I think an underrated aspect of the Neom project is that it was developed my western consultants- it's worth making fun of, but predominantly because it shows the lack of imagination with the western mind.
There's a difference between things that 1. seem absurd but fall in realms where fundamental understanding is pretty poorly developed and there is a broad possibility space for surprises and 2. things that seem absurd and fall into very well understood spaces where surprises don't really happen anymore.
Centrally planned economies and modernist urban planning fall squarely into the category of "ideas that were once groundbreaking but have since been pretty conclusively shown not to work for reasons that were pretty obvious in hindsight", and Neom basically leaps into the arms of several extremely well understood pitfalls with no plan to address them.
Add to this that the project can serve many individuals' personal ends without actually succeeding and you have the ideal recipe for money-incinerating white elephant boondoggle nirvana. This is not "ambitious but doable in principle". This is "obviously terrible idea borne of rich-kid hubris on an unprecedented scale".
> Maybe, every so often, do a deep dive into fact-checking something, even if you’re absolutely sure it’s true. Maybe if everybody does this, then someone will (by coincidence) catch the false absurdities, and then the social epistemology thing can work.
I do not recommend this. I did this with "Donald Trump is a Racist" and even though I hate the man and 80% of what he stands for, now half my in-group thinks I'm MAGA. For completeness, I went a little further down the rabbit hole than Scott did in his "You're Still Crying Wolf" essay, and started providing videos of black people saying Trump isn't racist.
On the other hand, much of my in-group isn't rationalists, so maybe that is where my error lies.
In any case, be prepared, after you do a deep dive fact checking things that are "obviously true" to your in-group, that you will have some very negative social interactions anytime you bring up what you found, if it doesn't turn out to be actually-true.
What you're calling an "absurdity heuristic" just seems like a very specific reframing of Occam's razor. An explanation like "there's a Saudi conspiracy to develop amazing construction technology and hide it from the rest of the world" is complex, whereas an explanation like "really rich people sometimes do stupid things with their money" is simple. It shouldn't be mistaken for a foolproof way to identify truth, but it's a useful shortcut.
This post looks like it's an exercise in something like ethics, or a values system. Rationality as a whole promotes a set of values (like any religion), but these values happen to be restricted to certain domains, namely processing and communication of information.
In this case, the 'absurdity' filter can be understood simply as, you have to use your value system to decide which hypothesis are worth exploring and which ones aren't.
If i try to point this out directly - that we are articulating and acting on values _just_ like religions do - people will usually say, no no, this is just about instrumental rationality. If that's so, then why 'obligations' to others? And what's wrong with 'the cowards 'way out?
> But I hate this answer. It seems to be preemptively giving up and hoping other people are less lazy than you are.
Words like 'lazy' and 'coward' have an obvious moral valence to them. Wouldn't a more fair interpretation be something like 'people are different, they will naturally explore different parts of the world, and you only have so much time and attention. You can't _not_ use your value system when determining which hypothesis to explore and test, and which ones aren't worth the expiration'.
Clearly, you are articulating norms and values here. You are trying to explore what is and isn't good, in the realm of reasoning and communication. But do value systems constrain anticipation? If not, then in this community they must live in the Jungian shadow; forever preventing us from fully introspecting what exactly the rationality project is about: defining a religion with truth as its deity.
"Extraordinary claims require extraordinary evidence." Doesn't that give us the good parts of the absurdity heuristic, and leave out the bias?
Relativity sounds absurd; but if you show me the textbooks and the experiments, I can see it'd take an awfully large conspiracy to be lying about it, especially when the math seems to check out. It's a big claim but there's lots of evidence offered, much of which I could check for consistency, some of which I could even replicate (Michelson-Morely!).
Neom sounds absurd, and there's no evidence offered that the Saudi king can make it work, and plenty of explanatory power in his own vanity.
In either case, we're just saying "okay, is there big evidence to go with this big claim?"
Admittedly, sometimes a claim sounds absurd because it sounds bigger than it really is. Relativity doesn't noticeably affect time under ordinary life conditions. Evolution doesn't literally mean a monkey gave birth to a human. AI safety doesn't literally mean the movie _Terminator_.
But extraordinary claims do require extraordinary evidence. "Absurd!" is just shorthand for:
"Why do you expect me to swallow that without first showing me evidence?"
If the absurdity heuristic is just "don't dedicate mental resources to things that have a tiny probability of being true/coming to pass" (with an asterisk about tail risks), I think the 1901 analogies shouldn't really make you apply it less.
Of the set of stories which would sound just as absurd as the internet one if told to someone in 1901, only a vanishingly small fraction came true. It may well have been rational to dismiss the internet story at the time, and saying otherwise may just be hindsight bias.
The word "absurd" is covering up at least two gradients of variation -
1) just how unlikely are we talking? A worldwide network of prostitute-dispensing-UFOs is unlikely, but it's practically a certainy compared to the likelihood that gravity obeys color. Relativity, _being true_, was considered very absurd for quite some time, and honestly, still feels a bit absurd.
"The Line" looks unlikely to succeed by its own metrics, as-is. But I might be willing to put down a bet at 1000:1 odds.
2) What timescale are we covering? in 1900, a worldwide network of adding machines exchanging representations of pictures _would_ be absurd, because several technological, economic, and social changes had yet to take place, and the 'absurd' thing wouldn't come to be for about a century.
"The Line" isn't a claim that "the world of 2120 will heavily feature engineered mega-cities that would look to the world of 2020 as if they were sci-fi dreams", it's a claim that _the thing we're breaking ground on_ will work.
---
The post seems to be asking "how 'absurd' is so absurd we can stop discussing it, or stop thinking about it, as a serious going thing?"
And on that, my rough heuristics would be
- On a timeline under 20 years, the cutoff for "too absurd to be worth discussing" might be somewhere around the 1:10,000 mark. This can be tweaked based on the cost involved.
- On a timeline over 20 years, the sky's the limit, but on the other hand, nothing can really be said with any certainty anyway, so it's _less_ useful to talk about.
In the example of "The Line", I proposed that it might be something like a 1:1,000 chance of success. This might be considered non-absurd, except for the fact that it's such a high-cost endeavor, in terms of human time, money, ecological impact, and the upside so mundane, that it _should_ be laughed out of serious discussion until its proponents have come back with a much stronger case for why it should be possible and valuable.
As another example - earlier this century, there was some buzz about having possibly discovered a reactionless drive (https://en.wikipedia.org/wiki/EmDrive). This was very very unlikely, that we'd discovered something that violated laws we believed inviolable, rather than being measurement error. let's say 1:10m. But, the potential upside was also very high - and the cost of taking it seriously was relatively cheap - some days or months of laboratory time. The EmDrive may have been absurd, but perhaps not _so_ absurd as to preclude serious discussions about it.
Yes, but he doesn't agree that it's absurd, which is a different threshold. (His exact words are "Like, I'm not saying Neom will be a huge success or whatever.")
He's accusing you of cognitive self-dealing (roughly, of straw-manning where you could have steel-manned), and I don't think your discussion above clears you of that charge.
It's one thing to say Neom's design is absurd, but quite another to say "murdering the people who previously lived in the area" in reference to a single man who, according to the Saudis (and AFAIK not denied by anyone), was armed and shooting at police. You don't need sophisticated philosophy to see how misleading this framing is.
I like Neom project. I hate living in cities although I currently live in the city myself. It is just because like majority of people I am not rich and for economic reasons, transportation issues, proximity to work, shops etc. I have to live in the city although I really feel that it is like a life in prison.
