar often depends on assumptions of great weather, or average weather, when we really want electricity whether the last week has been average, or particularly cloudy. You also need to factor in the monumental cost of batteries, as we want electricity at night too, and when we consume electricity is usually independent of when it’s produced.
In reality, while it’s fine to treat cost per kWh when dealing with small solar plants, as soon as solar gets to be above ~10% of the energy mix, or is used to replace existing fossil fuel plants, the actual cost to produce a robust system that’s almost equivalent (but still worse in many ways) to their energy generation capacity is 10-100x as much.
In Germany electricity prices are going negative more and more often. Meaning not only does solar not make money, fossil fuel plants that still need to run for the night actually have to pay to keep running, as they can’t easily shut down and power up without incurring significant cost. This raises fossil fuel prices, which may work further to incentivize green energy adoption, but the more negative prices it produces, the more we’ll have to pay for that throttle-able capacity.
It’s not an insurmountable problem, but a grid based on solar (or wind) would be 1-2 orders of magnitude more expensive than cost per kWh might make it seem.
"In Germany electricity prices are going negative more and more often. Meaning not only does solar not make money, fossil fuel plants that still need to run for the night actually have to pay to keep running, as they can’t easily shut down and power up without incurring significant cost. This raises fossil fuel prices, which may work further to incentivize green energy adoption, but the more negative prices it produces, the more we’ll have to pay for that throttle-able capacity."
Explain like I'm an idiot: why do electricity prices go negative? Is there no way to "discard" unused electricity?
They can discard used electricity in extreme circumstances, but generally it’s far preferable to either increase demand, or attempt to decrease the supply. The grid must be almost perfectly in balance at all times, which means keeping demand and supply in step.
The negative electricity cost usually only lasts a few minutes to an hour, around noon on sunny, windy days for example. For power plants, that’s a problem because it costs them a lot of money to shut down and start back up, and a lot of time to do so, which has its opportunity cost. Since they cost fuel to operate, their profitability is usually at a certain price above $0, although fixed price or minimum price contracts complicate that calculation.
Renewables basically cost $0 to operate, and maybe even cost little bit to curtail as that requires more complicated systems. Thus, the solar farms will keep sending the maximum amount of electricity they can produce to the grid so long as the price is positive. Unlike power plants, that will reduce production when prices are low, solar is completely demand-inelastic for all positive prices. Better to make $0.01 per kWh rather than $0 after all.
Negative prices are the way to de incentivize this oversupply. For the power sources that have high shut down/startup costs it makes continued operation less appealing, and for solar with no marginal cost it actually has an effect on that normally inelastic demand. Another solution power grids take is by literally paying producers to reduce production, which is less likely to cause economic troubles for the producer, but directly translates to higher electricity prices for the consumer.
It’s a signal that at times, we already produce too many renewables for our current demand. If we doubled the solar capacity, negative prices would stay for longer, and lower prices during peak solar production would also last longer. Profitability on solar originally assumed normal pricing during solar’s peak production, but as they actually get to be a significant portion of the grid, their $0 marginal cost will bring the price people are willing to pay for that electricity down to $0.01 when solar supplies most of the grid. This is good for consumers, but doesn’t bode well for justifying exponentially increasing solar investment.
Here’s a good article to learn more: https://open.substack.com/pub/gemenergyanalytics/p/some-consequences-of-solar-on-the?r=bbrjn&utm_medium=ios
The obvious way to get more demand in the system when the wind is blowing or the sun is shining is to reduce retail prices when this is happening and alert people via a smart app. Instead the market pricing is at the wholesale not retail level, so consumers have no incentive to use carbon free electricity.
We had a company like that in Texas. Then during the Snowpocalypse the Public Eye of Sauron gazed upon them. They were smote with its wroth.
There was a power company that just passed grid prices through to the customer.
But the impression I got was that they hosed themselves by continuing to provide power at grid prices to their customers, who were contractually obligated to pay for the power they consumed, but who didn't actually have the money to do that. This left the company with gigantic liabilities and no prospect of collecting the revenue that was notionally supposed to cover them.
So did every other power company at the time, the only difference is they nominally passed the cost to their customer while at the same time telling them not to pay it, while the others billed the usual $1 against the $1000 cost of power. Everyone had to swallow the emergency pricing. But if all the companies had actually been passing the high price to their customers, then price signals might have done their job and caused enough voluntary decrease in consumption to prevent the emergency.
In Germany, there's at least one retail electricity supplier offering dynamic pricing based on live market prices. The problem is: As a consumer, what are you going to do with this? Maybe you can schedule your dishwasher or dryer to run at 13:30, but the major loads like heat pumps or electric cars aren't as flexible: Heat pumps want to run in the early morning, to re-heat the house after a cold night, and electric cars have to be be recharged when they're at home, which is typically in the evening and at night.
I think the market still needs time to adapt. Employers can offer workplace car charging as a cheap perk. You can heat up (or chill?) a big tank of water as a kind of thermal battery if cheap electricity supply doesn’t correspond with home heating/cooling demand.
My point is that the biggest part of consumer demand is relatively inelastic and can't be shifted around much. If I charge my car while at work, I'll want to charge it whether the sun is shining or not, lest I pull an involuntary all-nighter.
If electricity prices were negative right now then I'd turn on all the heaters in my house. Then I'd turn on all the air conditioners in my house. Then I'd open all the windows just in case one turns out to be a lot more powerful than the other.
You're doing God's work, son.
Professional arbitrageurs operate at the wholesale level by investing in batteries, then outcompeting peaker plants at grid stabilization services. https://caseyhandmer.wordpress.com/2021/05/20/the-unstoppable-battery-cavalcade/ Given a few more years, I'm sure many other intermittency-tolerant industrial applications will spring up to take advantage of a literal free lunch.
Octopus does this in the UK, I get free electricity alerts the day before (as they can predict it based on the weather).
You can store energy efficiently by lifting water and using hydro (at the scale of a region, not at the scale of an individual, obviously).
Sadly this is entirely depend on local geography and therefore very limited. My country for example has reached its peak when it comes to hydro decades ago. It's not a solution.
These are two different things: some countries do exploit existing hydro resources to the fullest, but that doesn't mean that gravity batteries have been exploited to the fullest, or that gravity batteries would be very limited. What the actual limits to gravity batteries are - now that's a good question.
Gravity batteries that do not already use existing bodies of water are just limited in the same way as most things: they are too expensive and therefore currently pointless.
Just so that everybody knows what we are talking about: https://en.wikipedia.org/wiki/Gravity_battery
Can I infer from this that the current limits on storage capacity are entirely in the amount of energy that can be stored (or the ability to ramp up storage quickly) and not at all in the efficiency of storage?
Jenny Chase seemed to suggest that the reason we don't just run a desalination plant with any energy spikes is the cost of the desalination plant?
Storage efficiency is somewhat of an issue, as round trip efficiencies tend to be like 60-80% depending on the method. But yeah capacity is the bigger problem.
Desalination plants are expensive enough that in order to amortize the capital cost, you want to run them with a high capacity factor. The same is true for plants that produce hydrogen by electrolysis. Also data centers, aluminum plants, etc. Using electricity in productive ways tends not to be cheap. That makes it tricky to increase demand during production spikes.
How expensive are desalination plants? Google's bot says that the capital cost is 500 days of operation. That sounds very low to me, allowing a low capacity power and timing of electricity prices. Do you disagree with the factual claim of 500 days of operation or do you think that is a high capital cost?
Well, let's see if we can find a good example. Ras Al-Khair Power and Desalination Plant, completed in 2014, is the world's largest hybrid water desalination plant, according to wikipedia (hybrid apparently means it uses both reverse osmosis and multi-stage flash distillation; I'm not sure why). But it also includes a combined cycle natural gas power plant, so we have to separate that out.
EIA gives a combined cycle gas cost of $614/kW for 2015, for just construction.
https://www.eia.gov/todayinenergy/detail.php?id=31912
However, one source says total CAPEX for gas is about $1226/kW in 2024, about double that.
https://www.statista.com/statistics/243707/capital-costs-of-a-typical-us-combined-cycle-power-plant/
or from another source, $950/kW for an H-Class 1100 MW Multi-Shaft Combined Cycle - Multi-shaft configuration rated 1,083MW and 59.4% efficiency, $958 million total (950 $/kW installed) and 12.20 $/kW fixed O&M cost.
https://gasturbineworld.com/capital-costs/
Going with the middle figure of $950/kW, we can compare to Ras Al-Khair. Its total cost was $7.1 billion. The 2400 MWe of gas generation, we would expect to cost about $2.3 billion. So the desalination part is about $4.8 billion, give or take maybe 500 million.
The throughput is 1,036,000 m^3/day, so capital cost is roughly $4600/(m^3/day). You know, plus or minus a factor of 2.
But how much is that desalinated water worth? Wikipedia cites an article (humanprogress.org/desalinating-water-is-becoming-absurdly-cheap/) that says one plant in Israel is contracted to produce water at $0.41/m^3, but it seems to vary a lot with location. Other sources suggest costs "ranging from $0.50 up to $2.50 per cubic meter." (https://medium.com/@desalter/what-is-the-price-of-desalinated-water-and-how-does-it-compare-to-other-sources-of-clean-water-02f20a7b64fb)
If we take $1.50 as a reasonable midpoint and use simple division, that gives a payback time of about 3000 days (~8 years). In practice, you would have to subtract off any fixed operating costs as well as the cost of electricity, and then compare to the risk free interest rate. Even if electricity is cheap, at first glance it doesn't look like a great investment unless you can operate at a pretty high capacity factor *and* sell the water at a pretty high price.
What is the most efficient way to safely discard electricity when needed?
As for actually discarding electricity (which basically never happens for actually curtailing excess energy), turning it into heat is the only real way. That usually involves turning on what are basically big space heaters called Load Banks: https://en.wikipedia.org/wiki/Load_bank
Actually discarding excess electricity isn't usually part of normal operations though, as it's far easier to increase demand from consumers who are price-sensitive, or reduce supply directly at the plant.
As the original post writes, a lot is to do with how power plants run, such that they can't just turn on and off whenever they want without incurring cost, so it's cheaper to pay end consumers to just use electricity. Also, outside of Texas most energy markets will have a capacity market, paying generators to keep plants online in case of spikes in demand.
Most grids will have some capacity for using up excess electricity, but not massive amounts; the aim is for generators and retailers to balance supply and demand through accurate forecasting.
Electricity prices go negative for two main reasons:
- certain generators are paid whether the power is needed or not (eg the German feed-in tariff system means that solar and wind will get paid if generated). This is basically solvable by changing policy that was designed for lower generation numbers.
- it's worth keeping some generators online for later. As I recall this was a problem in California for some years, because the gas generators were needed in the evening, so had to keep spinning in the day to be ready - and hence generated some power, paying negative prices. I'm pretty sure this is now alleviated by batteries.
The only way to “discard” unused electricity is to put it in a battery, or run some powerful appliance or something. If the grid has too much electrical power running compared to how much is consumed, it messes up the voltage, which can damage lots of devices.
You could just build a giant resistor. But cooling it would be a problem.
But what if instead of one giant resistor you put a little resistor on every third power pole?
That’s fine if you always have excess electricity - but even places with excesses sometimes usually have the opposite issue. What you need is something you can turn on and off fast enough to balance generation and consumption. Gas power plants have been the common solution on the generation side. Batteries are a good option on both sides. Resistors seem like a backup option, but hopefully there’s something better.
Note that aluminum refining by electrolysis is essentially half a battery, like the recharge phase of a rechargeable battery. What I don't know is whether it can be done flexibly...
No mad scientist's lab is complete without one of these: https://en.wikipedia.org/wiki/Jacob%27s_ladder_(electrical)
The issue is that the German government committed to pay a certain amount for (parts of) solar energy regardless of market prices. So this part of supply does not react to price signals.
On the consumption side, there will soon be ways to take up the surplus energy. Several companies have noticed the price fluctuations and are now building huge battery farms (in the order of GWh), to buy up the electricity when it's cheap (or in the extreme, negative), and to sell it when electricity is expensive. Once they are ready (next year or so), the negative electricity prices will disappear.
Fun fact: some of the battery farms are built right next to the nuclear power plants that Germany switched off, because those places have the connection to the power grid already in place.
Have you looked up the recent trends in Australia, especially South Australia? Solar there is something like 70% of total annual production. They've been working with Tesla to install lots of Powerwall units as a distributed virtual power plant to smooth out demand. The long-term official plan is to overbuild solar to 500% of electricity demand, and consume the excess for charging EVs, electrifying various industries, and manufacturing synthetic fuels and chemicals.
Edit to add: Phase change materials for HVAC are another way we can profitably time-shift demand to whenever electricity is cheapest (either times of peak renewables output or overnight) or operate the system at a lower, more efficient heating or cooling rate without loss of comfort. We're really only at the very beginning of implementing ways of matching demand to supply, because historically we didn't need to.
I think you’re vastly overestimating the solar percentage in Australia. Solar is currently at 16% of total production according to a quick google search.
Eventually they probably will be capable of overbuilding to that extent, but that involves 5x the cost a pure kWh comparison wouldn’t give you. That, plus the extremely expensive batteries to have multiple days worth of storage for the entire grid, and the total cost of it is going to be significant.
The problem isn’t impossible to solve, but it is significantly more difficult and expensive than “solar is now cheaper than coal” seems to imply.
You're right, it's 75% wind+solar and not just solar. Solar PV is 24% as of 2022. My mistake, sorry. The fastest growth is in solar, though. And yes, it *is* much more complicated and difficult and expensive than just saying "solar is now cheaper than coal." Solar started being cheaper than coal in Chile in 2015 (approximately), and only now are all the other bottlenecks becoming feasible to overcome in optimum markets.
The rest of my comment is pointing that there are a lot of efficiency-improving, money-saving things people and companies can do today, and are starting to do, that look like energy storage (dispatchable opportunistic demand?) to the grid if you manage them properly, long before you get to building grid-scale stationary energy storage systems. We're not dumb enough to just blindly pay for 5x as much electricity as we need and build week-long battery energy storage, but yes, that would be the implausible upper bound. It remains to be seen which markets are clever enough to actually be smart about it and find the lower bound.
I can’t find this 75% number. If it were true it would indicate Australia has had to contend with this problem and done so successfully. The Australian government says solar is 16% of total generation and wind 12%.