Neom, the Line is a fantastic compromise. From one side you will be close to nature, even if it is only a view over the desert. Maybe they can put some irrigation and greenhouses to grow food for the city that will make the view slightly greener. On the other hand it would be a city infrastructure. The space between two walls would be like a street with shops, bars, offices and residential quarters. It is a fantastic project.
Now back to the reality, of course it seems absurd to think that it could be built with merely 500 billion dollars when the real costs are at least 10 times of that. Maybe if they could built the city only 17 km long but even that is unlikely due to corruption and general unprofitability of this project. But it is a different thing to admit that the project is not feasible for various reasons rather than to think that the idea itself is absurd. The US might never have free-for-all universal health care but it is still not an absurd idea.
It is like thinking that Egyptian pyramids are absurd. One estimate was that building a pyramid today would cost 5 billion dollars. It is a lot even today when we have big machines to move stones and project management theories how to organize big projects. In the past it must be unbelievably expensive and basically an absurd project. And yet, pyramids were built because some powerful people believed in the necessity of building them. Who knows if Neom could become a reality if some powerful people believe in its necessity? Or it could be some other seeming absurd project. Or in contrary, if they lose faith in this or another project then “absurd” becomes self-fullfiling prediction.
Comment on the comments: There are now at least two comments which say something like:
"<correct sounding general statement about absurdity>, as an example consider <thing I consider absurd/non-absurd but which a significant fraction of people disagree with me about>."
Come on surely you guys can figure out examples which all readers will agree with you on.
It looks bad that a group that happily discusses immortality, The Singularity, Roko's Basilisk, a universe turned into paperclips, etc., etc., cannot muster any sympathy for Neom.
It is possible for there to exist a statement X that is true, and yet also anyone who believes statement X is true is crazy, because craziness is judged on not based on whether the statements you believe are true, but whether the process by which you arrived at them is crazy. If in 1901, someone says that they know for a fact that in 100 there there will network of blah blah blah adding machines, you would be right to judge them as crazy, because there is no plausible non-crazy chain of evidence and reasoning that could lead them to this conclusion. If you are Einstein, who has just finished deriving the speed of light, but not yet told anyone about it, and someone comes to you with this statement about relative speeds that they know for a fact is true, concluding that they are crazy would still be reasonable, though you'd likely want to consider other explanations such as "they've been spying on me" and "they are a brilliant physicist that I somehow haven't heard of". For a more modern example, if you told me that you know for a fact that there will be a AI singularity by 2050, I would probably conclude that you are crazy, though I may also consider the explanation that you are wildly overconfident, or don't actually know what a probability is. Compared to if you told me that you believed there was a distinct possibility of AI singularity by 2050, and it was more likely than not, I would not call you crazy. Whether there does end up being an AI singularity by 2050 is irrelevant, because the process by which you arrived at the conclusion is what makes you crazy or not, not the actual conclusion. I also wouldn't hire someone to invest money on my behalf _just_ because they correctly predicted some low probability event far in advance, I would look at their process for generating predictions. If it seems like a process that is capable of consistently generating correct predictions, I will judge them a genius. If they (earnestly) tell me that it's because they are communicating with themselves from the future, or god, or "hidden energy", I'd probably judge them crazy _because they made a correct prediction_.
I think there is a difference between "I think this is true now"/"I think this happened in the past" and "I think this will happen/be true in the future."
My current theory on "The Line" is that someone mixed "Inspiration for the next important project" and "best recent video game featuring the middle east".
On the original quoted example about absurdity bias, I think that a smart and well educated person from 1900 would be able to distinguish the truths from the lies there if they were allowed to ask follow-up questions, thoroughly interrogating our time traveller on the details of the supposed future. What happens if the shade of blue is slightly off? What equations exactly govern the distortion of space as you approach the speed of light? What has changed the economics of male prostitution so much as to make this scheme economically viable? How do you convert lesbians into numbers?
You can make anything sound absurd by describing it in a sufficiently wacky way, but reality holds up under interrogation while absurdity falls apart.
(What if those were bad examples and you could make up an absurd scenario that _would_ hold up under careful scrutiny? Then it probably isn't all that absurd. You could probably convince the people of 1900 that the Austro-Hungarian Empire dominates all of Eurasia in 2012, but that's not absurd, it's just untrue.)
All you need to do is say that you don't see how it's possible (and give the reasons why you think it isn't possible). Saying "it's absurd" is essentially conflating opinion with fact. It's adding "something extra" that's not necessary, that's extraneous.
This is ubiquitous on the Internet, and in the long run doesn't add value.
Even "In my opinion it's absurd" is better, but still sub-optimal; you need to account for arguing from personal incredulity, and "I think it'd absurd" can run afoul of that problem.
I like to spend a minute on a back of the envelope calculation wherever relevant. If anyone is seriously proposing X, and I'm writing a blog post criticizing X, I will apply at least a BOTEC level of analysis.
Kingdom Centre in Saudi Arabia is 76.8m * 302m * 37.8m and cost US$453 million in 2002, multiplied by 1.65 to adjust for inflation through 2022. This comes to $850/m^3. I previously calculated ( https://astralcodexten.substack.com/p/model-city-monday-8122/comment/8116431 ) that the Neom Line would be $58/m^3 if they spend $1T. So it's off by an order of magnitude, but maybe still possible to do the project if they can sell the condos for more than the cost of construction. They could build 1% of the line, sell condos and office space to get their construction costs back with a profit, and rinse and repeat until they run out of buyers. Or with enough loans those steps could be somewhat parallelized. Real estate developers usually use tons of leverage and sell houses whenever they're completed rather than waiting for the entire subdivision to be completed.
I think Scott's prediction "Saudi Arabia builds a structure at least 100m x 100m x 1000m before 2040 or the Singularity, whichever comes first: <1%" was very overconfident. 100x100x1000 would cost only $8.5 billion at Kingdom Centre's $/m^3. MBS can easily spend that much on whatever he wants to build so long as it's possible with current technology. I would put the odds at 50% for the same prediction.
Why be wary of the social epistemology? Do you hope to contain all human knowledge in your brain one day? Why would you hope to be able to do epistemology by yourself? I think rationalists should embrace social knowledge and epistemology, treat humanity as one huge complex sentient machine.
When I read the bit about Lee de Forest’s fraud prosecution I was intrigued and wanted to find out more. However the link from Eliezer’s article is now dead and other sources I found suggest that while he was found not guilty, his business partners were not so fortunate.
The actual quote provided seems pretty convincing evidence for the proposition ‘some people found his claims so absurd as to be prima facie evidence of fraud’ but I have no idea where it comes from.
It looks like the CA HSR project has slightly more than doubled in price and added a few more years since then, and Musk has shown that he can in fact make money selling cars, but there really hasn't been any more support for his Hyperloop idea.
I feel like the quoted example of absurdity bias does not fit well here. The examples he quoted are clearly formulated to be absurd, mostly by being oddly specific. Take the black sphere one, for example: These might simply be floating houses or teleporter endpoints. It's quite obvious that some of these houses would also house prostitutes and therefore you'd be able to order one from there.
It's the same thing about the internet example: This is even a stretch as it is - is one of the primary use-cases for the internet really lesbian porn? Why specifically mention encoding or the 69 meme? I'm not even sure which one is referenced here. If, on the other hand, you told that persons that we found a way to make our computers/calculators really, really fast and connected them all up and suddenly that person will probably not be as skeptical. Then you might also add that people also use it for sexual content and it turned out that's something that's quite desired, depending on that persons view they probably won't even be surprised.
The presentation of Neom doesn't fall into the same category, since they're not presenting it by making it intentionally overly absurd and focusing on specific, minor details to detract from the somewhat reasonable big picture. They're completely serious. They also don't seem to bother to explain the somewhat ambitious aspects like the super fast travel and the strange shape. Giving these aspects a pass because, if you just formulate hard enough, you can make everything sound absurd is not a good argument IMO - it's just not an apples to apples comparison.
---
On whether we should accept absurdity as an argument, I think this is actually pretty simple: It's basically a (slightly smug) way of saying "I'm gonna need some evidence to believe this". If your point isn't really important for a decision, as this stated opinion probably is, I don't think it's necessary to dig deep into a topic when the fact that this is pretty absurd is obvious to the casual observer. It'd be different if your opinion did matter a lot (a judge shouldn't be as easy to accept this argument, for example, and neither should your investment advisor) or if some people you trust point out that you might be wrong with this.
I felt as though the absurdity argument against Neom was a combination of multiple arguments, which are not entirely spelled out because any of them would work and it's assumed that most readers can recreate multiple of them without help:
(1) People who are experts at this have looked at this and declared that it will be much more difficult than is claimed.
(2) There is lots of low-hanging fruit: obvious changes that could obviously improve the design. One example mentioned in the comments is turning the thing on its side: it's easier to build a 200 m tall and 500 m wide city than a 500 m tall and 200 m wide city. The reason you build up is because there's not enough land in the city center, but there's plenty of land here. Another example is that it could follow the coastline instead of being perfectly straight with a mountain range in the middle. It doesn't follow the coastline because that's where all of the palaces for the Saudi royal family go.
(3) The project follows a known failure mode. This particular failure mode is called High Modernism and has been well described in Seeing Like A State by James C. Scott (1998).
(4) The project involves a common human experience, where our intuition should be trustworthy. In this case, Neom wants to be a city. We can ask: is this a city that I would want to live in? I'm guessing that the answer to this for Scott is no. This is a less solid argument because different people might think different things.
(5) There are other underlying problems that could ruin the project does not address. In Neom's case, the biggest problem is finding people to live there. They're not going to come from Saudi Arabia itself: the country is barely above replacement fertility rate (2.28 children / woman) and falling, and other cities in the country are also expanding. So they're going to have to attract foreigners. If it's a status project, they probably want elite foreigners, not Syrian or Yemeni refugees. But a lot of global elites do not want to live under the Saudi's authoritarian and Islamic government. Even for those who are willing, Neom has to compete with other elite-seeking Middle Eastern cities like Dubai, Riyadh, Abu Dhabi, and Jeddah.
I personally think that Scott is underconfident that something will be built (it should be >1%, maybe 50%) and overconfident that a significant number of people will end up living in whatever is built (it should be <75%, maybe 25% ... although maybe 50,000 is small enough that Scott's estimate is reasonable). Building a structure that is at least 100m x 100m x 1000m is technically possible with enough money. The Saudis have enough money. So something will likely be built. But persuading people to come will be more challenging.
1. When someone makes a facially absurd claim it is reasonable to dismiss it out of hand until they explain in detail why it is not absurd.
2. When a large amount of money backed by smart successful people is invested in something that seems facially absurd you should carefully consider two possibilities:
a. It is not as absurd as you thought. For example, I thought the idea of Elon Musk successfully and profitably creating an electric car company from scratch was absurd. I was wrong.
b. The stated and actual purpose of the investment are different. For example, in third world countries many large infrastructure projects have stated economic goals that are facially absurd. They fail at those goals. But they do an excellent job of meeting their actual goals to funnel money into the pockets of well connected people.
I do not think NEOM is a graft - the people paying for it are the best connected people. It is interesting to consider what a "failed" or scaled down NEOM project might actually accomplish.
For example, could NEOM eventually morph into a planned city (like Brasilia) that operates as an almost extra-territorial enclave within Saudi Arabia complete with world class universities that recruit globally, offer full rides, are decided non-woke, and admit based on merit alone? Could those be the core of a Silicon Arabia? Could NEOM have, in effect, a separate peace with Israel so it can bring in Israeli investment and expertise? Now that would be interesting.
One of the first steps in making sure you're applying the absurdity heuristic appropriately is just to make sure to steelman the position you think is absurd. E.g. "I know that monkeys can’t turn into humans, this is so absurd that I don’t even have to think about the question any further” sounds a lot less absurd when you understand that monkeys turning into humans isn't what evolutionary biologists are claiming is going on.
Monkeys magically transmuting into humans would be absurd and our anti-evolution guy would be right to dismiss it as absurd if that were the claim.
I think the missing piece here is that beliefs come with levels of certainty. The question of how far to go down the chain is about the degree of certainty that you want to achieve.
The claim that Neom is absurd has one level of confidence attached to it, but, as you point out sometimes absurd seeming things happen. Ok, so you work out the construction costs and that makes it even more unlikely it will happen, you supplement with data about the fact that Saudi construction costs aren't much cheaper than elsewhere and you can increase that level of confidence.
Now that isn't to say you can always increase confidence by digging deeper. That's why you stop when you get to assumptions like: but what if the universe is all a dream or God intervenes and violates the laws of physics. Because digging into those questions doesn't actually (or shouldn't) result in substantially increased confidence about your answer.
Yes, at the end of the day you can't get below your own priors but you can replace a judgement like:. silly sounding things tend not to work, which might only give you a relatively weak degree of certainty with judgements that increase your degree of certainty.
(Not that since probability distributions assume logical omniscience you have to supplement them a bit with some kind of judgement about the likelihood of there being some further deductive argument that's more compelling ...so I've played a bit loose with the term certainty which isn't quite just moving to a more extreme probability what it really is is something more like: evidence that further accessible information or consideration of argument will not lead to a refinement in your probability judgement...or at least that you don't expect the effort to be worth it).
Maybe a better way to put this is the following:. why isn't the obvious answer that you are balancing the rewards and benefits of further inquiry the right one?
I mean, you have certain judgements as to how likely further effort will be to refine your judgement and some payoff function regarding the degree of certainty. Surely, if you had to bet your life on the success or failure of the project it would make sense to expend more effort in research.
At each level you can't do anything but rely on your judgement (what would it mean to do otherwise) but part of that judgement is an evaluation of how likely further effort is to produce a more reliable answer. Do we need anything more than that mechanism?
The quote in the post by Eliezer says "This doesn't mean that anything can happen. Of all the events in the 20th century that would have been "absurd" by the standards of the 19th century, not a single one - to the best of our knowledge - violated the law of conservation of energy, which was known in 1850. Reality is not up for grabs; it works by rules even more precise than the ones we believe in instinctively."
I agree with his overall point that our current understanding of physics is such that some things seem incredibly likely to be universal laws. I have a bit of a nitpick, however, with his precise example of conservation of energy. It doesn't really affect the overall argument but I figured some people here might find it interesting nevertheless.
In classical mechanics, energy is a number which can be easily computed in principle and is conserved. In quantum mechanics, not all states can be associated with a definite energy. However, there *are* certain pure-energy states, and all states can be written as weighted combinations of these pure-energy states. The analogue of energy conservation in quantum mechanics is that the (technically, squared magnitude of the) relative weights of these different pure-energy states do not change over time. In particular, the average energy (computed via these weights) is conserved, and for large systems we can identify this with the classical energy, thus letting us recover the classical notion of energy conservation. This all works perfectly well when we observe isolated quantum systems.