Wind doesn’t have the same problems as solar, but it has that are equally difficult its own due to the unpredictable nature of it when it makes up a significant portion of electricity generation.
Not Australia. South Australia.
From official national gov't statistics: https://www.energy.gov.au/energy-data/australian-energy-statistics/data-charts/australian-electricity-generation-fuel-mix-calendar-year-2023
From the SA state government: https://www.energymining.sa.gov.au/consumers/energy-grid-and-supply/our-electricity-supply-and-market
Ah, was looking at Australia as a whole.
It looks like the majority of their renewable generation is from wind: https://www.statista.com/statistics/985429/australia-energy-generation-capacity-in-sa-by-source/ which doesn't have the same issues as solar in that the generation is more randomly distributed. It seems they are also maintaining full natural gas capacity, even if they rarely use it. Current electricity cost per kWh seems to be ~$0.34, or even as high as $0.45, depending on the source.
This tracks with the discussion though. Wind doesn't have the same limitations as solar, but neither the rapidly decreasing costs. It only works in windy areas (and wind is pretty consistent over the entire year), but still requires almost full capacity from non-renewable sources (in this case natural gas). The end result is some of the most expensive electricity in the world.
Personally, I'm cautiously optimistic, but highly doubtful the SA model can be replicated anywhere besides high wind, low population density areas of the world. The immense electricity price tag isn't exactly tantalizing either.
A lot of people like Nuclear because it requires no additional non-renewable capacity, is completely throttleable and has low marginal cost of production, leading to very low electricity costs depending on the financing structure during construction.
South Australia does have 75%, but they have higher prices and rely on the larger grid in a similar way as Scotland
The Net Zero Australia report is favourable enough to wind and solar that they think there's nearly no case for nuclear, yet they also project that energy costs would rise from 6% to 8% of GDP (not certain, but very distinctly recall it being two percentage points)
A real energy revolution from wind and solar would have to involve batteries doing incredible things, which they could based on solar
Yes, the state has interconnects to other states that can export up to ~20% of peak power output currently, and yes they're trying to increase that further in the future, because why wouldn't they? You solve the easy problems first. Heck, there's talks of building multi-GW undersea cables from Darwin all the way to Singapore, because that's slowly becoming a viable way to decarbonize Singapore. Long distance interconnects increase maximum penetration of intermittent resources locally *and* overall by reducing output variation due to weather, time of day, and other factors.
As for the GDP prediction: How much of that is increased prices, and how much is increased demand? Australia has a lot of heavy industry it wants to electrify, and then of course you have to add in EV growth too. And while I have my doubts about a lot of specific projects and timelines, they're genuinely trying to increase green hydrogen, ammonia, methanol, and other chemicals production for export, which would eat up quite a lot of electricity production (and, assuming they're responsive to real time energy prices, would help minimize required battery storage).
As another example: see how the Port of Melbourne is trying to become a green methanol bunkering hub for methanol made in Tasmania (to start with). There are shipping companies ordering dual-fueled and fuel-agnostic ships, and this is an incremental step towards using one problem (high renewables penetration) to reduce another (shipping emissions).
I'm not saying everyone can do exactly what they're doing. I'm saying there's a path, and under optimal conditions the path is finally becoming traversable, and the number of obstacles that get in the way is steadily going down. Making the path viable in more places still requires tech advancement and an enormous amount of hard work, but nothing so revolutionary we can't see a way to it by normal engineering and investment processes.
I seriously doubt multiple days worth of batteries are necessary. Single-digit hours of peak capacity is enough ninety-some percent of the time, and seasonal variation or the occasional week of randomly bad weather can be planned around in other ways.
You’d definitely want to look more into it then. Unless you’re comfortable with accepting rolling blackouts, you either need other capacity (like natural gas in the case of the Australia example) or a very large amount of batteries to be confident you won’t be running out of energy. You need enough capacity to provide from the afternoon, to the upcoming morning, which is generally ~12 hours. Much more in the winter for northern hemisphere countries since the days are shorter and provide much less energy.
Alternate generating capacity is part of what I mean by "planned around in other ways," yeah. Synthetic hydrocarbons are just another kind of battery. Other part is retooling energy-intensive stuff to make hay while the sun shines.
Four hours of batteries is enough to shift peak supply at noon to peak demand in the evening, which seems like a good start, and twelve hours is still a lot closer to "single-digit hours" than it is to "multiple days." Doesn't have to be perfect, or all paid up front - arbitrageurs will gladly spend their own money to build more batteries so long as daily price cycles let them rapidly recoup the investment, while battery and solar-panel manufacturers will keep finding ways to make more, better, cheaper, so long as the arbitrageurs keep buying. No need to subsidize all that, it'll keep spinning self-funded until short-run variability is as smooth as it can reasonably get and overall demand for electricity is saturated.
Synthetic hydrocarbons can function as a kind of battery, but they are very inefficient, they only give back about 30% of the energy you put into them. They only make sense for energy intensive applications that can't use anything else.
I think you're underestimating the range of tools available to mitigate the problem. My own take is that we will need "days" of "energy storage" but that this looks very different from what either you or JamesLeng are describing.
First, you obviously understand that the difficulty here depends a lot on latitude and local climate. I assume remote, small, and near-polar communities will probably need generator backups at a near-1:1 rate, but they are so small it doesn't matter climate change-wise if they do keep burning fossil fuels, and it doesn't matter energy-demand-wise if they need synthetic hydrocarbons or hydrogen or whatever. So let's assume we're talking about a typical middle-latitude northern-hemisphere city with a mix of sunny, rainy, and (in winter) snowy weather.
In 2070, after all currently existing assets have been replaced, what does its grid look like?
Well, first, it'll be a mix of all the available energy resources. Let's ignore that for now, though, and assume it's all wind and solar, even though no utility in that kind of climate is stupid enough to do that. Let's say the worst case is you need about four days' worth of energy storage, average load.
You'll probably have 4-8 hrs of lithium (or by then, possibly sodium) ion, which will get you through a typical night and day. Then you might have flow batteries for another 4-20 hrs if anyone manages to scale them the way that the fundamentals suggest should be possible. If not, maybe more lithium, but probably not more than 12 hrs total. Most likely near the low end of both ranges, since solar peaks during the day and wind at night, wind and solar are somewhat anticorrelated, and PV is cheap enough that you can overbuild enough for average capacity to be sufficient even in winter. This will get you through 95% of the year or more, assuming you do enough overbuilding of cheap solar panels to deal with a typical winter day.
So what's next? Well, realistically any actual grid will have some proportion of hydro, geothermal, waste-to-energy (gasifiers are 4-5x more efficient and much cleaner than traditional boilers attached to incinerators), and if we're lucky nuclear. In many regions this will be enough to stretch the above storage to the point that you only need more a couple of times per year, if at all.
Elsewhere, we need to consider that "average load" now includes a whole lot of electric vehicles. I know people hate the idea of vehicle to grid because they assume it'll be done badly, which, fair enough. But a lot of people are going to be willing to get paid to partially discharge (or just not fully charge) their cars a few times a year, as long as you warn and/or politely ask them the day before. An average EV stores the equivalent of ~2-4 days' worth of power use by the average house, and many households will own 2 or more. So if 50% of people let the grid borrow 50% of their EV capacity a couple of times a year, that gets you an extra ~1-2 days of distributed storage.
If that's still not enough, then consider that heat pumps and heat storage and phase change materials can load-shift HVAC and hot water demands for residential, commercial, and industrial (including process heat) users. Companies are already building these in some places because even today it pencils out economically.
And if that's still not secure enough, consider that we're now talking about problems that will occur once a year or less. At that point, you can in fact just pay people (especially commercial and industrial customers) to use less power, and this can be done at much bigger scale than we do it today - especially as we electrify more and more of our industrial process energy needs.
And if even that isn't enough for your standards, you're more than welcome to buy a generator you'll use once every couple of years. Or buy very cheap multi-day batteries that, like Form Energy's, charge slowly over months to be ready for the rare occasions they're needed.
It’s a nice thought experiment based on vibes, but that isn’t really a useful critique. “We’ll probably have these technologies and they’ll probably be enough” doesn’t cross the gap from theory to reality.
You need to be able to have a functioning grid for (at minimum) 3 cloudy, low wind days in the winter. That means either having a very large amount of full batteries on standby containing days worth of full grid power (possible, but extremely expensive), or a throttle able energy source like fossil fuels, nuclear or hydro.
The suggestions you make will certainly help, but they’re the sort of small improvements on the grid that won’t make a meaningful total difference in creating a renewables based robust energy grid.
We must also keep in mind that in places where solar barely makes up double digit percentages of generation capacity, we are seeing negative prices at peak solar. I wonder how the economics of a new solar plant will work out when they literally can’t sell energy at peak production for more than a tenth of a cent.
At the end of the day, 2070 is too far to make any meaningful predictions or even have a useful discussion about. I’m talking about near term problems with solar that are effecting the world today, not the hypothetical technologies in 2070. Energy prices are going negative, and current examples of renewables-based grids require significant base load that’s usually taken up by fossil fuels plants on standby. Only countries fortunate enough to have ample hydro are able to cross that bridge to majority-renewable electric grids, and that’s not a repeatable thing. It’s great people are thinking up potential solutions, but as of right now they are just ideas, and we aren’t close to solving the baseload capacity problem with non-throttleable renewable energy.
A Danish study estimates that solar 'capture price' (*) in northern Europe will decline to 20% at 20% of generation, while Spain can do 40% at 40% generation - https://www.eifo.dk/media/vozjbeo4/capture-rate-analyse.pdf
Batteries could alter that very significantly or entirely. While I get the impression that physically they don't have the same low floor as solar panels due to just involving more material, according to ChatGPT batteries and panels seem pretty similar in weight per capacity (8 hours of batteries vs solar that can likely produce for 8 hours)
(*) What they would get paid in a spot market vs the average price due to high supply when they are producing. As mentioned in another comment the challenge with just building things to soak up the supply is the capacity factor of high capex electrolyzers or the like. Energy costs might be 40%, if you make that free you gain 40% but if you run 50% of the time your capital cost is doubled. Basically energy isn't a single factor unlock, although it could grease the wheels
Another major consideration is energy production throughout the year. At least in Germany (and presumably even more so the farther north you get) solar production in the winter is almost nothing. Thus, you still need to maintain a completely functional grid without the existence of solar, batteries or not (unless we’re planning on building 10-20x capacity for total demand).
In essence, we have to pay for throttleable energy sources that are independent of the weather that can satisfy demand no matter what. Those plants have very high fixed costs, so growing solar to a significant portion of the market just means the cost for electricity outside of solar production will be significantly higher. Arguments for solar over nuclear usually don’t take into account what would actually need to be built for solar and wind to replace fossil fuels.
Not anti-solar or anti-renewable btw, I just don’t see how we can actually have a renewable grid with current trajectories. As far as decarbonization goes it’s great, but certainly won’t get us anywhere near net zero.
You could probably run a cable from the Sahara to power Germany.
https://en.wikipedia.org/wiki/Desertec
Moved nowhere since 2009. Expensive (the Mediterranean is deep), plus being dependent on countries like Algeria for electricity doesn't sound too attractive. Spain could work, but the political/cultural differences and mutual distrust between Europeans and Arabs are too strong.
On the number of pages in the Federal Register -- that can be a very loose proxy for the regulatory burden, but here's something I wrote in 2003 about how loose that can be: https://stuartbuck.blogspot.com/2003/12/regulation_31.html
To quote:
Many final rules do not create brand new regulations in areas where no regulation existed before. Many, if not most, final rules simply replace a pre-existing rule by modifying it in light of changed circumstances or new research. To take an area with which I'm deeply familiar, the FCC issued an new telecommunications rule in August called the Triennial Review. It was massive. But all that the Triennial Review did was tweak telecom regulations that already existed, and which themselves already took up several hundred pages (including accompanying explanations) in the Federal Register from previous years. In fact, the effect of the Triennial Review was to loosen, if only slightly, those pre-existing regulations.
Imagine this hypothetical: The Federal Fruit Commission issues a Notice of Proposed Rulemaking in 2004 asking for comments on whether it should ban the sale of green bananas. This takes up 50 pages in the Federal Register. In 2005, the FFC issues a final rule banning the sale of bananas that are more than 75% green. The rule itself is two lines long, but is accompanied by 100 pages of explanation. In 2006, the FFC has realized that the grocery industry hates the rule. It issues another NPRM of 50 pages asking whether it should change the rule to apply to 50% green bananas. In 2007, the FFC issues another two line rule, with 100 pages of explanation, changing the previous rule to read "50%." In 2008, the FFC realizes that people are still unhappy, and that the rule doesn't really accomplish any public good. It issues another 50-page NPRM asking if the rule should be eliminated. Then, in 2009, the FFC finally gets rid of the green banana rule, along with 100 pages of explanation.
If you looked at pages in the Federal Register, you might say, "Oh my goodness, the FFC has produced 450 pages in the past 6 years just on green bananas. Regulation is out of control!" But nowhere are there "450 pages of rules" that apply to green bananas. In fact, at that point, the green banana rule would no longer exist, and the regulatory burden would be zero. Of course, the 450-page number might be useful if you want to approximate the societal effort spent on dealing with federal agencies -- i.e., monitoring rulemakings, writing comments, etc. But "450 pages" would not represent the actual regulatory burden on anything.
If that example seems fanciful, there's a famous administrative law case in which the FDA spent literally 9 years debating back and forth over whether "peanut butter" should be 87.5% or 90% peanuts. (!) See Corn Products Co. v. Department of Health, Education & Welfare, 427 F.2d 511, 513 (3d Cir. 1970). These hearings produced 7,736 pages of transcript -- all over the peanut content of "peanut butter." For all the Federal Register pages expended in that proceeding, there was only one very simple rule that kept being modified.
> Of course, the 450-page number might be useful if you want to approximate the societal effort spent on dealing with federal agencies -- i.e., monitoring rulemakings, writing comments, etc. But "450 pages" would not represent the actual regulatory burden on anything.
Keeping abreast of all that revision churn, constantly needing to re-do the cost-benefit analysis of disrupting established workflow for compliance with the new rule vs. evading or corrupting the inspector vs. fighting a legal battle to get the rule changed again, IS regulatory burden.