The tricky part is when we also consider wavefunction collapse. Let's say we accept the Many Worlds interpretation of QM. Then, the observable reality around us is just one of a truly enormous number of wavefunction branches. Quantum mechanics only applies exactly to the *totality* of the branches, not separately to each wavefunction branch. Thus, the above argument for the conservation of the distribution over energies applies when you consider *all* of the many worlds but not necessarily when you only consider *our* world, which we inhabit. If we were capable of knowing the true distribution over energies of just our world, it would probably constantly keep fluctuating due to frequent collapse events. All we can say is that when averaged with those of the other worlds, the distribution over energies will be conserved. And yet, if we cannot observe the other worlds, is this definition of energy conservation even useful? One can avoid the question by saying that we only care about idealized, isolated physical systems, not the whole universe treated quantum mechanically. And yet, the latter is what we were trying to seek insight into in the first place...
In Kierkegaard, there are a few separate epistemological stages. This post is a really nice gateway into his whole thing!
Immediacy: You are completely naive and unreflective. You don't yet know the difference between being and seeming. Only babies and very special adults are like this. If you're reading a blog post like this, you're not in immediacy.
Aesthetic: You've discovered reflection; that your ideas about things are different from the things themselves. You've found that reflecting critically on your own and others' ideas brings you closer to the truth, so you do that a lot. Eventually, you're in a prison of increasingly-subtle self-referential reflections, unable to get out of your own head. This is when you question all of your assumptions. He talks about Descartes' "doubt everything". Analysis paralysis, basically, and ultimately, despair.
Ethical: You can't sit around all day reflecting. You've got to do something sometime. What you do should be good for everybody and follow some kind of absolutely valid universal rules. You become conscious of yourself as having eternal validity, through your relationship to these universal rules. This is Scott's evolution denier; he doesn't need to question any further, he knows what he needs to know. But, since you can't actually live up to the rules (you're not perfect, and anyway life will trap you in impossible situations), you're still in despair.
(Here I'll pause to note that the fundamental tension described in this post, between naïveté and infinite regress, is very identifiably the same as the tension in Kierkegaard between the Aesthetic and the Ethical. How do you negotiate between these two? The answer is...)
Faith: Not as easy to describe. Still conscious of your eternal validity, still aware that you're sometimes wrong about stuff, you nevertheless choose, act and grasp hold of your life. You make a "leap of faith" which is more like a dance move than like an Indiana Jones puzzle. You know you're wrong to do it, but you do anyway. And somehow, you win! Or, if you don't, you don't notice much. For a Christian, this is something like thinking, "God put me here, now, and He gave me the knowledge I have, and set things up so I'm faced with this question; if I genuinely do my best, even if it ends up that I'm totally wrong, my sin is atoned for by Christ." Or as Kierkegaard says, you come to rest transparently on the ground of your being.
So, I don't know if other ways of looking at the world have answers to the questions raised in the post, but Christianity does, in my opinion, and Kierkegaard is the philosopher who articulated it best.
(Disclaimer: everybody is always wrong about Kierkegaard, I am not an authority, Kierkegaard would punch me if anybody believed what I said just because I said it. He would punch you if you believed what he said just because he said it. What he wanted least of all was disciples, which makes my fanboying pretty awkward. If this is interesting to you at all, read Kierkegaard yourself!)
Wait, this doesn't seem to make any sense to me. You're not deciding a binary here... not really. A good (rationalist) way to approach this would be saying "This is absurd, and (therefore) very unlikely to happen/be true" Pointing out that historical prognostications sometimes sound like they're unlikely to be true doesn't mean we shouldn't still realize that some theories are dramatically less likely to be true than others.
An absurd retrodiction: a dude with $200M (from the sale of Paypal, from which he was kicked out for "lacking a cohesive business model") would disrupt two overregulated capital-intensive industries all over the world within a decade and a half.
One line of reasoning I didn't see in the comments is using base rate. In such a case, it's not obvious what would be the relevant reference class, but I feel that "mega-projects that planned to spend similar amounts of money" and "building a city of that size from scratch, by centralized fiat" are good contenders. From what I've read by Bent Flyvbjerg, the odds are not good that this will be delivered on specs, on budget, on time.
Back to the question of the absurdity argument, I think that "if this project worked, it would be so far from its reference class that it's absurd to imagine it could succeed (in the near future, barring any major breakthrough)" is a reasonably applicable and successful heuristics.
Re. "Maybe, every so often, do a deep dive into fact-checking something, even if you’re absolutely sure it’s true. Maybe if everybody does this, then someone will (by coincidence) catch the false absurdities, and then the social epistemology thing can work." :
YES. Taking random samples of life is a tragically under-used technique. It's especially useful if it becomes a social norm.
Neom is something that you come up with if you're playing a video game where the goal is to pack as many people in as closely as possible, while damaging the environment as little as possible, while not respecting any constaints of reality. It's like a parody of the average consultant's mindest.
I realize it's Saudi Arabia, but presumably people still want to go outdoors in Saudi Arabia. I suppose you can take a few steps away from The Line and look back at the wall... and then walk back inside...
Scott, I read your post as saying, Saudi Arabia thinks they can do this but I think it's really really unlikely.
Seems like an accurate assessment of the situation, and not a claim that no way no how maybe somehow Saudi Arabia has figured out something fantastic that will have ten million people living in a straight line
I think there's a practical exercise to quickly examine these sorts of biases; I call it a "pre-mortem" exercise. The exercise is simple: write two short memos, each 0.5-1 page in length. One is titled "why it worked" and the other is titled "why it failed". Circle back up and compare. If you all agree, those are prioritized strengths and weaknesses. If you disagree, you've unearthed differences in your assumptions! Or, in this case, differences in your absurdity lenses.
This exercise was largely inspired by the 'Safire Memo', the backup speech for if the Apollo 11 landing had gone wrong. Some credit also goes to the observation from the book Strategy Beyond the Hockey Stick, which observes that in big organizations that fell from the heights (Nokia, Kodak, etc) even when leaders individually knew that the new technology would win the organization still failed to bet that way.
Missed the chance to title this "The Neom-Central Fallacy, or, The Most Absurd Argument In The World".
More seriously, I think any sort of absurdity heuristic or bias becomes less relevant as the...complexity waterline gets raised. Without going into __any__ specific examples *cough*, the world just seems a lot more complicated today than [arbitrary past cutoff]. I guess some of this is an inevitable part of growing up - one can only hold so much information in the brain, and the rate of information-addition to the world far surpasses that. I think the slope of that latter line matters, though.
Even though in many ways it's easier than ever to Deep Dive: Much More Than You Wanted To Know, if the rate of increasing Unknowns and Unknown-Unknowns outpaces the rate of generating Knowns, then the total space of non-knowledge never shrinks. A whole lot of this can be safely written off via rational ignorance. But when even immediately practical and relevant things start to become Absurd, well, there's a problem. That's not knowledge that one can safely ignore; life's table stakes keep getting bigger. And there's so much bad-faith social epistemology now (classic update method from antiquity! insert evosci pablum here) that I'm kinda surprised to see it recommended as a potential fix. That seems to rely on one's ability to avoid a whole bunch of *other* biases...
Anyway, this feels like a post that coulda been avoided merely through slightly different word choice or a quick edit. "Define terms please" is next to spellcheck for godliness. The general free-for-all confused melee in the comments seems to indicate that while a springboard post to discussing Absurdity Bias was a good idea, Neom isn't a particularly good category member to base it off. Possibly doesn't belong in category at all.
Neom isn't a bad idea because it sounds impossibly expensive and difficult to build with today's technology. Any proposal like that might happen to get an Elon Musk to make it cheaper and change the whole equation faster than anyone could have imagined.
Neom is a bad idea because no matter how advanced our technology is, even if we're literally building Dyson spheres and are immortal and can build giant skyscrapers the height of Mount Everest in a day, a linear city is always going to be worse than a city that is literally any other shape.
Irrelevant nitpicking ahead: EY's examples in the Stranger Than History quote are bad.
In 1901 there were maybe physicists like Plank who had an inkling to what Einstein will propose only 4 years later because it was based on theories and experiments already available like the Michelson–Morley experiment of 1887 suggesting the Lorentz invariance of the speed of light. Plank was definitely not shocked by relativity.
Also in 1901 the world was crisscrossed by telegraph lines, telephones networks were growing and there were plenty of newfangled sophisticated mechanical calculators like the Millionaire of 1892 and electric tabulating machines using punch cards made since 1890 by Hollerith (his company will eventually by a founding part of IBM). Telling someone that in a 100 years there will be a network of electric calculators connected by wires would have not been a shocker and I'm pretty sure it was predicted by some at the time. And they did had all kinds of porn fiction in 1901, including lesbian.
It's not just about absurdity; it's about Bayesian probability. If you collect everything we know about physics and technology, and plug it into the Bayes' Rule, then we will discover that painting yourself a specific color is incredibly unlikely to reverse gravity. Of course, we could be wrong; but we would be foolish to make long-term plans or investments based on such a thin hope.
The tricky part here is that if we told someone from 1901 about the Internet, and he plugged in everything he knew about science and technology of the time into Bayes' Rule, he'd be forced to conclude that the Internet is a scam. He would be factually incorrect, but his conclusion would be fully justified. But, if he stuck with his rigorous thought processes (and somehow lived about 100 more years), he might be one of the people who got to *develop* the Internet.
There is no magic oracle that you can consult to get 100% accurate answers about any question. The best you can do is balance false positives and false negatives; and this means that you would occasionally miss the one cool revolutionary idea that will change everything. In exchange, you gain the kind of steady progress that gets you from steam engines to the Internet in only 100 years. It's a worthwhile tradeoff.
This post is IMHO completely unnecessary (and, sadly, showcases why "the substack quality has fallen compared to SSC"). Any modicum of "common sense" (horror!) would give you the same conclusions in 20x less words. This is simply not interesting and not why I read ACX (less and less these days, for reasons above).
I honestly don't understand the concern here. I mean, I do understand the object-level concern, I just don't see how the problem doesn't simply generalize to "P(A|B) = [P(A)*P(B|A)]/P(B), all the rest is commentary".
I mean, the only way I can conceive the problems posed in the post as problems is when I forcibly reduce the available options to simple binaries. Like, is "absurd" supposed to mean "literally impossible" - or just "having a very low probability"? Is "questioning your assumptions" a yes/no option where you either require them to have a strict logical proof, or require no proof at all - or is it more like having probabilities assigned to them, and calculating the need to revisit them from the current context? Isn't the evolution guy's problem the fact that he already decided he "knows" something, instead of admitting and processing new evidence?
In each of those cases, reducing to a binary appears to be the obvious failure mode, and eliminating that failure mode an instant improvement. (Now, in each of those cases, there's also no simple way forward after that. But I, for one, find reassuring the conclusion that there's no simple trick and we should just all follow the usual epistemic best practices.)
(I don't understand the problem with social evidence either, the vast majority of each of our individual knowledge was provided by other people, one way or another, there's no escaping that. Not to mention it would be extremely impractical to escape. Humanity runs on trusting others to do things one cannot do oneself. It got us far, certainly farther than each of us trusting only one's individual experience would have gotten us. It doesn't feel like a cop-out to me, maybe because it seems like a description of what's already happening, rather than a prescription of what should happen.)
I think I'd defend "use absurdity as a heuristic" in basically all cases. Taking Yudkowsky's examples:
Special relativity is completely absurd, and people should dismiss it out of hand until presented with a stack of evidence. No-one does this, because its so strangely beautiful you end up wanting it to be true.
If you explained the internet in more detail, most reasonably intelligent people in 1901 would probably accept that the invention itself would happen if it was possible. They'd accept the possibility wasn't absurd if they has a period-appropriate grasp of electricity. Anyone who was being intellectually honest would accept the porn thing as well.
The answer to the hidden example is that everyone thinks their own moral opinions (possibly all opinions?) are correct. If they thought they was wrong in some way, they'd change them to whatever they thought was right. "Moral consensus kind of drifts around haphazardly" is an idea that's clearly true and not absurd in the slightest, but people reject it because they want morality to be objective.
More generally, if you hear a very complicated argument for something ridiculous, you should still dismiss it because its more likely that the argument is wrong in some way than that the ridiculous thing is true. Neom is a bloody stupid idea because it's meant to be a desert city that's one block wide and the length of Ireland. If you show me a really good argument for why this is a good idea, then the right thing for me to do (even if I find the argument utterly convincing) is to reject the argument nonetheless and fall back on it still being a bloody stupid idea and you being a sophist.
There's a difference between it's impossible and it's impractical. A difference between being on the edge of known physics and being central to everyday experience.
There's very little which is uncertain about this project from an engineering and physics perspective. The 2 main areas I'd be worried about would be wind loading (absolute and differential) on a surface that large, and the environmental impact. The first can be mitigated by over-specifying the design for that force (and we already know how to build moment frames). I presume the latter will involve a lot of sand piling up in annoying locations which can be addressed with standard ground moving equipment. If nothing else, it will be a great way to learn more about environmental science.
What makes the project absurd isn't the physics/engineering which, granted, a lot of people are unfamiliar with. It's the parts that people are familiar with - everyday living in urban environments. And approximately everybody seems to think this isn't a place they would want to live.
Everything connected together with doesn't have to be terrible. The Las Vegas strip has a lot of major attractions connected together with a monorail. Lots of varied businesses connected indoors can work very well - the Minneapolis Skyway system is fantastic. But for individual large structures the advantages are limited. Skyscrapers exist because the demand for an area means the cost of land is higher than the cost of building upwards. Malls exist because there's an advantage of having lots of otherwise-unrelated stores in close proximity creating a network effect. And Amazon Warehouses exist because there's a need to be able to store a wide selection of stuff in an area where it can be individually boxed up as needed. What's left? Airports, because the minimum cost of decent runway is best amortized over a lot of flights which are themselves best served by a lot of jetways?
This is in contrast to other wacky projects where the objections are related to engineering factors like cost rather than experiential. Approximately nobody is objecting to Elon Musk's tunnel building projects based on the idea that "tunnels can't work" or "nobody wants to take tunnels". A lot of people are familiar with subways, highway tunnels, etc. The objections are on the engineering side about the ability to reduce the cost of development for the given service level/capacity and safety required.
If this was simply an infrastructure which pre-deployed subway, water, sewage and other utilities in the desert under a "if you build it, they will come" model it would still probably be a gigantic waste of money, but at the very least it would be preparing for a future which might possibly occur and meet the needs of people.