About supersonic flights: this one is probably a technical problem rather than regulations. Aerodynamics are so different once you go supersonic that you're basically building a whole different aircraft, and one that's way more expensive to fly (see how expensive fighter jets are per pound compared to passenger planes). And most good usecases for supersonic flights are transatlantic or transpacific anyway, so the rule about sonicbooms over land probably wouldn't completely block it.
(See also the construction physics post on the history of supersonic passenger flight attempts).
Does anyone have a good summary of why concorde stopped? It went between London and Washington, D.C. or New York City.
From wikipedia;
>They cited low passenger numbers following the 25 July 2000 crash, the slump in air travel following the September 11 attacks, and rising maintenance costs: Airbus, the company that acquired Aérospatiale in 2000, had made a decision in 2003 to no longer supply replacement parts for the air
Seems like it was for a reason other than "makes peoples windows rattle". Maybe the 2000s were a perfect storm against supersonic flights anyway.
Brian Potter’s piece on it is good. TLDR the economics of it were always pretty bad and the post-9/11 decline was just the final nail in the coffin.
https://www.construction-physics.com/p/why-did-supersonic-airliners-fail
Thank you! As soon as I saw the question I was "I know I read a great post about this, but for the life of my I can't remember which blog".
Wow this is fascinating;
"the Concorde used a series of pumps to move fuel between different tanks to adjust the location of its center of gravity (the fuel was also designed to act as a heat sink, keeping the aircraft cool at supersonic speeds"
Quick google (not high confidence) suggests that there are something like 50-100 flights between LA and NY every day.
If its' true that supersonic flight cuts that trip from 6 hours to 2 hours, then you only need 1/3rd as many planes in the air to make those trips. Is the cost increase for supersonic flight more than 3x the cost (per seat I guess) of existing planes?
You don’t quite get to cut down to 1/3. Planes need an hour or two on the ground between flights to deplane, clean, and board. Cutting from 7 to 3 is closer to reality. (And given that the supersonic speeds only apply to the cruising part of the flight, and not the takeoff or landing, it might be more like cutting 1 hour ground plus 1 hour takeoff landing plus 5 hours air to 1 hour ground plus 1 hour takeoff landing plus 1.6 hours air.)
European ULCCs (particularly Ryanair) have managed to get it down to around half an hour on average. If a supersonic aircraft has similar capacity to Concorde then the deplaning and boarding could be done in a similarly expedited timeframe as Ryanairs 737s. If they are built to 777 capacities then you of course end up with longer turnaround times.
Supersonic aircraft would have higher maintenance and refueling burdens though.
The Concorde cost about 3-4 times as much per passenger seat to operate as the 747 and about ten times as much as the 737, so yeah, it still doesn't pencil out unless passengers are paying a large speed premium.
(This is just counting operating costs, not development costs. Development can make planes much cheaper - as demonstrated by the cost gap between the 747 and more modern 737 variants - but is much more expensive for supersonic jets, so that adds another cost barrier for them).
(to clarify, one reason the 747 is so much more expensive to operate than the 737 is its longer range; the Concorde's range is closer to that of the 737, so in practice that 3-4x factor is still underselling it)
The single largest expense for airlines today is fuel, followed by labor and the cost of the airplane itself. https://www.iata.org/en/iata-repository/publications/economic-reports/airline-fuel-labourcosts/
Energy scales with the square of velocity, so if you are going 3x as fast you 9x the energy, so 9x the fuel to reach cruising speed. But drag scales with the cube of velocity, so you need 27x the energy to keep yourself at that speed. You can reduce that a bit with more aerodynamic design and by flying higher, like Concorde did, but fundamentally supersonic flight is going to need a lot more fuel. That's where the extra cost will come in.
This is a little chicken-and-egg, no? It's less cost-efficient to do supersonic passenger flight, at least in part, because we haven't devoted billions and billions of R&D dollars into it in the way that we have for subsonic passenger flight.
Also, the fighter jet comparison isn't useful here unless you can show two things:
1. Military passenger aircraft have similar cost per flight hour per pound to civilian passenger travel
2. Subsonic fighter jets didn't cost more to fly per pound than subsonic passenger aircraft
I'm pretty sure 2 isn't true but I don't have any data I can point to, and I'm certain that 1 isn't true.
Yeah, to be clear fighter jets are a directional gesture, they're more expensive to fly for multiple reasons.
We can use the Concorde as a better comparison baseline (it had a *lot* spent on development costs). It's nowhere near fighter jets costs but it's still about 10x the cost per passenger mile as a 737.
In my opinion the market for scheduled supersonic flights is a bit limited because it competes with both subsonic first class and subsonic private jets. Supersonic which still requires getting to the airport 2.5 hours early does not save much time over a private jet which requires getting to the airport 10 minutes early, and there is yet more wasted time in having to fly on the airline's schedule. And you can have meetings and get work done on a private jet, so your flight time is productive. And far more comfortable.
Compared to first or business class, you'll have more departures to choose from subsonic, and a much more comfortable work environment.
That leaves private supersonic as a potential market, which would be viable I think except that politically it would be a non-starter in a lot of places.
The other problem with Concorde was short range -- it did the trans-Atlantic hop which only takes five hours subsonic, not long enough to be annoying (especially in first class).
A supersonic jet which could do a 14-hour trans-Pacific flight in six hours, that would be more tempting. But not necessarily physically viable, let alone economically.
The SpaceX concept videos showing Starship doing point-to-point travel, anywhere to anywhere within an hour... that might actually be a more viable future than trying to push through the air at Mach 2.
Yes, and especially with things like Netjets and JSX democratizing and making the private flying experience cheaper, it'd be really hard for supersonic to compete.
I really want to understand the psychology and motivations behind people who block everything when there is no obvious financial interest.
There is a whole ecosystem of activists and regulators, who block houses because of their effect on pond life or solar farms because newts don't like them. They must be self aware enough to know those are ridiculous and it is not as if the newt lobby is funding lawyers and activists or influencing officials.
You would be surprised how high up the conspiracy goes: https://t.ly/EaQaM
A newt?!
Perhaps he'll get better. https://www.youtube.com/watch?v=ZB5ig6vpQug
Most people do most things for interests that aren’t immediately financial. If someone is doing something that you see as a threat to your way of life, you’ll want to oppose it, even at some cost. Doesn’t matter whether that’s killing an animal, or putting a shadow from an apartment building on your vegetable garden, or teaching your kids that they can change their gender, or trying to replace your government.
I wasn't only talking about financial interests. I can understand situations where there is some tangible benefit to a human or charismatic mega fauna but there has to be some limits to circles of concerns.
Regrettably, some people simply like power over others, and that can be a substantial motivation in otherwise senseless blockage actions.
Because some people believe pond life has value. You don't, that's cool. But they do. And it's profoundly unhelpful, this constant reversion to "these people want things for no reason."
No one is a nihilist. They just don't share your values.
A non zero share of people are nihlists.
> No one is a nihilist.
Speak for yourself.
I believe nihilists can't exist, and that they believe this, too.
But I've seen nihilists. They killed Donnie.
They can be wrong because of facts they don't know or have failed to consider, or they can be wrong because their values are wrong. I guess you're saying their values are wrong, and that is why they are choosing a pond over a solar farm. I prefer the hypothesis that their values are good, but they don't know the facts, or they have failed to consider them.
Some people actually do value the environment. And some people are cynical nimbys who declare a laundromat or parking lot "historic" so it can't be built on.
Because it's easy enough to think that there's no harm done and that the houses or solar farms can simply be built somewhere that *doesn't* have newts.
Also, if you're the chosen and known Newt Crusader, you basically have to fight every newt battle there is, no matter how trifling, because there's a little voice at the back of your head saying that if you start going "Well, maybe the newts can take it on the chin this time a little bit", then the screw-the-newts people will go "Look, even the Newt Crusader agrees that the newts can suck it! Kill the bastards!" ("Newts" can, indeed, be a great deal of things here - such as abortions, guns, AI models, countries within a military alliance etc.)
My sense is that mostly people object to things over aesthetics, and the pond life or newt stuff is just a tool to block the thing, not the actual reason for wanting to block it.
I have also often encountered the mindset of "if it doesn't benefit me, then I am against it" — basically, the attitude that one's political power to object can be used to extract some concessions of whatever kind.
Giant airships are only awesome and inspiring if you know that a person on board is in command.
Self-driving giant airships are sad.
What if you can get giant airships to autonomously hunt down other airships, to steal its cargo or something if you want a purpose more than the plentiful lulz? Robotic sky piracy sounds pretty awesome.
https://store.steampowered.com/app/115800/Owlboy/
Contains literal robot sky pirates.
If those giant airships are moving hundreds of tonnes of cargo faster than ships and cheaper than airplanes, are carbon neutral, and are being moored, loaded, and unloaded robotically, that seems truly wondrous.
Though I'd want a manned one for my own personal use also, of course. One that's nuclear powered and can dispatch a submarine.
I love the concept but I don't think the economics pencil out. They're only a little faster than ships at significantly higher cost, and it seems like ships generally could go about that much faster if they needed to but it isn't worth the extra cost. They can reach places ships can't, but there they have to compete with both trucks and trains. There are probably a few places where you could make it work because the transport infrastructure is bad or nonexistent.
Forget the flying cars, bring back airships.
And the associated fine dining, of course! Slow food as it takes at its fastest two days to travel between Europe and America, but who needs to be in a hurry when travel is this luxurious?
https://www.youtube.com/watch?v=s8BqxWbX_r8
https://thehistorypress.co.uk/article/zeppelin-hindenburg-transatlantic-workhorse/
"Mary Day Winn of the New York Herald Tribune wrote: ‘ The real glory of Zeppelin travel … is its freedom from seasickness. It is the smoothest form of motion I have ever known, just a continuous floating, with no rolling, no dipping, and almost no change of levels. The sound of the engines can be heard only faintly – a low, steady murmur barely entering consciousness except when it slows up. There is no vibration.’"
The obvious tourism market for Airships to me always seemed to be air cruises over land. Instead of being slower point to point travel (where it is always going to lose out to the convenience of fixed wing jet aircraft) have an air cruise over the Grand Canyon stopping off at major tourism points like a cruise ship and so on. There is clearly a demand for cruise ships but not for ocean liners, so why not do the cruise ship market but in places with no ocean! If I had access to the huge amounts of capital that would be required that would be my business venture.
I think that would be a great idea, you could catch the 'luxury travel experience' market, and it would be 'green' travel by comparison with airplanes and helicopters.
The main problem is the safety factor; the Zeppelins had a great safety record right up until the moment they didn't, and since the Hindenburg was destroyed in minutes, that isn't appealing on grounds of safety considerations.
If the problem of "how do we ensure the giant flammable disaster waiting to happen doesn't happen" could be solved, then Land Cruises on the luxury blimp could definitely be a market! As the "Tasting History" video explains, the low altitude the blimps travelled at meant that passengers could almost reach out and touch the sights, and the smooth motion meant no turbulence or similar problems.
Imagine being able to cruise along *inside* the Grand Canyon like a drone, for instance:
https://www.youtube.com/watch?v=4Y8eHOp9ah4
I was under the impression that the safety issue was solved with helium although the greater required size of helium airships may make them uneconomical I suppose.
I thought so too, but as you say, nobody seems to have picked up on the untapped market for airship cruises. I do think the spectre of the Hindenburg disaster hangs over the entire area, the same with nuclear power and the bad reputation it has due to several incidents at power plants.
But lo! Out of curiosity I looked it up online, and at least one brave pioneer is doing this very thing!
https://oceanskycruises.com/
Prices are a bit eyewatering, but if you're sufficiently flush with cash, why not? And this is Swedish, so I'm sure American entrepreneurship could work out a way to cut down prices to the middle-class budget (less luxury but still something special). Say $5,000 per person but you get everything included (meals, drinks, etc.) for a seven-day land cruise. Sounds good to me! Something you'd save up for, or that the entire family would chip in to give Mom and Dad (or Grandma and Grandpa) that holiday trip of a lifetime.
"The demand for our North Pole Expedition has exceeded our best expectations and we are now only accepting requests for interest on a waiting list mode for our Adventurer departures. However, we still have limited availability in our Pioneer programme which also secures your place within the first departures while also giving you the opportunity to invest in us. Cabins for our North Pole Expedition start at 2.000.000 SEK (~$200,000 USD) and departure dates are subject to the manufacturers’ calendar.
What is included and excluded from the price?
Included: roundtrip transfer in a snowmobile between your hotel in Longyearbyen and the airship, 2 nights onboard the airship in your private cabin, services of a professional expedition leader and crew on board, meals, beverages (alcoholic and non-alcoholic), activities and guided excursions and related equipment.
Not included: flights to/from Longyearbyen, accommodation in Svalbard, airport arrival and departure taxes, passport and visa fees, travel insurance, medical evacuation insurance, personal items, optional activities outside the main North Pole Expedition program (pre and post programs, additional ground services in and around Svalbard), gratuities"
Wow you weren't kidding about it being eye watering! 5000 seems like a good target, more than an ocean cruise (while its still a novel technology) but not so much so that its out of reach for the middle class.
A brief stroll through the history of rigid airships, almost all of them not named "Hindenburg", will show that the safety issue is mostly that the things are too big and flimsy to survive major storms, and can't fly high enough to pass safely over them. I am not aware of this problem having been solved.
If you just want to use it for tourism, that might not be a dealbreaker.
Unfortunately, helium supplies are very limited and expensive.
The rarity of extractable helium is a pretty tricky thing to actually know because of the US Strategic Helium Reserve sell off vastly suppressing the price and making it uneconomical to extract and thus we just release huge amounts from things like natural gas wells and don't even attempt to exploit it. Its fairly similar to the articles about all of the worlds Rare Earth Metals being in China when the reality is there just wasn't really a business case to survey deposits until recently.
The use of helium in dirigibles would be an absolute drop in the ocean compared to modern uses of the gas though. Its like when people complain about aviations carbon emissions, despite seeming like it would be a massive issue the inherently small market makes it miniscule.
According to https://www.federalregister.gov/documents/2023/07/24/2023-14425/modernization-of-special-airworthiness-certification I think it might *technically* be possible to get a thousand-ton zeppelin classified within the recently expanded category of "light sport aircraft," so long as it didn't have cabin space for more than four people, and met some other standards.
Once the bugs are worked out, and revenue from superior door-to-door air freight service has recouped development costs, just swap out the cargo bay full of intermodal shipping containers for some modular luxury apartments.