(I got an email about the comment I'm replying to here, but I can't see the comment on the actual page. I wonder if it'll show up once I reply to it? EDIT: comment system is bugged; this comment was made as a reply to a reply to my top-level comment, here: https://astralcodexten.substack.com/p/absurdity-bias-neom-edition/comment/8185050)
Wow, you're totally right. Chapman is in a very Kierkegaardian place. His solution to the whole problem, "non-systematic situated meaning-making", he draws from Heidegger, who was inspired by Kierkegaard. So it looks like he's getting some Kierkegaardian insights at second hand.
Hegel had noticed the problem of nebulosity, that no system is actually true and right, and had responded to it by making a system of systems; this giant uber-system which would account for and rationalize all possible systems, and explain how and why various systems arise in various times throughout history. Hegelian thought had completely taken control of Europe, and much of Kierkegaard's writing was rebelling against it. So Kierkegaard talks a lot about the deficiencies of this kind of naive systematizing.
As someone who is so excited about prediction markets and, by extension, the wisdom of crowds, I am a little surprised at the degree to which you dismiss the "social epistemology" dimension. I'm not saying that you can't attempt to address the problem with your individual brain, but surely for most issues it is preferable (and much more practical) to rely on social epistemology, considering how many assumptions we are faced with every day. Is the Earth round? Will my car explode while I am driving down the freeway? Do vaccines contain microchips? I don't have the expertise, nor the time to establish the precise level of consideration to give all of the these questions.
"Some people try to dodge the question and say that all rationality is basically a social process. Maybe on my own, I will naturally stop at whatever level seems self-evident to me. Then other people might challenge me, and I can reassess. But I hate this answer. It seems to be preemptively giving up and hoping other people are less lazy than you are.”
I don’t think this is about being lazy. I think it’s about inferential range. Slack in inferential range has a clear tradeoff, which you well describe in Studies on Slack.
The easiest absurdity dismissal is that Neom is presented as this giant thing all at once. No city (maybe Brasilia?) ever just got planned and millions of people moved there. Now throw in a totally untested architectural model for this city, this style of living. It would not be absurd to build one section, in the most desirable location (near the sea) and see if it works, see if people like it. Then build another section and so on. It may not get to the length of Ireland. They may have learnings as they go and as residents give input. Newer sections may be designed differently. But if the economics or the functioning of this thing only works as a giant whole, it seems doomed.
The goal here is probably to minimize instances where we lock in a wrong position and refuse to update further without wasting too much time on analysis, I would guess. This is the position I would try to adhere to, at least, because like other humans I can't optimize for only the truth without serious loss to everything else I consider worthwhile.
I think the trick for things outside my fields of expertise is splitting up the heuristics, rather than serious study:
-Is it in my field of expertise, or is it something that looks like it is but is adjacent? I see a lot of experts make this mistake, where they comment on something in a related field, and end up wildly off base because they mistook it for their own field.
-Do experts in the field say its impossible? Being fringe is one thing, being impossible is another. Not infallible, but good enough particularly for hard sciences. I'm not going to necessarily understand *why* something is impossible without a background in the field, so it's impractical to learn it myself.
-If the claim appears particularly absurd: Do I understand the claim correctly? Some claims are as insane as they appear, but this is worth a double check.
I think all of that has to happen past the initial gutcheck of absurdity, though, so it seems more like a series of backup heuristics. I'd also be concerned about the opposite; blandly accepting a statement because it seems reasonable. I have a feeling the "reasonableness heuristic" is the more deadly of the two, but I have no evidence to back that up.
My point is that what you wrote didn't pass muster as something a NEOM proponent would read and say "yeah, that's a good point. I still think we will succeed, but if we fail, it will be because of these things Scott highlighted".
That example didn't have to exist for the argument to be wrong, but it was convenient that it did exist.
Another IMO not-very-good argument was to call the project impossible. Jumping from "this is very hard, I don't think the saudis can do it" to "this is impossible". Now, someone might say that this is a nitpick.
And yet, I think the difference between "very very hard" and "impossible" is everything we've ever done as a species that has been worth a damn. A sensemaking error that causes one to confuse the two is a sensemaking error that makes all the difference between a thriving civilization and one that stagnates.
I believe impossibility claims should be handled with care, ideally built on things about as certain as the laws of physics, because they're powerful tools of epistemic warfare
"But how do you, pondering a question on your own, know when to stop because a line of argument strikes you as absurd, vs. to stick around and gather more facts and see whether your first impressions were accurate?"
Here's the thing: you can just say "looks super hard, not convinced they can do it". And if pressed to make a decision whether to buy discount NEOM property, you can just say no. You don't need to jump to claiming impossibility to do that.
I even described a path towards constructing a stronger argument, by focusing on the implementation path rather than the endpoint.
And so, in a way, this piece makes the same mistake as the one on NEOM. You go in depth about the absurdity heuristic, and the result reads to me like you're not actually engaging with the point at hand: using weak arguments to support a position, regardless of the validity of the position itself. Had you just said "seems pretty hard to me, I'd bet against it" I would have been in complete agreement.
It really feels like much of this would be resolved much better with a conversation.
Speaking of which, I'd be pretty happy if you read my analysis of your ivermectin piece, ideally letting me know where I'm misunderstanding you and/or correcting the elements you agree are in error. I'm learning a lot while doing it and I still think there's a lot of good to be done by engaging in open-ended discourse about the topic at hand.
Back in 1969, I saw Paolo Soleri's exhibit at the Museum of Modern Art in NYC showing plans and models for what he called arcologies. These were immense buildings, as tall as skyscrapers and running across the land for miles and miles. They were an an update of Corbusier's Radiant City or Buckminster Fuller's urban structures with an attempt to think of a city in ecological terms, as forming its own world and lessening its impact on the natural world. I think the name came from architecture and ecology. The idea was that such a linear city could provide housing, controlled environment agriculture and recreation, a transportation system and all the infrastructure needed for urban life with a minimal ecological footprint.
There were a lot of big thinkers in the 1960s and a lot of fascinating new ideas, many of which we take for granted today. Arcologies are not one of them, though I noticed a reference in Neuromancer. Neom, for all its grandiosity, offers the same appeal. Saudi Arabia is not noted for its pleasant climate, abundant fresh water supplies, or agricultural production. An arcology could offer better climate control, more efficient water use and possibly a platform for urban farming. Neom is not absurd or inconceivable by that standard.
I think the main flaw is that the design does not seem to provide a pathway for organic urban growth. What is the minimum unit that could be built as a useful prototype, something more than Habitat '67? How would it work as a test bed? How would it fit with the nation's traditional urban life? How would it deal with the large number of foreign workers who are major part of Saudi Arabia's workforce? Then there's the matter of the variability of oil income, but I'm not a Saudi staff economist, so that isn't my problem.
It helps to remember that absurd comes from the mathematical term surd or irrational number. These were a real problem in mathematics for a long time, but nowadays we work with surds all the time and barely notice their absurdity.
(*) Soleri's work informed a lot of the model design for Star Wars. Compare some of his arcologies with some Star Wars battle cruisers and the like.
I think Alexandros Marinos is actually right in the case of the Neom post. I believe it's usually most productive to approach a disagreement as such:
1. State the opposing side's view, as far up in the Graham hierarchy [1] as possible
2. State your point of view, as well as possible
3. Rebut your opponent
4. Optionally make fun of your opponent
As far as I can see, 1 was done minimally, 2. was part of the post, 3. was only minimally present (absurdity heuristic), but it was *heavily* entangled with 4 (which is probably the worst kind of entaglement possible in a disagreement: Even singling out the the sentences "I disagree with this because of the absurdity heuristic. I will now proceed to make fun of Neom." would have been more productive.