"revenue from superior door-to-door air freight service has recouped development costs"
They're talking about doing that in Canada - it does have a problem with "large areas of the country are inaccessible except by aircraft, everything has to be flown in, a head of lettuce thus costs $15 (apparently Canada had a lettuce shortage crisis due to having to import lettuce from California which was hit by drought and bad crops, resulting in shortages)".
https://financialpost.com/commodities/agriculture/how-to-fix-canadas-lettuce-problem
"In the last few weeks, Dean has considered himself lucky if he gets a quarter of the cases he normally needs, and actually, that’s fine, because at these prices — $6.99 for a head of iceberg, $14.99 for a bag of three romaine hearts — nobody’s buying it anyway.
Article content
“The price points are insane,” said Dean, who runs Mike Dean Local Grocer Ltd., named after his dad. “I won’t put a margin on it because we’re already looking ridiculous.”
The reason for the shortage and the ridiculous prices, at least this time around, is Impatiens Necrotic Spot Virus (INSV), a bug-borne disease that has wiped out a massive chunk of the lettuce harvest this fall in California’s Salinas Valley. The region, known as the world’s salad bowl, grows more than half of all lettuce produced in the United States. This season, however, lettuce producers reported losing as much as a third of their crop, according to the Grower-Shipper Association of Central California."
If lighter-than-air craft can bring down transportation costs, they'd be very happy about that.
https://oceanskycruises.com/airships-for-cargo-transportation/
"What are the Problems?
Transport is the main challenge in bringing goods to remote areas. Goods are either transported by trucks over winter roads to inland communities, by seasonal summer voyages to coastal settlements, or by air freight, which is the only year-round cargo service. All three of these options are costly, not always reliable and can be limiting based on weight and weather.
Due to climate change, these forms of transportation have been drastically affected. In the last decade, the availability and reliability of transport has diminished, and northern communities are suffering. The earth’s temperature continues to rise, which shortens the winter roads season, and creates unpredictable sea and weather conditions.
This is only going to get worse as the years go on. The remote communities are desperately needing another solution for the transportation of goods."
You might be on to something here JamesLeng! It could be well worth your while to look for investment opportunities!
(And if the airships don't work out, you can always fall back on lettuce farming for the Canadian market 😁)
Well, I can't fairly take credit for the idea: https://www.elidourado.com/p/airship-industries
To be fair, this was written in the era of very noisy and very shaky propeller airplanes. Modern jets are incomparably more comfortable.
I don’t quite buy the story of “perpetually losing for 50 years”. Why didn’t the highly optimistic, pro-growth, anti-government-regulation attitudes of the Ronald Reagan administration manage to put a dent in any of this during the 1980s? Reagan had a pretty astounding popular mandate and hated big government regulation.
I think the image at https://substackcdn.com/image/fetch/w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F848f0ce8-2637-4fde-9372-3cf9e9b1cc37_1200x900.jpeg is pretty explanatory here.
Stuart Buck, above, made a compelling argument that the number of pages in the Federal Register may be an exceptionally poor proxy for "amount of regulation."
It looks like you’re right about this — Reagan didn’t do all that much deregulating, even though he talked about it, it was kind of a neglected part of “Reaganomics”.
You might be interested in Yglesias' recent posts on neoliberalism and the era in which it rose to prominence. This is freely available: https://www.slowboring.com/p/what-was-neoliberalism
and this one I haven't actually read in full because it's paywalled for subscribers: https://www.slowboring.com/p/shackling-the-state
I always feel like Russia will not be able to go solar - we're too far north, and have too many cloudy days where even the most efficient solar panels will produce far below their top capacity. On the other hand, Russia is pretty much the only large country in such position, so if everyone else goes solar, we may continue to burn fossil fuels and use nuclear power (as well as use hydro power, and maybe geothermal in some rare places) for a long time without much negative effect. Economically, of course, solar revolution in the rest of the world might be a disaster for Russia, but on the other hand, oil is needed for more things than energy production, and with cheap energy, the demand for it might not fall that much.
Another thing of note: I often hear about energy prices going negative in Germany, but I also read that metallurgical and chemical companies (BASF, primarily) are closing plants in Germany, in part, because of cost of electricity (after loss of cheap Russian gas). Can anyone explain to me how this two facts can co-exist?
BASF needs consistent energy prices. If energy prices are negative 10% of the time and 50% higher than they were in 2019 the other 90% (or some other distribution with higher mean and higher variance), they lose out.
They even lose out a little if the mean stays constant and the variance goes up since it makes their expenses less predictable (though I doubt that's the main effect here).
Basically power prices in Germany are negative *sometimes* (ie on a sunny or windy day) but they're high on average. This impacts the economics of running a high-capex factory like BASF's, which needs to be run more or less constantly.
However it's also bad for the economics of wind and (particularly) solar, because the negative prices happen when they generate.
It's something we will figure out ways around.
Germany built a lot of renewable energy infrastructure when it cost much more, I think their electricity price is triple the fossil grid. A high renewable system also needs a lot of transmission and backup
Germany is locked in to high prices for some amount of time while the grid effect means that it would take radically cheap batteries to unlock much cheaper whole system costs
They are actually coupled. About 20 years ago, Germany committed to pay very high prices for electricity produced by solar panels. The deal was risk-free and high-gain for investors, so all companies in the world stopped whatever they were doing and built solar panels for Germany. instead. (Ok, I *am* exaggerating here, but only slightly. It completely spiraled out of control. I think the cost for German electricity consumers has been something like 200 billions.
The effect is twofold:
- The solar companies will be paid for the electricity from their 15-20 year old panels, regardless of market prices. Essentially, the German government pays them directly. This means they never have incentives to switch off production. This sometimes leads to negative market prices for an hour or so, because even then they have no incentive to switch off production.
- Afterwards, the German government collects the money from all consumers of electricity. So companies like BASF pay twofold: first, the market price. Second, the insane amounts of money that the German government pays for 15-20 year old solar panels. The second part is what makes electricity much more expensive in Germany than in other countries.
Fortunately, the subsidies were only guaranteed for 20 years and will run out in the next 5 years or so. It's not even a big share of all installed panels in Germany, so it was a *really* stupid policy 20 years back. (And it did not look like a good idea even back then.) But on the other hand, at that time Germany was pretty much unilaterally responsible for a world-wide explosion of production of solar panels, all of which were installed in Germany. Without that, mass production of solar energy and the rapid decline of costs may never had started. So good for the world, I guess.
This post talks about some applications for abundant energy: https://caseyhandmer.wordpress.com/2023/10/11/radical-energy-abundance/
I remember reading a much more comprehensive list of ideas, but unfortunately can't dig it up.
I also recall a convincing blog post on the limits of reducing the cost of solar energy. It argued we're near the end of the exponential because the cost of mass-producing modules will stop being the limiting factor (other costs like land, installation, power electronics, maintaining the grid, etc. will dominate).
Yeah, but this is ridiculous. Casey Handmer is a dreamer, not an engineer. Terraform Industries doesn't have a tech to produce liquid fuel directly from solar, otherwise he wouldn't need to write with so many unsourced charts.
The reason that the cost of solar power doesn't go to zero - or to $10/MWh - is indeed that the cost of solar panels is now a fairly small part of the cost of building a solar plant. In markets that don't have import tariffs on China (ie not the US or India), a solar module currently costs about 10 US cents per Watt. The total capex of even a large solar plant, however, is in the range of 40-60 US cents per Watt - because this has to include mounting systems, racking systems, the inverter, at least some grid connection costs even if you get lucky and get a perfect site right next to existing transmission, and then land and arrangement costs.
Solar's going to be cheap but not that cheap.
If racking/mounting is a big part of the cost, could you just wrap the modules in a more rugged coating at the factory and then have the install just be dropping them on the ground? Since the modules themselves are getting so cheap, slightly lower efficiency per-module doesn't seem very important.
There's a company called Erthos trying to do this. I'm not a fan.
The design lifetime of solar modules is over 25 years, so you do want them to be installed on a structure that will fully resist wind and allow air cooling on the backside, access to the junction boxes and cabling, rain cleaning, and vegetation management. Depending on the specifics of your environment, it's likely that just plunking on the ground will end up with them blown away, buried, or covered in plants within a few years unless you're constantly working to prevent that.
There are some designs of system for flat roofs where the modules basically come on plastic trays you can with rocks to keep them in place - but the roof then needs to support that.
I know the comparisons are going around comparing PV panels with roofing material (but comparing high volume bulk purchased PV with consumer retail roofing).
But given the mechanical needs of a PV panels (resist weather and wind and heat, while staying sealed), my mental model is we shouldn’t expect panel costs to drop below wholesale bulk roofing material costs. Even if the cells were free.
And they need to be mounted about as well as those same roofing materials or they will blow away…
The elephant in the room is China. The current emphasis on progress seems like a precursor to the 1990s competitive mode, where Silicon Valley chip companies were "paranoid"¹ about Japan taking over, thus ensuring that the Internet boom happened in Silicon Valley and not in New England or somewhere in Europe or even in Asia. Progress studies is a mislabeling of what is ultimately competitive studies.
You can't defeat the God of Straight Lines, but every half-century, you'll find that a declining Empire will have an efficiency decade with some Thatcherian soul-searching and belt-tightening. For a time, some important numbers will reverse. Never in a million years would I have believed the U.S. would return to being the top oil producer and reduce its dependence on OPEC, for example.
YIMBYs are winning? As far as I can tell, YIMBYs have turned California into Austin (or is it the other way around?) by expanding low-density housing, thus making it more attractive to yuppies, which inevitably raises housing prices, therefore bringing more homelessness.
[1]: "Only the Paranoid Survive" was the famous mantra of the former CEO of Intel, Andy Grove
"Self-driving taxis have a big advantage over self-driving self-owned-cars: they can operate 24-7 and never have to park."
Really? So where you live, demand for taxis is the same for each time of day? No rush hours in the morning and evening? Or do you plan on having the idle taxis driving around aimlessly, clogging the streets?
I think the point is that they don't have to park right next to the places people want to go, so you can put in some concentrated parking facilities and not need to have mandatory parking next to your house *and* next to your job *and* next to the shopping mall. This is similar to how buses or school buses (or taxis!) don't really contribute the problem of finding a parking spot during rush hour.
There may be some advantage to that, but you need to factor in the additional traffic by the taxis coming and going to their centralized parking (im addition to driving around empty much of the time to pick up the next passenger). Also, they're essentially still cars, so they take up just as much real estate per passenger as cars, while buses take up much less space.
Edited to add: I don't doubt that robotaxis would reduce the demand on parking space compared to individually owned vehicles to some extent, because fewer of them would be needed - how strong that effect is depends on how strongly pronounced rush hours are. I don't think a massive rollout of robotaxis would reduce overall traffic - quite the opposite.
Two points:
1) There is NO world where driving automation doesn't make congested streets MORE congested. Some hype artists like to claim otherwise but it's just not credible: so long as we don't have Star-Trek-style transporters, physics and geometry mean that roads are finite and more activity will congest them. And automated driving makes using roads easier at the margin.
2) We should be excited about the application of driving automation to traditional transit. If every bus was automated, AND it had its own right-of-way, you could run buses at subway frequencies if you wanted, and affordably too (modulo your city's definition of 'affordable').
If the fleet of robotaxis involves a dispatch server keeping track of where they all are at all times, and all the other cars which any of them can see, using that data to plan routes and calculate appropriate passenger fares, it would be fairly trivial to implement precisely calibrated congestion pricing.
1.) Sure there is. The world in which a large part of traffic problems are variances in human driving habits.
In Dallas, back in the ... 80s? 90s? It has been a while. Anyways, in Dallas at some point in relatively recent history, highway traffic had a pattern that was baffling analysts: Every, let's say it is Wednesday, at 2:00 PM, there was a massive backup on a key highway. (The exact details escape my memory and aren't important.)
They studied the problem for a few months, sent up helicopters to watch the traffic. And, after extensive expensive analysis of the problem, they figured out what was going on: An older driver had a weekly hair appointment, got on the highway going SIGNIFICANTLY slower than all the other drivers, and created a traffic problem.
To solve the problem, they sent a driver every week to take the driver to the hair appointment. And their mysterious traffic problem vanished.
This is not a small thing; traffic is full of variations on this.
Traffic is also full of coordination problems that automation can provide a path to resolving. A car does something dumb, everybody brakes, and now you have a wave pattern embedded in the moving traffic which is wasting fuel by making people brake for a hazard that hasn't been there in half an hour.
Huge amounts of traffic management is basically making sure that the traffic is sitting in a place that doesn't cause downstream problems - preventing gridlock. If the cars themselves can be coordinated to avoid gridlock, this problem goes away.
A couple of jobs back, there was a massive traffic snarl in my daily commute caused by an on-ramp that went for a mile before turning into an off-ramp; if you got there early, people used the exit lane to pass slower-moving traffic. After ten minutes or so, this turns into "people use the exit lane to pass the traffic brought to a stop by people using the exit lane to pass the traffic".
Nevermind basic driver inattention - a driver who misses their turn and has to do two u-turns is an additional load to the system.
There are many worlds where automation can easily make congested streets less congested, and we live in one.
Two things:
1) You seem to be describing a world where ALL cars are automated, so there are no human drivers to occasionally pump the brakes when they shouldn’t. Sure, in THAT world, there’s potential for improvements, but we don’t live in it now… and I’m not certain that we ever will. If we do ever get there, no one driving today will live to see it, so it’s academic.
2) You also seem to be describing a world where making driving easier through automation changes nothing else. But of course that’s not so: make it easier on the margins, and people will do more of it. It’s supply and demand… or more precisely induced demand: https://en.wikipedia.org/wiki/Induced_demand
I’m no fan of congestion either, but it’s a necessary by-product of automotive trips. We can mitigate it, but never eliminate it, nor even signficantly reduce it (at least not in prosperous areas).
1.) I was responding to a specific person making specific claims. If you want to argue about -entirely different- claims, we can have that conversation. But even if you want to talk about -just these- claims, note that nothing I said requires every driver be automated, as limited coordination is still superior to no coordination.
2.) Question: Are there currently, right now, empty roads with no drivers? Why? Shouldn't they be filled by induced demand?
Obviously not; demand is not, in fact, perfectly elastic, consuming every inch of available road. Indeed, I spent ten years in a city which had lost a substantial part of its population base, which, combined with infrastructure built for a much larger population, made driving incredibly pleasant there. (Excepting areas of new construction, which were built with the newer, lower population in mind - which resulted in traffic concentrating in newly-constructed areas.)
Induced demand applies when there is a greater demand for transportation than there is available supply. Sometimes you can't reduce congestion, because every improvement results in more drivers - but this is not, in fact, a universal phenomenon, and even where it does arise, the demand is not, in fact, infinite.
More than this, however, even if an improvement merely results in more people being transported - this is still an improvement! It's a huge improvement! People are getting places they wanted to be, which they previously could not get to!
You can't compare buses to cars because buses aren't a realistic substitute for cars.
Very few people choose to take buses if they have the option of simply driving there, because buses inevitably go from somewhere that isn't exactly where you are to somewhere that isn't exactly where you want to be at a time that isn't exactly when you want to go, in the company of people you wouldn't choose to hang out with.
A self-driving taxi *is* a reasonable substitute for your own car so it's a fair comparison. There would be some need to construct additional infrastructure to cope with the huge influx of self-driving cars into the city centre in peak hour, but it's a lot cheaper than building trains.
Elena is right: dedicated facilities for robotaxi maintenance, storage, and charging are a plausible outcome (they won't ACTUALLY go 24/7 though the business incentive will certainly push as close to that as possible).
Street parking is also possible, since the robotaxis will obey whatever the local parking ordinances are... but street parking is antithetical to what makes cities great, so this would be something to discourage.
At some point—which won't be in my lifetime, but a man can dream—the proportion of automated vehicles as a percentage of total road users will be so high that robotaxis might simply park on freeway lanes between, say, 0200h and 0500h; just like street parking, but not in a place where they discommode anyone.
In the distant future, when automated vehicles are
I'm not sure why you think that "street parking" is bad but "parking on the side of a freeway" is good. That's basically the same thing.
Are they? Street parking impedes pedestrian and cycling use. Freeway parking, were it allowed, would not.
As other people say, you can always have them drive to a parking lot way outside of town. But there could also be more creative solutions - maybe you can swap out the passenger chassis for a freight chassis + seat for one deliveryperson, and Amazon could use them to deliver packages overnight.
Actual deliveryperson seems inefficient. Could probably be replaced with a robot arm for shifting cargo within or immediately adjacent to the vehicle, plus recharging cradles for a few quadcopters capable of carrying, say, a 30lb box thirty yards, then ringing a doorbell.
We can do better. Once you've got the robot arm, just have it throw the packages in the general vicinity of the destination. You often wouldn't even have to slow down!
For some applications that'd make sense, but a lot of parcels are going to be fragile, or have significantly nonuniform density, high drag relative to their mass, or other inconvenient ballistic properties, and a lot of destinations will have awnings, pedestrians, or other obstacles. Also, an arm rated for throwing (alongside other cargo manipulation, so it can't be a simple trebuchet) would be a significant engineering challenge, whereas picking things up and setting them down more carefully can be adapted incrementally from warehouse bots.
Acting as a mobile base for quadcopters already covers most of the "without slowing down" benefit. Truck might dispatch a flock of short-range drones into a residential subdivision while cruising past on the main road, each drone with a package for a separate house, and drive on to complete other deliveries before returning to pick up the drones minutes, hours, or days later, or allowing them to find their own way to some other collection point - made easier by the fact that they only had to cover a short distance with the heavy payload.
Sure, that's the "safe" and "legal" way to do it, but wouldn't you much rather "Move Fast and Break Things"?
Breaking the things you're being paid to deliver, and the windows and bones of people you're delivering them to, is not a viable business model even in absurdly favorable regulatory contexts.
I think the tfp plateau from 1970 onwards is just homeownership in the classic yimby/Georgist sense that promoting speculation of limited assets leads to innovation slowdown. But environmental and labour regulation also had some effects. I think we could get the productivity back with sensible regulation if we had a lvt/yimby/densification boom
TFP measures GDP growth vs. inputs and is inflation adjusted. I think this calls into question how accurately we define inflation. As an example think about software like video games. Their price has increased hugely over the last but their complexity has increased orders of magnitude. One could argue that inflation should be considered negative for many types of software. What about the reliability and efficiency of automobiles? The decision of what to consider inflation and what to consider as actual growth seems critical to the measurement of TFP.
HAH! Should have read further before commenting...
This objection was discussed at the conference, but I think answered by a graph showing that manufacturing productivity has been stagnant since 2005; that seems harder to fudge.
I am not sure what this means though. I would not be too surprised if the manufacturing productivity of product X had stagnated. If the product is profitable and the market isn't growing then there may not be tremendous incentive to improve productivity. On the other hand, if the measurement is on a continuously changing mix of products then it seems that there is ample opportunity for errors. In my own, semiconductor field, production is now transitioning to 12" wafers (4" was the norm when I started in 1991). The transition from 8" to 12" alone results in a 30% increase in output (die) for the same cost (productivity). I just feel that there is plenty of room to be skeptical about these types of measurements, both in there accuracy and in their applicability.
Not 30%. Close to 100%. It’s area, not diameter.
It is a 125% area increase but 30% cost savings. Masking time is unchanged. Silicon, dopant, etchant and photo resist usage and process times are all unchanged. The exact cause of the cost savings isn't entirely clear. I guess, the building itself, as well as, implantation and diffusion steps yield 125% more die per (each). However, quite a lot of equipment must be updated to 12".
I see what you mean, agreed.
> As an example think about software like video games. Their price has increased hugely over the last but their complexity has increased orders of magnitude.
Has their price increased? My sense is that the nominal price of games has stayed pretty flat over my lifetime, which means the real price is dramatically lower after accounting for inflation despite quality and complexity both going up.
I suspect that a lot of the quality of life impacts of tech/software/the internet/etc are dramatically underestimated by these kinds of statistics since they're generally cheap or free.
I am guessing that my gaming career started prior to yours. I think high end games were $29, back in the 80's despite coming on 12 floppy disks. You are absolutely correct though that this means decreased price in inflation adjusted terms.
Yes that was my thought as well - video games have famously resisted inflation
Yeah, a standard AAA game was $60 for a very long time (I think it started with the XBox 360) and has only just now risen to $70.
Regarding the “Public Transit” section: This is what everybody who tells you to “get tough” on homelessness means. Getting rid of the people who suck makes things suck significantly less. It is very easy to set up a simple test that eliminates the people who suck. You can just charge people money. Homeless drug addicts who can’t hold down a job and constantly commit crimes don’t have any money.
In my dream, cities replace all existing taxes with a flat per-day fee for existing within city limits. You could call it a "congestion fee" except it would apply to people rather than cars.
San Francisco has a $15.9 billion budget, and roughly 1 million people in the city on any given day, so the fee would need to be $43 a day. Residents wouldn't mind, because it would replace all their existing taxes. Tourists and business travellers wouldn't mind, since it's a small fraction of the cost of a hotel room. People who commute from outside city limits *would* mind, but you could probably have some sort of reciprocal arrangement for people who live in neighbouring cities and pay their fees there. The only people who would mind would be the people who currently live in the city without paying taxes.
Mandatory crime insurance would be a little like that https://www.overcomingbias.com/p/who-vouches-for-youhtml
Just coming here to be an annoying crank and remind everyone that 1971 is the year the US fully ditched the gold standard and started our modern freewheeling monetary policy experiments.
This comment made me laugh
>The United States could produce all its power by covering 2% of its land with solar panels - for comparison, we use 20% of our land for agriculture, so this would look like Nevada specializing in farming electricity somewhat less intensely than Iowa specializes in farming corn.
I think this actually understates the case. The US uses ~3TW of primary energy. Fully electrifying everything would probably increase that somewhat, maybe to 5TW or even 10TW depend on whether we manage to be smart about it. 2% of US land area covered in solar panels would be 200,000 km2. At current panel efficiencies and typical capacity factors, that would give us 50TW, 10x what we need. Beyond that, there is growing (but much slower than I'd like) awareness of agrivoltaics enabling dual land use. This seems like it should be a tradeoff, but in many case the result is higher yield and lower water consumption for crops, plus electricity production to increase farmers' profits. For some crops, research is showing the extra shade improves nutritional quality. This gets even better if we manage to get the kind of semitransparent super cheap perovskite cells as standalone modules instead of how people are starting to use them now, which is as the top layer in tandem modules, because then they can go directly above crops instead of between rows without impact availability of the frequencies plants can actually use.
And I see someone mentioned the value of desalination as a use for cheap energy. I'd add wastewater recycling and indoor farming as other clear use cases. "We're losing cropland to climate change" becomes a less severe threat if you can say "We can drop a ten story building anywhere on the planet and power the HVAC and grow lights and water recycling and water supply with nearly-free electricity, with a year-round growing season and 2-10x the yield per floor area of a typical outdoor farm."
I agree with your point, but its even stronger than you think: electrifying everything would *decrease* primary energy. Primary energy includes fuel. Electric vehicles are much more efficient than combustion ones, and heat pumps are 5x more efficient in primary energy use than nat gas furnaces. We can likely *decrease* primary energy use by electrifying, not that we necessarily should.
Oh, for sure. Primary energy consumption will decrease for a whole lot of things when we electrify. In other cases it'll be nearly one to one. In the remainder it might be a rather significant increase, especially in some chemicals, fuels, and industrial processes, at least initially.
I suspect in the short to medium term the latter will balance out or outweigh the former, and in the long run we'll Jevons Paradox ourselves into much higher total energy consumption by the time we fully decarbonize. For example, most aviation (which is kind of the worst case scenario application) is still going to be kerosene based for at least the next 20-30 years, there's not enough biomass sustainably available to make SAF so we'll need a lot of electrofuels, and that alone could easily eat up the equivalent of 60% of current total world electricity consumption to eliminate 2.5% of current world CO2 emissions.
>Self-driving taxis have a big advantage over self-driving self-owned-cars: they can operate 24-7 and never have to park.
But demand varies over the day, right? Like, either you have the number of cars needed to meet demand at 3am and almost no one can get one at peak hours, or they are all wasting fuel and creating traffic to drive in circles all night, or they park somewhere?
I would guess the answers to that are either 'they can go park outside the city where space isn't at a premium' or 'they can be converted to other use at night such as package delivery'. Those seem plausible to me intuitively, but I wish I knew what the experts actually-thought-about answers were to this question.
As the expert Scott spoke to, I can say that both answers you offer are plausible but the former is more likely. Regarding the latter: even today people have experimented, but it turns out a vehicle optimized for moving passengers is NOT optimized for moving goods, and to try to do both is to do neither well.
Regarding the former: 'where space isn't at a premium' might be IN the city; design a parking garage for robots to navigate rather than humans and you can be MUCH more efficient. And such garages will be needed in any case for charging (they'll all be electric, of course) and for maintenance: human nature being what it is, robotaxis will likely need to be cleaned and groomed a few times a day.
Perfect, thank you, that is exactly the level of analysis I was looking for :). Much more efficient parking that also does charging/maintenance makes a lot of sense.
They can park in my driveway overnight if they like, in exchange for a five-dollar voucher for my next trip.
Overnight parking really doesn't seem like a big problem.
MattY on the case for more energy (https://www.slowboring.com/p/energy-abundance) does some work to answer the "what would we do with cheap clean limitless power"
What happened in 1971? Looking at the graphs, the key date could actually be a couple of years later, in 1973. That's when the price of oil suddenly tripled and stayed high: https://ourworldindata.org/grapher/crude-oil-prices. I guess that had a major, enduring effect on the economy.
When you adjust for inflation, prices go back down to pre-1973 levels by the 80s, see https://inflationdata.com/articles/inflation-adjusted-prices/historical-oil-prices-chart/ .
Yes, that graph certainly looks like the inflation-adjusted price of oil fell back to nearly the pre-1973 level. The graph I'd used for inflation-adjusted oil prices looks meaningfully different, though. It's from the BP Statistical Review of Energy, so I hope it's reliable, but I hadn't linked to it above because it requires downloading the pdf, but here it is: https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/energy-economics/statistical-review/bp-stats-review-2022-full-report.pdf. It's on page 26. The inflation-adjusted price stayed over 3x the pre-1973 price for around a decade, then fell to about double the pre-1973 price in the mid 1990s, which interestingly is when productivity started to rise a bit faster.
https://en.m.wikipedia.org/wiki/1973_oil_crisis
Thank you for the link with several inflation-adjusted graphs of oil prices.
"Once you convert a problem to 'let’s manufacture billions of identical copies of this small object in a factory', our natural talent at doing this kicks in, factories compete with each other on cost, and you get a Moore’s Law like growth curve."
More canonically this is called Wright's Law, or "learning curves", whereby each doubling of the total produced units of something tends to produce a similar ratio reduction in the cost. Solar panels go down about 20% in price every time we double the total number of produced units:
https://ourworldindata.org/learning-curve
Solar panels are currently producing about 4% of US electricity, which can give a ballpark estimate for how much cheaper they will get. Assuming constant electricity demand, we'd be saturated with cheap electricity in five doublings (8, 16, 32, 64, 128) at which point solar panels would cost (0.8^5) or about a third of what they cost now, which is still really great in terms of electricity cost but not 1/50th of what it is now as alluded to in the post.
(If we did dramatically increase our energy consumption as the costs drop - which wouldn't be very surprising - we might squeeze out a couple more doublings of total capacity and therefore get down to, like, a sixth of the current cost. Much further than that in a short period of time would surprise me.)
Would this mean that solar would never get past 1/3 current cost, or just that we would have to double amount produced by waiting for time to pass rather than by scaling up?
Also, is this a totally different thing from Moore's Law, or does that also work by doubling the number of chips produced?
The best mathematical formulation for the solar module learning curve is that every time cumulative production of solar modules by our species doubles, the cost comes down a certain percentage. Sources usually put this between 20% and 30%; my calculations put it at the upper range of this (and currently some really weird stuff is going on with solar module prices don't ask).
This is a way to express that solar modules will never stop getting cheaper, but it's harder to squeeze out costs from a $0.10 per Watt module than from a $5 per Watt module, so the rate slows down over additional experience.
Moore's Law is a specific case of the experience curve which assumes that the rate of cumulative production doubling is really rapid, to give a time relationship. It's even more empirical than a straightforward experience curve.
In any case the big issue with reducing solar total costs is now the non-module components, which are mostly not simple manufactured things.
I think you are conflating number of installed panels with yearly production. Your "five doublings" would apply only if all of the 4% had been installed at the current "productivity."