Unfortunately, it seems like entangling 3 & 4 is psychologically easy to fall into (I've too often fallen into that trap, and imagine that'd continue if I garnered more status), and gives you the most clicks, so avoiding it is *hard*. I've seen some comments suggest that the ACX posts lean more into this than the SSC posts did, this feels true to me, and I want to *gently* point to CONSTANT VIGILANCE [2].
I made an estimate of how much lead we'd need for a national lead-acid battery. It's more lead than is known to exist in the world.
I made an estimate for how much a national lithium ion battery would cost. It's over 100 trillion dollars, and it would take a large fraction of known lithium reserves.
I made an estimate of how much pumped hydro storage we'd need, instead of batteries. It's absurdly large, the national reservoir size would be equal to draining Lake Tahoe just to store 3 days of power for the country.
So far, the most common response has been:
"You're just knocking down strawman arguments, of course we wouldn't solve the problem with 1 thing".
When my point has been more like:
"Solving green energy and intermittency looks really hard and every solution looks expensive. We need new solutions, or maybe we need nuclear"
My favourite absurdity is how my job involves making a special type of rock do my bidding by feeding it patterns of lightning. Lots of my time and effort is spent making sure that the lightning is just so, otherwise the rock misbehaves.
Quite a few people are preoccupied by worrying that the machines with these rocks at their heart might one day go out of our control and kill us all.
I think that the virtuous rationalist path is most likely to do a large amount of research on Neom down to the construction physics level, then write a Much More Than You Wanted To Know post, then get feedback from others, add corrections as needed, and then rinse and repeat because you can never be sure that you didn't miss something.
The problem with that is that we are human, we don't have the time, energy, nor inclination to do deep dives, and reality has a surprising amount of detail (http://johnsalvatier.org/blog/2017/reality-has-a-surprising-amount-of-detail). So, we rely on heuristics based on our priors. This includes the absurdity heuristic, even though that heuristic's poor track record is a major theme of the rationalist canon (ie, how heliocentrism, evolution, special relativity, quantum mechanics, etc. seem so absurd).
I think the idea is that there is usually never a royal road or even a muddy goat path, but there is always a compass. That compass always points in the same direction, labeled "Do more research, and listen to more criticism and feedback to make sure you are not going astray". When looking at the direction the compass is going, we see difficult terrain, lots of skulls (https://slatestarcodex.com/2017/04/07/yes-we-have-noticed-the-skulls/), terrible weather, and it goes uphill both ways.
I don't actually think we should be on the most virtuous rationalist path all of the time. After all, we have a limited budget of time, energy, money, etc. However, even following it a little bit of the way is probably good.
For instance, instead of "Neom is so absurd! Ha ha ha!", it would be better to say "I think Neom is absurd because of A, B, C inefficiencies, which if we look at ordinary construction costs would cost $X, and make it hard for residents to get around because of Y. Even if the Saudis have super-secret extra-efficient construction technology (unlikely because of Z), constructing a more ordinarily-shaped city with walkability and transit in mind would still cost Alpha % less and increase mobility by a factor of Beta for reasons Gamma." At least for me, sometimes the absurdity becomes more apparent when the critic lays out the facts than when the critic points and laughs.
As for how far we should follow the compass of Rationalist Virtue given our time/energy/money/inclination constraints, maybe here there is really not only no royal road nor goat path, but also no compass. I suppose it depends on things whether we are trying to persuade or get to the bottom of things, how important is it personally to know the truth (ie, getting Neom wrong costs nothing to anyone here in the West, at least for now), and how certain we already are (this last one can still go wrong, since closed-minded bigots are very certain of their beliefs).
The point is that we should always see absurdity as a cop-out. Maybe we can agree that it is the best we can do for now given a large multitude of constraints, but we should never accept it as a good argument in and of itself. If we stop at a certain region when making a Map of Knowledge, we should never mark that region as "Here Be Absurdity". Rather, we should mark it as "Here we gave up", or maybe include a pointer to a cache of arguments someone else made. In the latter case, we would let the reader/future mapmakers decide if the pointer points in the correct direction.
Also, I specifically disagree with "in order to dismiss it as absurd, I need to explain why the Saudi government would waste $500 billion on an obviously absurd idea. This is easy: their king is a megalomaniac, plus people are afraid to voice dissent."
Yes, MBS really is a megalomaniac, and yes people really are afraid to voice dissent, and yes, all of those are good reasons why the Saudi government would waste such a sum on an absurd idea. However, these are not by themselves reasons to dismiss Neom, and Scott himself noticed the red flags of Bulverism and bias arguments. To properly dismiss Neom, it would be better to discuss urban planning, economics, and construction physics.
Neom seems more like a concept with some strong statements, but we've seen similar grand scale plans fail before or simply never come to fruition. Does anyone remember those artificial land palm trees? They're...not doing so well and were never finished. So until ground is broken or a highly detailed multi thousand page report is issued, then we can dismiss any concept car or concept arcology building out of hand based on priors that such things never come to market or look very different if they do.
It seems like we might match our level of effort in dealing with a specific claim to some dgree with the amount of specificity being offered.
Do they have a detailed list of supplies, purchase prices, and hours of effort by various categories of builders to construct their price estimate with comparison rates of real projects? Or is it just a number pulled out of thin air? If there are some 'high level' graphs then that's pretty low specificity and lacks data to support it from prior building costs in the area, as Scott mentioned.
Now you don't want to get sucked into a rabbit hole by some crazy person who writes endlessly long lists of things and infer that somehow makes what they're saying any more or less likely.
But it is easy enough to make some guess as to how seriously you choose to take any pile of estimates and if it is too complex, to rely on experts to analyse such a complex argument. If you see a brochure with 30 pages of pretty pictures and pie charts....you can push back to say it is absurd relying on your superficial priors...but if you have a 3,000 page plan then a bit more effort will be required to analyse it for overly rosy or pessimistic projections. This is how political think tanks generally work to promote highly biased and often false ideas using detailed reports to support the foregone conclusions their sponsors wanted to promote.
I'm going to offer an anecdote that might be slightly relevant, but I sincerely hope is not. Mostly I just want to tell this story....
Sometime around 2007-2008, a friend of mine was hired as a contract welder and fabricator in (I think) UAE. He was working on the enormous stage for the wedding of the Princess. Six months. Enormous.
He had several trenchant observations about the culture there, but the semi-relevant one in this case is about "labor conditions". 'Berto was working steadily, but saw there was no way all this work could get done, so he asked up the chain for some help.
A few days later, he got some help. A hundred (100) Sri Lankan laborers showed up, and were put at his service. These guys had only the clothes on their backs, and no skills at all. They were put to work sometimes pounding in screws with hammers. They worked frantically hard. 'Berto was deeply troubled by the implications of this. After six days of this, there was the collective day off. Then on [Monday? after the day off anyway], a hundred COMPLETELY DIFFERENT Sri Lankans took their place. This went on for almost two months.
My friend prayed the whole time that each hundred laborers were not just fed to the sharks after six days.
The fact that he considered that might be possible, well it might have a bearing on construction economics in the Arab World.
"Marginal spending on medicine has zero net effect" is only absurd if you don't know what "marginal" or "net" means, right? It means "we're spending roughly the correct amount of money on medicine" (whereas marginal spending having negative net effect would mean we're spending too much, and positive that we should spend more). Or am I missing something?