> this would look like Nevada specializing in farming electricity somewhat less intensely than Iowa specializes in farming corn.
This is probably not actually what it would look like; there are losses in transmission, so it's better to have generation within a few hundred miles of the consumers. Of course this is offset by northern states having less sun, and some places having more cloud cover, but the difference between solar irradiance in the southwest vs. the north is only about a factor of 1.5, so it doesn't take much in transmission losses (not to mention the cost of building and maintaining the long-distance infrastructure) before it's better to make solar farms local.
> less likely to end with major cities being converted to radioactive ash
Nuclear reactors can't explode like a nuclear bomb. A bad meltdown looks like everyone in the city getting radiation poisoning, not a fireball or any buildings getting destroyed.
I'll edit it for clarity, but another problem with nuclear reactors is that terrorists can steal the nuclear material (although this may be more true in the minds of regulators than reality)
That's true. Also potentially more concerningly, rogue countries can use a civilian nuclear program as source material and plausible deniability towards building a bomb.
Its why we have to ensure renewables are as cheap as possible, the developed world is always going to be antsy about exporting dual use technologies (which nuclear infrastructure is) to unstable countries.
They can steal it but they can't do much with it unless they also have centrifuges to enrich it to weapons grade.
They *could* use gaseous diffusion, though it's far less efficient.
Even more remarkably, a bad meltdown looks like three gigawatt scale meltdowns at Fukushima, where it is conservatively estimated that 50 people would have died from cancers with no evacuation and 2000 did due to the evacuation
It is now safer in health terms to live in the exclusion zone than live in the relatively clean Tokyo city air!
For similar reasons the Chernobyl evacuation zone ceased to make sense after just a few years as the most dangerous isotopes all have a short half-life, by definition.
There are some isotopes which don't put out much energy, thus are relatively persistent, but end up concentrated by biological processes into vulnerable organs or tissues and thereby cause disproportionate harm. Strontium-90 substituting for calcium, and Iodine-210 in the thyroid, I think are the most infamous ones? Even those only last so long, though, and the fewer types are left, easier it gets to filter.
Strontium-90, yes. Iodine has a short-lived Iodine-131 (8 days) and a very long-lived Iodine-129 (16 million years) https://en.wikipedia.org/wiki/Fission_products_(by_element)#Iodine-127,_129,_131
Caesium-137, like Strontium-90, is a multi-decade half life isotope (30 years https://en.wikipedia.org/wiki/Fission_products_(by_element)#Caesium-133,_134,_135,_137 ) but
>Cesium is uniformly distributed throughout the whole body similar to potassium, and it does not accumulate in any one particular part of the body like iodine
https://www.ncbi.nlm.nih.gov/books/NBK594664/
Cobalt-60 is really nasty stuff too, with a 5-year half-life.
For what it's worth, people *are* currently working on battery-powered airplanes. Most likely, trans-oceanic flights will never be electric, but city-to-city shuttles easily could be. The gravimetric energy density of the most cutting-edge cells (silicon or lithium metal anode) is good enough that it's feasible, at least theoretically for small planes.
That depends what you mean by city-to-city. If you're talking 300+ mile trips, it is very very unlikely to be possible to get any service or safety level approaching what we have today. You'd need 8x battery energy density from what we have currently just to get a very limited capability aircraft which would never sell, and by the time we have that kind of energy density in batteries we will likely have hydrogen or sustainable aviation fuel alternatives.
There are some projects out there but nothing that isn't basically just fraud, the most egregious example being the Eviation Alice.
The other main advantage of hydrocarbon fuels in aircraft is that the aircrafts weight decreases considerably as the flight progresses. This is most commonly seen when aircraft that have to return to their departure airport after a moderate mechanical issue have to stay in a holding pattern to reduce their landing weight. This makes air travel cheaper as you can load an aircraft far beyond its maximum landing weight with the expectation that the weight will be lower when it reaches its destination. This is a pretty big barrier for battery powered jetliners.
Just to add: one way to assess whether a project is viable is by looking at how much the airframe without batteries weighs, and how much weight they are planning on taking off and landing with. So, an airframe like the ERJ175 weighs something like 50,000 lbs, and max takeoff weight is 85,000, and max landing weight is 75,000 lbs. So structurally they need a 50,000 lbs airframe to sustain landing at 75,000 lbs, for a 1.5 to 1 max landing weight to airframe weight. If they could land with more weight, they would either raise the max landing weight (to make more money) or lighten the structure (to make more money).
Most battery-electric airplane companies won't publish their MTOW/MLW or battery weight numbers. Eviation did have theirs published at one point, with a ratio of empty weight to MLW of 3.67:1, compared to the 1.5 to 1 for the ERJ above that is just obviously not possible.
Hmm, could you jettison the empty batteries (attached to a chassis with a parachute) once you reached cruising altitude?
That would add a lot of weight on takeoff with parachutes etc, and the battery packs would have to be very strong to withstand the impact, you’d need frequent inspections, and you’d need to retrieve the batteries after every flight which would be an enormous cost.
A better option would be an energy beam to power the aircraft from the ground so that your battery is just for reserve or cruise.
I have about 90% confidence level that Li-metal anode secondary cells will never be made safe enough for commercial use, much less aviation. The dendritic growth problem is an extremely tough nut to crack.
So I've spent my professional life studying solar, and I'm essentially a solar enthusiast, but it does sound like the ones at this event were not being intellectually honest. Solar cannot cost $1/MWh, even if the panels were free and of the maximum theoretical efficiency, because you need to install them properly, wire them up properly, and maintain them; this isn't a huge cost but it does become meaningful when the cost you are discussing is under about $100/MWh (currently the price of solar modules is only about 20% of total capex, depending hugely on size of your solar project etc). Solar is cheap, but it isn't that cheap. Even reported prices close to $10/MWh are from the Middle East and are pretty much fudged to make the bidders look good.
However there are plenty of places where if you could pay a solar plant a guaranteed $35/MWh, someone would build you a solar plant. Then the reason why you don't immediately build something to use it - like a desalination plant or electrolyzer - becomes the capacity factor. These are expensive kit, and solar capacity factors are at most 35% (I mean, even in theory you could only run them in the daytime, ie 50%. Yes you can add a battery but that's now even more kit!). This is at least theoretically possible but the economics aren't the slam-dunk you might think, because of capex.
What mostly we (ie the real solar industry) is banking on is making demand more flexible. Batteries, but also charging electric vehicles in the day, deep-cooling industrial fridges, maybe shifting datacenter loads to parts of the planet that are currently enjoying renewables.
Terraform Industries is not a serious company; if it really had the tech it claimed (making liquid fuel from sunshine), it would be an instant win, but it can't. Synthesizing hydrocarbons from air and water vapour is just... not a thing at the moment.
Thanks; how do these numbers change for a large solar plant vs. rooftop solar?
I'm pretty sure you don't want details so: a residential rooftop solar plant costs roughly three times what a utility-scale solar plant costs, in most countries. This is mainly because of sales and marketing costs to individual customers and also because you normally need scaffolding, people need to travel out to do the work individually, the inverter needs to be smaller, etc.
A commercial solar plant (ie like 1MW on a Wal-Mart) is between the two on cost. Some economies of scale but not *as* cheap. Except in China because the state organises bulk purchasing schemes.
Regarding demand, one generally overlooked sink that many people have is hot water tanks! At least in the UK most hot water tanks can be heated either from a natural gas boiler or an electric immersion heater, and the water stays hot for a long time. You can't convert it back to electricity, but it's a useful overflow mechanism that most people already have and is easy to set up.
"Synthesizing hydrocarbons from air and water vapour is just... not a thing at the moment."
What do you mean? They are literally doing it. Or do you mean that doing this *at market prices* is not a thing?
There are many companies around the world who are serious about making green hydrogen, ie from running electrolyzers on renewable energy. (The biggest volume electrolyzers are, of course, in China). It's do-able but expensive and honestly there's just not that much demand for green hydrogen, but they're ramping up major capacity.
There's also interest in making this hydrogen into ammonia to make it easier to store and transport.
There is very little interest from serious players in making hydrocarbons. One issue is that there really isn't much carbon dioxide in air, and if you start with concentrated CO2, you haven't started "from air". It's actually relatively hard (energy-intensive, capex-intensive) to get high-concentration CO2 though it's definitely not impossible (that's why CCS has become CCUS, ie carbon capture, utilization and storage). But most discussions of Sustainable Aircraft Fuel, SAF, are talking about biofuels which are something different.
This is why I am dubious that a US startup which publishes very dodgy solar numbers has simultaneously cracked low-capex electrolyzers AND making hydrocarbons from air plus hydrogen.
Stick a couple of wires in a cup of salt water. You just made a low capex electrolyzer. And making hydrocarbons from carbon dioxide plus hydrogen is a trivial chemical process.
None of that means that this approach will be competitive! But to act as if it's some impossibility is rather silly.
I would not discount the existing criticisms of solar, namely that they take up space. In theory we can cover our roofs in solar panels. In practice it means converting the lovely farm or pasture outside of town into industrial infrastructure -- it's easier for a single company to just buy an old cornfield and cover it with 10,000 roofs worth of solar panels than to convince 10,000 individual homeowners to individually get permission for, purchase, install, and maintain rooftop solar. There is a *really strong* negative sentiment in rural places towards solar power now, specifically because developers are buying up family farms and paving them with concrete and silicon. There's the feeling that a farm benefits the local community, while the solar plants are just using the land to power an Amazon datacenter so tech workers in cities can get rich. People value land and its usage at more than just the pure economic value of it.
Maybe I'm biased because I'm currently reading *The Origins of Woke*, and it's front and center in my mind [1], but it seems obvious to me that Civil Rights law could have contributed to the Great Stagnation. When hiring based on merit became heavily restricted/turned into a dystopian nightmare in the latter half of the 20th century, it would make sense that our collective competence/growth/progress would suffer as a result. It obviously wasn't the only cause, but I'm not sure why Scott didn't mention it at all.
[1] My *Two Arms and a Head* book review horrified Richard Hanania, so I figured that reading his work was the least I could do to make it up to him.
Another possibility, related to civil rights law and crime, is a decline in agglomeration. Before 1970, the NYC metro area's population had always increased, but it went from increasing by 11.2% from 1960-1970 to decreasing by 4.1% from 1970 to 1980.
https://en.wikipedia.org/wiki/New_York_metropolitan_area
That may be a very local effect. Urbanisation has continued to increase. https://en.wikipedia.org/wiki/Urbanization_in_the_United_States
The suburbs and small towns would still seem to count as urban rather than rural for that graph at the top.
Even per Hanania, it seems unlikely that that would directly contribute to much stagnation. Sure it's a constraint, but Hanannia notes that markets are incentivized to optimize within that constraint. That could, take the form, for example, of relegating diversity hires to lower skilled roles, while keeping higher skilled roles more meritocratic.
Indeed, as Cremiuex discusses in this thread (https://x.com/cremieuxrecueil/status/1848828338456113528), the dollar value of an additional point of IQ has increased in later American cohorts, as high compensation jobs and low compensation jobs have become increasingly sorted between high IQ and low IQ individuals.
That's still extra friction, transaction costs. Time and mental energy and communication bandwidth the hiring manager spends on tactical maneuvers related to the prospect of being sued is, one way or another, coming at the expense of their actual value-adding job. Farmers can try to "optimize within the constraint" of siege warfare, but it'll hurt the yield.
The Great Stagnation is a West-wide phenomenon, though. Sure, other countries have laws resembling the CRA, but the issue is still generally not as front and center as in American politics.
>Over-regulation was the enemy at many presentations, but this wasn’t a libertarian conference. Everyone agreed that safety, quality, the environment, etc, were important and should be regulated for
Obviously everyone's entitled to their own opinion, but given that the post itself acknowledges that overregulation of nuclear cost literally millions of lives, which is orders of magnitudes more than the sorts of archetypal environmental disasters that such regulations are trying to prevent, and given that the post suggests that regulation could be (part of) the reason the US isn't twice as rich, which would equate to a cost of tens of trillions of dollars, maybe it's worth rethinking the paradigm of "just have the right amount of regulation," which has cost trillions of dollars and millions of lives, and give libertarianism a second thought.
Was there any discussion on progress and stagnation in mental health interventions?
"I asked some of the YIMBY leaders there what they were doing - did they have blackmail material on Governor Newsom? They said politicians had finally realized that housing prices were in crisis, started groping around for solutions, and they’d been there at the right time and worked really hard to get the message out."
This reminded me of Milton Friedman, who would say that politicians/regulators take whatever ideas are in vogue. That's why you need to keep your ideas in the public discourse all the time. You never know when the current ideas will become obsolete and people will start searching for new ones.
I'm attempting to coin a term for the progress studies lexicon: Q̶u̶a̶t̶t̶r̶a̶ Qattara depression techonologies.
These are techs (like solar and EVs were, or what airships could be) that require subsidies to break out of the local optima and reach a new, better local optmia. I think our system is sometimes too much of a greedy algorithm and gets stuck in local optmia too easily. I'd love to get progress studies folks thoughts on this.
https://medium.com/@bobert93/qattara-depression-technologies-26723f5b362f
Coining a term is rarely accomplished through brute force. The term itself is clunky, as you noted, and uh, misspelled in this post. Less link spam please!
I suspect to coin a term requires high status, which I lack. But I think it's good to try anyway and see if it sticks (very unlikely I'm sure).
Id suggest low status and dyslexia, people mimic my typos
Excellent - I have both of those things
Qattara: a Great Depression, or the Greatest Depression?
I think the chart showing California's energy use on a typical spring day is interesting. I am surprised to see that batteries are a significant factor, even for a relatively short period of the day.
However, it is important to note that residential electricity rates in California are the second highest in the nation (Hawaii has the highest), and are triple those in Louisiana, which has the lowest rates. California also has a lot of desert, which is ideal for solar. Further, I wonder what the chart would look like on the day of highest demand.
"The objections failed because Progress Studies is the same type of field as Gender Studies"
Ouch! Damning with faint praise!
Though by the later descriptions, the get-together seems to have been at least fun and at worst, no worse than these kind of talking-shops generally are.
Really? That just sounds like damning with harsh condemnation to me.
It depends if you think Gender Studies are worthwhile or not.
Scott seems to have a generally favourable impression of the Progress Studies get-together, so he probably means it non-negatively.
Love the idea of Progress Studies and the summary of this conference! It's a great antidote to my worries about the election and gives me at least some reason to be optimistic, even if Trump were somehow to return.
But not sure how to take the couple mentions of "Ralph Nader, Jane Jacobs, Rachel Carson, hippies, protectionists, and all those people", which seemed to be negative. Jane Jacobs is invaluable. Rachel Carson was important. Nader, in his early days, did some good and important things as well.
Scott isn't implying that embracing Progress Studies means rejecting their legacies is he? Wouldn't the Progress Studies we want incorporate their insights rather than reject them?
Someone/something can have had a positive initial impact, but lead to an overcorrection in the other direction in the long run. The answer probably depends on which of those things you're referring to as their legacy.
That's certainly fair. I think, for example, Ross Barkan's recent Times article (https://www.nytimes.com/2024/09/30/opinion/the-power-broker-robert-moses.html) attempting to rehabilitate some of Robert Moses' legacy is fair, although I think its framing as a flaw in The Power Broker was wrong (and I wouldn't accept it as flaw in Jacobs reasoning either.
The Power Broker definitely acknowledges the positives of Moses' legacy as well as its flaws, and I still believe the core of Jacobs' argument about how cities works is spot on as well. Just as I also think that Carson's cry for us to be conscious of the environmental consequences of our actions and Nader's argument that we shouldn't simply trust companies or the market to ensure product safety shouldn't be discarded either.
Having said that, I fully accept the argument that groups with an interest in stopping progress are abusing mechanisms like environmental reviews and community reviews to simply stop development. If Scott is just saying we need the pendulum to swing back to prevent these abuses (versus getting rid of these mechanisms altogether), I'm fully on board with that.
This is wrong in so many ways. (but it does make me wish I'd gone to this conference)
>Is it theoretically possible that - since there aren’t existing fusion-related regulations - we’d have a fresh chance to fight the same battle and maybe win this time?
This battle has already been fought over the last couple years. The NRC voted to regulate fusion power plants like any other facility that uses radioactive materials (e.g. hospitals), rather than apply all the regulations for fission reactors. Then Congress passed a law saying the same thing.
The real problem with fusion is not regulation, it's that it still needs a good 10-15 years of tech development (which might or might not work out well) before we can even start the process of scaling it up to a significant chunk of the grid. Also we don't know how much it will cost.
>Between 2010 and 2019, the nuclear/solar cost comparison fell from $96 / $378 to $155 vs. $68 (yes, nuclear got more expensive).
Nuclear did not increase in price by 50% over 10 years. Very few nuclear plants were being completed during this time, so the data is noisy. They're not a fungible commodity like solar panels; this is like saying the cost of space stations has gone up because the ISS cost more than Skylab.
>Solar used to be limited by timing: however efficient it may be while the sun shines, it doesn’t help at night. Over the past few years, this limitation has disappeared: batteries are getting cheaper almost as quickly as solar itself.
This limitation definitely has not disappeared. Battery storage is cheaper than it used to be, but still about $300-400/kWh and only expected to fall about another 50% by 2050 [https://www.nrel.gov/docs/fy23osti/85332.pdf, see figure ES-2]. My guess is the OWID figure doesn't include packaging, installation, facilities, finance costs, or the power and control equipment to connect to the grid.
At $200/kWh and electricity revenue of 5c/kWh, even if you assume the input energy is free, you would have to cycle the battery about 4000 times to pay for itself. That means you're cycling it just about every day; so storing daytime energy to use in the evening is great, but anything longer term is out.
>If these trends continue, solar power could reach $10/megawatt-hour in the next few years, and maybe even $1/megawatt hour a few years after that. This would make it 10-100x cheaper than coal, and end almost all of our energy-related problems.
$1/MWh is laughable. Transmission costs alone would be far more than that. Even at $10/MWh, further decreases in generation cost aren't worth much because other costs become dominant.
>In fact, nuclear has been increasing ever since the ‘60s, when it cost about $20 (in current dollars). This is proof of concept that nuclear can be much cheaper than it is now.
Eh, kind of. Those old prices are pretty fake because the manufacturers sold plants below cost as loss leaders to increase their market share.
>If we did everything right - got really good regulations, innovated hard, switched to the most promising type of nuclear reactor - we could not just get the cost back down to $20, but go even lower - as low as $1 per MWh within a decade or two.
No, again $1/MWh is not real and it will not be real in the foreseeable future.
Re: batteries - According to this article:
https://substack.com/home/post/p-149971818
Costs for systems on sale today are already below the most optimistic projections for 2050 from that NREL analysis. How confident are we that storage isn't going to get a lot cheaper than the NREL prediction?
Thanks, I hadn't seen this. But I also don't really think it's representative of a full grid-connected system. Compare [https://tag-en.com/tagenergy-acquires-99-9mw-battery-storage-facility-from-res-as-rapid-growth-in-uk-continues/] about a 100 MWh facility that just turned on that was bought (100% stake) for 65 million pounds, i.e. about $840/kWh. Or [https://tag-en.com/tagenergys-second-uk-battery-storage-facility-operational-as-construction-completed-at-chapel-farm/] from last year, $390/kWh. Or this, [https://thebulletin.org/2024/05/600-million-battery-storage-project-meets-resistance-in-california/], $600m for 600 MW which if that's about 1200 MWh comes out to $500/kWh. It's hard to get solid numbers for who's paying what, but $148/kWh (which is a projection by Clean Energy Associates, not an actual project cost) doesn't seem particularly reliable.
But to answer your question, it could get a lot cheaper than the NREL projection without getting cheap enough to really change the use case. Even a 10x decrease will need hundreds of cycles in its lifetime to justify itself, so it won't (can't) be adequate for unusually long periods of high demand or low supply.
We need the TIC cost to get to ~$50/kWh for batteries to be actually cheaper than NG at North American prices for diurnal arbitrage (assuming an average of 0.5 cycles per day).
For true seasonal storage (1/yr) to make sense, we are needing more like $1/kWh…
In my following of fusion news I find that companies are definitely sensitive to regulation. It seems to be a key parameter to limit the on site inventory of tritium, due to the potential emergency planning zone and risk of regulation. I believe the NRC ruling is that pilot plants will be regulated under particle accelerator and medical application rules while fleet power plants are likely to have some kind of fusion specific regulation
Helion has mentioned that one reason they might have two different kinds of machines is that you could have the energy producing D-He3 machines near populations with virtually nil tritium inventory. The D-D machines producing tritium and He3 might end up on an island somewhere. At scale that might not pose much cost for them but I feel like it's a real issue for the D-T reactors
It does look really promising for regulation, for example kind of a slam dunk in Britain who are being aggressive with fusion and have formalized a favourable regulation setup already. The US will likely follow so it's probably not a high risk. With tritium's mobility in the environment it does seem plausible to me that ALARA type regulations could impose a significant cost (despite its extreme harmlessness in actual fact)
>I believe the NRC ruling is that pilot plants will be regulated under particle accelerator and medical application rules while fleet power plants are likely to have some kind of fusion specific regulation
The NRC did consider a hybrid approach that would use the byproduct materials rules for "near term" fusion reactors, and left open the possibility that future fusion designs would be more dangerous somehow and need more regulation. But the actual decision, and the language in the Advance Act, make the byproduct approach permanent and ensure the courts can't overrule it. See here: https://heatmap.news/climate/advance-act-fusion
The YIMBY website I was on in 2010 was really into Jane Jacobs, so it's interesting that they're seen as opposites now. I think the synthesis is that Jacobs had some good ideas about how to build dense cities that people actually like to be in, but the anti-urban sprawl stance is a luxury we can't afford with today's housing situation. Any new housing is better than no building.
Early YIMBY generally focused more on "Let's build good cities rather than bad cities", but they underestimated the strength of build-nothing NIMBYS and the severity of the housing crisis. It doesn't matter how attractive your city is if people can't find anywhere to live. Hence California's population drops as house prices keep going up.
> Hence California's population drops as house prices keep going up.
I see this as due to decreasing household size, specifically baby boomers whose kids have grown up and moved out. Boomers keep their large houses instead of downsizing, while the kids can't afford to stick around and move out of state.
>(this means the claims that “AI companies are going to leave California!” or “Elon Musk is only supporting this because his AI company isn’t in California!” were even more deceptive than I thought - leaving California would make no difference!)
As one of the people that made this case, I wanted to offer that I sincerely believe it. It was not intended to be deceptive. Consider: Elon Musk hates CA and is moving his companies away (though not xAI yet), AI companies are actively avoiding regions due to AI regulations i.e. the EU, AI offers a level of control and power that pork does not, and Elon is doing extreme things to catch up in AI (e.g. securing lots of H100s, building an AI datacenter in Memphis in record time). When it comes to directing the flow of power, I am extremely wary of how power brokers behave.
I do think alternative explanations, including your own, are reasonable. I recognize it is perfectly consistent for Elon to be afraid of AI, try to control AI, and want more regulation of AI. Admittedly, fitting all these assumptions together is a strange puzzle: AI is powerful, yet people obsessed with power would not deceive to get more of it. AI is powerful, but companies wouldn't change their operations to get more powerful AI. People are anything but perfectly rational in these circumstances, so I am open to lots of explanations. I regret you believing the arguments I made are in bad faith.
Very suspicious that the day computer time starts on is also when NEPA began. We must just be in a counterfactual simulation of how things would've gone if progress slowed down. (/s)
But will things go back to normal in 2038?
>Nuclear is in an even worse situation, as you can tell from the recent $96 → $155 price rise mentioned above. This is partly because - as Matt Yglesias writes - nuclear plants are held to an “As Low As Reasonably Achievable” safety standard. Suppose an company invents a nuclear plant which is twice as cheap and twice as safe as existing models. They might like to try selling it for a bit below the cost of existing models, then pocket the profits. But the regulator will say that because it’s cheap, they have more money to spend on safety, and demand safety improvements until it costs just as much as existing models (even though the new model was already twice as safe).
No, this is not what that means. At some point I guess I have to write an actual article about this, but here I'll just point out that people who are in the process of building nuclear plants complain about plenty of regulations that make their lives difficult, and ALARA is not generally one of them. Also it applies to how a plant is operated, not how it's designed or built.
>The solution would be to enshrine into law some specific safety standard for nuclear - maybe 1000x safer than coal power.
A standard like this has existed since the 1980s. It says living near a nuclear power plant shouldn't increase a person's accidental death rate or cancer death rate by more than 0.1%. This has actually contributed to decreasing operating costs quite a bit for existing nuclear plants. Nobody has really built any new ones in the US except for Vogtle (which was a disaster for many other reasons) so the jury's still out on what impact it has there.
I have been reading and listening around in the Nuke space for a couple of years now and Scott's article/your comment strikes me as exemplary of recurring discussions one sees there.
I want to highlight this here because I think these debates are indicative of an ideological rift within the nuclear community that goes much deeper than mere disagreements about regulatory details or the economics of small modular vs large reactors and is not always easily discerned from the outside where people might think of nuclear essentially as one blob on the political map.
Basically - as I perceive it - in the nuclear community you have two camps. You have a nuclear start-up adjacent group who roughly believes that overregulation and lack of learning are nuclear's main problem, They aim to address this via new (small modular) reactor designs that can be factory-built as much as possible on the economical side and lobbying for deregulation on the political side. If one wants to read such a perspective I recommend Jack Devanney's Substack who argues - for example - for underwriter certification for nuclear power.
Importantly, these people are libertarian-leaning and fundamentally believe that nuclear is held back by regulation but if unshackled could be extremely competitive on the free market - perhaps even leading to a "to cheap to meter" energy abundance situation.
The other camp is more aligned with the traditional nuclear industrial-scientific-regulatory complex. While they agree that there is a problem with regulation in general, they think that this has much improved since the 70s/80s. They are also very skeptical that significant learning can ever be achieved in the nuclear industry and view SMRs (Edit: more precisely, advanced nuclear/Gen 4) mostly as warmed-up versions of tried and failed designs from the 50s and 60s. Instead the only way to make nuclear cheaper is building bigger reactors (>1000 MW) to achieve economies of scale. According to them the main problem with nuclear currently is a lack of strategic industrical policy. Nuclear know-how and supply chains have withered to a degree that we have simply forgotten how to build new nuclear. This perspective can for example be found on the Decouple podcast which I highly recommend. In Germany, the main proponent of this perspective is the nuclear historian Anna Veronika Wendland who also comes recommended.
Crucially, these people don't believe that given changes in regulation in a free market nuclear would necessarily succeed. They believe that cheap fossil fuels, e.g. the shale revolution, are as much to blame for nuclear's demise than regulatory issues.
But they still think that nuclear ought to be our primary choice of energy generation. They think nuclear is "cheap" but in ways that the free market fails to appreciate. These market failures include:
- financial markets and privatised utilities are incapable of funding very high CapEx (10-20 billion), very long term (60-100 years) projects such as new nuclear plants;
- electricity markets insufficiently reward clean, stable, always available electricity;
- markets on their own will not ensure that the supply chains necessary to keep the industry alive are maintained (the problem being that a build-out of 1-10 large nuclear plants is enough for many countries and ought to be done in series to be cost-effective, but after that workers leave the industry and know-how is lost).
These are nuclear socialists (in the social democrat tradition) who want a government-driven nuclear industrial policy including state-funded build-out programs such as the French Messmer plan and national champions such as Russia's Rosatom.
I think it is important to keep track of these fundamentally different perspectives or else one gets a skewed picture of what nuclear advocacy is all about. Scott clearly only got to hear the libertarian perspective at the Progress Studies conference.
Now who is right? Personally, my view is that the nuclear socialists tend to score a lot of tactical victories against the nuclear libertarians in debates. They tend to come from "inside" the industry and therefore are just much more knowledgeable about individual issues like ALARA or whatever. The libertarians are also just biassed against government and tend to take every absurd regulatory story they hear at face value. For example, check out the episodes on "What went wrong at Vogtle" with James Krellenstein where he goes into (excruciating) detail to show how the NRC actually performed rather well according to its mandate and did everything it could to help get the plant built.
This mirrors Scott's semi-regular anti-FDA posts on here which regularly involve him getting the facts wrong on some supposedly terrible thing the FDA did.
But I still think that strategically, the nuclear libertarians are right. Whatever the details, the core problem really *is* that nuclear is overregulated. To be precise: I think the parallel universe in which we continued to build nuclear power plants as we did before the mid 70s, achieved ~50-100% nuclear electricity generation globally but had ~10-20 Chernobyls would be much preferable to our current trajectory. Moreover, I don't think that is a likely scenario. Much more likely nuclear would have gotten much safer without getting extremely expensive through the knock-on effects of scaling and *energy abundance*. In an energy-abundant world our economy and technology level would have been so radically transformed that solving nuclear safety would likely have been a trivial problem.
I refuse to believe that we are now forever doomed to the current path. It ought to be possible to remove the insane anti-nuclear bias introduced by our peculiar nuclear history (imagine if the first contact we as a species had made with nuclear would not have been the bomb but a power plant) from our public psychology. So pro-nuclear politics ought to consist in doing the work of educating people about the real benefits and risks of the technology, making the argument for lighter regulation and thereby enabling nuclear to compete on the free market. If the winning approach will be small or large reactors, who cares?
Thank you, I think this is a good breakdown of different perspectives within the nuclear world. Clearly I fall a bit more toward the second camp than you do, but rather than open up that debate in general, let me point out one motivating factor that I think is pretty commonly felt in the industry (and has been from the start). Radiation is scary, partly for contingent historical reasons like the bomb, partly because it can give rise to particularly gruesome imagery (the radium girls, Louis Slotin's nine days as a zombie, Eben Byers' jaw falling off), and partly because it creates a threat of cancer specifically. I have heard that people go to much greater expense to avoid cancer than other ways of dying (can't find a source right now). I think public fear is a legitimate reason to have stringent safety standards, even if that fear might not be entirely rational. From the outside, to people like Devanney, this probably looks like regulatory capture or self-justifying bureaucracy. There may be a grain of truth to both of those. Public opinion has improved lately, and maybe if it improves enough then the world will be willing to tolerate another Chernobyl or two in exchange for cheap electricity. But a number of countries over the years have turned against nuclear power catastrophically, most recently Germany. Requiring proof that reactors are very safe was always meant to be a defense against that, though obviously it's not a perfect one.
Something I feel could point a finger to ALARA (or not) is Germany's experience in running nuclear plants for $20/MWh vs the US $30/MWh multi unit cost
I posed this to ChatGPT and it jumps out at me that it mentions a leaner workforce due to regulatory environment. I've heard it mentioned that the US pads the plant workforce as a means for spreading out annual worker doses, increasing the number of people who can sub into a higher dose task throughout the year
Yeah, the idea of having more workers in order to reduce dose does make sense as a consequence of ALARA. I would be surprised if it accounts for a majority of the difference in operating costs between countries, but I don't actually know.
I would be very interested in a write up on what ALARA is because I have also hear more or less this exact explanation from a lot of different sources that all seem relatively reputable.
This article is actually quite good: https://www.flyability.com/blog/alara
If you read this it's clear that ALARA is mainly relevant to trying to limit the dose to radiation workers. In particular, I'd like to call out the following consideration for determining what is reasonable:
"The economics of improvements in relation to the benefits to the public health and safety—will the cost of those tools be offset by actual harm reduction?"
That is, in principle, ALARA corresponds to cost-benefit analysis (i.e. maximizing expected utility). In practice, usually nobody does these calculations. People use a qualitative approach to consider options for reducing dose and whether they are reasonable under the circumstances. Whether the company or the plant is profitable is not a consideration. But if, for example, there were a technology that would make it extremely cheap to reduce all plant workers' doses from 4 rem per year to 2 rem per year, the nuclear plant operator would be expected to implement that, because the benefit outweighs the cost. All of this seems perfectly sensible.
On the other hand, the expectation presumably is that this cost-benefit calculation is based on the Linear No Threshold model of radiation harm, which is not realistic. But I want to emphasize that the lack of realism is mainly for very low doses and dose rates. The NRC uses a cost-benefit ratio of about $5000 per person-rem. Workers at a nuclear plant are allowed a dose of up to 5 rem annually, and they may receive it all at once (for example, when doing maintenance on a highly radioactive piece of equipment). There is fairly strong evidence that a rapid 5 rem dose will raise cancer risk by an amount comparable to that $5000/rem value. In practice, most radiation workers never get near that. If the industry were spending tens of millions to reduce ten thousand workers' doses from 1 rem to 0.5 rem, then that would be justified under LNT and maybe not justified under a more realistic model. I think there may be some of this going on, but the amounts we're talking about are not a big fraction of overall operating costs.
Recently, there was a bit of a kerfuffle about the possibility of applying ALARA to design/construction of advanced reactors if they applied for a license under the NRC's proposed new Part 53 framework. I think the intent was good (that designing the plant with e.g. extra shielding to reduce worker dose would make it unnecessary to take special measures during operation). But nuclear advocacy groups objected, and the NRC ultimately agreed with them and clarified that ALARA would remain an operational principle, not a requirement that applies to design or construction.
In conclusion, I really think ALARA, while not perfect, just isn't that big a deal, and there are lots of other regulations that would be better targets for the ire of nuclear enthusiasts.
If there were a law like that for coal power, we'd have no coal plants at all!
I think there are some fundamentals where I'm more optimistic about a radically low nuclear system cost
I was going to say a canonical wind and solar grid requires transmission to triple, but a quick search seems to indicate tripling that would increase system costs by 25%, not severe. This does still create a regulatory vulnerability which involves high voltage lines flying over states that neither sell into or buy from them due to the point to point nature of DC lines
Transmission plus distribution does add to 50% of delivered electicity costs meaning that a high capacity factor onsite power source could deliver radically lower costs. Solar could do this with batteries while nuclear has ~done it with less regulation. I think pretty equal weight should be given to a "$10/MWh" nuclear or solar future
I believe the solution with nuclear would rely on a system that fully balances regulator incentives for safety and economy. This exists in the commercial shipping insurance system where the insurers define a regulatory and inspection system that you have to meet to get insurance. Too strict and the insurer loses business, too loose and they risk losses. This would also require a dose rate based rather than linear radiation harm model, while the market system would itself erase ALARA, as the reactor operator would not buy insurance that came with it attached
Under this proposal, what happens if, say, an Iranian insurer offers cushy rates and minimal questions asked, conditional on handing over the occasional ingot of highly-enriched "waste" ?
Just physically, nuclear inventories are closely monitored. It's technically possible for small distributed amounts, probably from several plants over a large time frame, under the inventory precision to add up to a critical mass (if reprocessed, Iran went the easier route of building an enrichment plant). However it's not practical to my understanding
More generally, if it was legal to sell your insurer nuclear waste, it would also be legal to sell it in general, which is not part of the proposal! Nuclear plants do have insurance providers without a regulatory aspect but they cannot give spent fuel to them
Nuclear plants (in the US at least) have very limited ability to get insurance. Liability insurance is only available up to about half a billion dollars, which is peanuts compared to what they would owe in a Fukushima-type accident. (They also "self-insure" by agreeing to each pay some of the costs if any of them has an accident.) But ultimately the government covers the bulk of the risk. It seems like insurance-based safety rules only work if private companies are willing to provide that insurance.
It would require adopting a non linear radiation harm model that incorporates repair rates
In this post Jack Devanney says with his still conservative SNT model the radiation harm at Fukushima is "less than ten million dollars"
https://open.substack.com/pub/jackdevanney/p/snt-versus-lnt-at-fukushima?r=q4k5f&utm_campaign=post&utm_medium=web
Yeah, Jack Devanney likes to say things like that. But even he admits that we'd need wholescale tort reform to make it work in this country, so good luck with that. His implied policy for Japan would have involved spending many millions on radiation monitors and dosimeters prior to the accident (i.e. at every nuclear plant) and then refusing to compensate the people who lived nearby because their doses weren't that high. All indirect damage (like loss of business for local farms or fisheries, let alone psychological effects) he would prefer to ignore. I'm still not sure what he thinks should have happened with the cleanup costs, which are nearly as high as the liability. Maybe he thinks they should have left it to rot, let radioactive material dribble out forever, and given tiny amounts of compensation to people who decided to live nearby anyway. I think that sort of response would be a perfect way to convince voters to make nuclear power illegal, as they have done in many places in the past.
The SNT model, by the way, is not "still conservative." It's guesswork, and fairly optimistic guesswork imo. He assumes a 1-day repair period, and he claims we don't see chronic harm below 20 mSv/day (which is about 1000 times the natural dose rate even in high-radiation places like Denver). He's right that LNT is unrealistic and highly conservative, but he doesn't seem to think it's possible to go too far the other way. The society of health physicists has argued in the past (though it's not their current stance) that a reasonable dose response model would calculate cancer risks linearly but only for annual doses above 50 mSv or lifetime doses above 100 mSv. That would make a huge difference (compared to LNT) in number of expected cancer fatalities, but it would still suggest that the people living closest to Fukushima benefited from temporarily relocating, and that many deserve some compensation.
[Health Physics Society, “Radiation Risk in Perspective: Position Statement of the Health Physics Society” (PS010-2), adopted January 1996, revised July 2010, McLean, Va.]
The compensation scheme pays everyone according to their dose above some cutoff. His concept involves some form of local approval for the plant to go into operation in return for both building it and to agree with the compensation rates. Likely on the lines of paying a special tax rate or funding schools
I believe the compensation also pays non-radiation costs like business disruption. Based on SNT there would be no motivation for large scale replacement of topsoil and the like as the radiation compensations would be far less expensive. The Fukushima exclusion zone is about three times safer than Tokyo when comparing radiation rate to air quality
If his model is accurate the benefits would outweigh the expected potential costs very easily for the local people. One potential controversy is that it pays a flat rate for dose with no special compensation for developing cancer, which is not provably related and unlikely. This would again be worked out in the contract with the local community in order to accept that
"It feels like the United States, after a fifty-year binge on over-regulation, has woken up, wiped the vomit off its chin, noticed it’s lost half its net worth, and started to consider doing something else. I am equally confused why it took so long and why it’s happening now."
Wild-eyed, half-baked hypothesis: maybe the people who remember the real environmental degradation that took place prior to 1970 (e.g., the Cuyahoga River catching fire, acid mine drainage turning streams bright orange throughout Appalachia, dead zones in the Chesapeake Bay, etc) became overly enamored of regulation as just the price we pay for clean waterways. But now that the US has to a great degree shed its dirty industries anyway and the people who remember the world pre-1970 are dying off and being replaced by those who don't, the mood is shifting.
It’s not so crazy an hypothesis. Environmental concerns drove a lot of regulations..
The water in our local harbour used to turn blue when the leather factory released its wastewater into it. Nowadays the factory is gone and the water is (relatively) clean.
People who weren't around during the 70s before the big push on environmental regulations really don't know the other half of the story. Manufacturing facilities in the western world don't release pollutants into the air and water today, but that's not out of the goodness of their hearts or civic-mindedness, it's down to regulations and legislation being applied.
+1
Re. "Why is solar improving so quickly?": Two overlooked points. One, solar has been heavily subsidized by governments, at every level: research, development, production, and retail. Two, solar has low costs now only because coal and nuclear plants can kick in at night or on cloudy days. On a cloudy day, a solar panel's output doesn't drop linearly with the sunlight; new solar panels that I've tested have all dropped to zero output even at noon when the sky is cloudy. There is a minimum of solar power they need to overcome the resistance in the panel to produce any energy at all.
I recently calculated that, if we replace fossil fuels with electricity, then for the US energy grid to store enough solar power for one cloudy day across America, we'd need enough lead-acid batteries to hold 80,293 gigawatt hours of power. A lead-acid battery stores about 35 watt-hours per kg. So we'd need about 2,294 billion kg of lead for batteries. There are only about 95 billion kg of lead left on Earth that we know of. The situation with lithium is much worse.
We can, however, store energy kinetically. For instance, build dams to create lakes, and use electricity to pump water up into the lake, and then let it out to make hydroelectric power. Hydroelectric is about 90% efficient, and big electric water pumps are over 90% efficient, so we should get > 80% efficiency out from the output of the solar panels.
Pumped hydro is great in places that already have dams with big reservoirs. But we pretty much dammed up all the best places many decades ago, plus the environmental concerns around dams are more acute now than they were back then, so building lots more of them in the US is implausible.
Non-water gravity batteries are also an option
I did see a cool idea a while back for storing gravitational energy by cutting free a giant cylinder of rock in the ground and then pumping high pressure water underneath it to make it rise up above grade. The advantage is that the cost should scale with the surface area, but energy capacity scales with the mass, so the square-cube law sort of works in our favor (though not as much as with dams). Downside is lots of expensive digging, among other things. Also I'm not sure how you structurally stabilize the rock once it's towering up in the air.
My company, Form Energy, is trying to help with this using massively scalable and affordable iron-air batteries!
https://formenergy.com/technology/battery-technology/
carbon nanotube nanostructures have the potential to use mechanical stress to store energy at similar density to lithium-ion batteries, using much more common atoms. But that technology is way beyond our current tech level.
Accelerator driven nuclear could easily be throttled up and down as the inverse of solar supply, but in that scenario it'd probably cheaper to run the nuclear 24/7 and dispense with the solar.
Lithium batteries use on the order of 1kg of lithium per 10kwh of capacity, so 80TWh would use on the order of 8 million metric tons of lithium, which is only ~40x current annual lithium production. As with all natural resources, proven reserves keep expanding and the amount available depends a lot on on whether the price makes it worthwhile to explore for more.
Utility-scale battery storage is already down to ~$120/kwh, and these batteries can cycle 3000x, so ignoring discount rates that's 4 cents per kwh. The commodity lithium is <10% of the cost of the batteries. At CNY 70,000 per tonne of lithium carbonate, that's about $105k per tonne of lithium, or $10.50 for the lithium in 1kwh of capacity. So as the manufacturing gets more efficient, there's room for battery prices to decrease at the same time as lithium prices go up and lithium supply increases because of the increased lithium prices. This happened in late 2022 when lithium was expensive, and then a lot of new supply came on line and lithium prices crashed by almost 90%. See the history of failed predictions around "Peak oil" for an analogy to supposed constraints on lithium supply.
Just tried to reproduce your lead conclusions, don't think its right:
USA uses more like 12,000 gWh in a day, not 80,293.
A lead-acid battery stores about 35 watt-hours per kg of battery. Lead is only 30% of this weight so its actually more like 110 wh/kg lead.
So we would actually need about 100B kg in your scenario.
Meanwhile, while there's 95 billion kg of lead reserves, it's lead resources that are more akin to "lead left on Earth that we know of" and that is 2000 Billion kg.