I think the main thing this discussion tells us is that there are many ways to analyze a project, and people are EXTREMELY convinced that their "rational" way is the single canonical rational way...
- One dimension of this is the Saudi government and its various out-group properties. The all-consuming issue to about 90% of the commenters, regardless of specifics about the project.
- A second dimension is the cost. The all-consuming issue to about 9% of the commenters.
- What very people appear to be interested in is the city qua city aspect of this; yet to me that's the part that obvious renders it "absurd" regardless of bias or otherwise.
Consider the pictures as drawn. Not how you you want save the project by redefining, the project as described.
What you are living in is a TUNNEL 170km long. You are living in a skyscraper/shopping mall that you can never leave. Is that what people want, especially the kind of people who rhapsodize about this sort of city?
How are things delivered in this city if there are no cars? How are buildings repaired if there are no open spaces (eg roads) that act as emergency locations for cranes and building equipment? How is air pumped into the city and heat removed? (All this stuff can be done for skyscrapers because they have huge open spaces on the sides where the equipment lives and can vent, but according the video two of the sides of each "skyscraper unit" don't exist, and the other two are mirrored glass with no ugly chillers or pumps.
What are the fire escape plans?
Do animals and birds live in this space along with the trees? How does that play out in terms of animal waste and falling leaves?
etc etc etc
This stuff is considered absurd because there are no intermediate proofs of concept. Yes, you can grow one tree in a skyscraper, but no-one is growing forests in skyscrapers. Every real-world image I see of this sort of thing ACTUALLY has the trees on the outside of the building, so the problem is moot.
It's not quite the same, but the logic here is somewhat like "we know how to build gardens, therefor how hard can it be to build Biosphere 2?" And we know how that turned out.
Now if you want to say that the images in the video are all nonsense, well sure. Then we can have a different sort of argument. But the argument at least some of us (1%) are having is with the concept as presented by the video. And we call absurdity because of these eminently practical things: air, heat, life forms, repair, logistics, building evolution, ...
I think an underrated aspect of the Neom project is that it was developed my western consultants- it's worth making fun of, but predominantly because it shows the lack of imagination with the western mind.
There's a difference between things that 1. seem absurd but fall in realms where fundamental understanding is pretty poorly developed and there is a broad possibility space for surprises and 2. things that seem absurd and fall into very well understood spaces where surprises don't really happen anymore.
Centrally planned economies and modernist urban planning fall squarely into the category of "ideas that were once groundbreaking but have since been pretty conclusively shown not to work for reasons that were pretty obvious in hindsight", and Neom basically leaps into the arms of several extremely well understood pitfalls with no plan to address them.
Add to this that the project can serve many individuals' personal ends without actually succeeding and you have the ideal recipe for money-incinerating white elephant boondoggle nirvana. This is not "ambitious but doable in principle". This is "obviously terrible idea borne of rich-kid hubris on an unprecedented scale".
All that, and not a single remark about drawing bright lines in the sand.
> Maybe, every so often, do a deep dive into fact-checking something, even if you’re absolutely sure it’s true. Maybe if everybody does this, then someone will (by coincidence) catch the false absurdities, and then the social epistemology thing can work.
I do not recommend this. I did this with "Donald Trump is a Racist" and even though I hate the man and 80% of what he stands for, now half my in-group thinks I'm MAGA. For completeness, I went a little further down the rabbit hole than Scott did in his "You're Still Crying Wolf" essay, and started providing videos of black people saying Trump isn't racist.
On the other hand, much of my in-group isn't rationalists, so maybe that is where my error lies.
In any case, be prepared, after you do a deep dive fact checking things that are "obviously true" to your in-group, that you will have some very negative social interactions anytime you bring up what you found, if it doesn't turn out to be actually-true.
What you're calling an "absurdity heuristic" just seems like a very specific reframing of Occam's razor. An explanation like "there's a Saudi conspiracy to develop amazing construction technology and hide it from the rest of the world" is complex, whereas an explanation like "really rich people sometimes do stupid things with their money" is simple. It shouldn't be mistaken for a foolproof way to identify truth, but it's a useful shortcut.
This post looks like it's an exercise in something like ethics, or a values system. Rationality as a whole promotes a set of values (like any religion), but these values happen to be restricted to certain domains, namely processing and communication of information.
In this case, the 'absurdity' filter can be understood simply as, you have to use your value system to decide which hypothesis are worth exploring and which ones aren't.
If i try to point this out directly - that we are articulating and acting on values _just_ like religions do - people will usually say, no no, this is just about instrumental rationality. If that's so, then why 'obligations' to others? And what's wrong with 'the cowards 'way out?
> But I hate this answer. It seems to be preemptively giving up and hoping other people are less lazy than you are.
Words like 'lazy' and 'coward' have an obvious moral valence to them. Wouldn't a more fair interpretation be something like 'people are different, they will naturally explore different parts of the world, and you only have so much time and attention. You can't _not_ use your value system when determining which hypothesis to explore and test, and which ones aren't worth the expiration'.
Clearly, you are articulating norms and values here. You are trying to explore what is and isn't good, in the realm of reasoning and communication. But do value systems constrain anticipation? If not, then in this community they must live in the Jungian shadow; forever preventing us from fully introspecting what exactly the rationality project is about: defining a religion with truth as its deity.
"Extraordinary claims require extraordinary evidence." Doesn't that give us the good parts of the absurdity heuristic, and leave out the bias?
Relativity sounds absurd; but if you show me the textbooks and the experiments, I can see it'd take an awfully large conspiracy to be lying about it, especially when the math seems to check out. It's a big claim but there's lots of evidence offered, much of which I could check for consistency, some of which I could even replicate (Michelson-Morely!).
Neom sounds absurd, and there's no evidence offered that the Saudi king can make it work, and plenty of explanatory power in his own vanity.
In either case, we're just saying "okay, is there big evidence to go with this big claim?"
Admittedly, sometimes a claim sounds absurd because it sounds bigger than it really is. Relativity doesn't noticeably affect time under ordinary life conditions. Evolution doesn't literally mean a monkey gave birth to a human. AI safety doesn't literally mean the movie _Terminator_.
But extraordinary claims do require extraordinary evidence. "Absurd!" is just shorthand for:
"Why do you expect me to swallow that without first showing me evidence?"
Typo: extra parentheses
"This is easy: their king is a megalomaniac, plus people are afraid to voice dissent)"
If the absurdity heuristic is just "don't dedicate mental resources to things that have a tiny probability of being true/coming to pass" (with an asterisk about tail risks), I think the 1901 analogies shouldn't really make you apply it less.
Of the set of stories which would sound just as absurd as the internet one if told to someone in 1901, only a vanishingly small fraction came true. It may well have been rational to dismiss the internet story at the time, and saying otherwise may just be hindsight bias.
The word "absurd" is covering up at least two gradients of variation -
1) just how unlikely are we talking? A worldwide network of prostitute-dispensing-UFOs is unlikely, but it's practically a certainy compared to the likelihood that gravity obeys color. Relativity, _being true_, was considered very absurd for quite some time, and honestly, still feels a bit absurd.
"The Line" looks unlikely to succeed by its own metrics, as-is. But I might be willing to put down a bet at 1000:1 odds.
2) What timescale are we covering? in 1900, a worldwide network of adding machines exchanging representations of pictures _would_ be absurd, because several technological, economic, and social changes had yet to take place, and the 'absurd' thing wouldn't come to be for about a century.
"The Line" isn't a claim that "the world of 2120 will heavily feature engineered mega-cities that would look to the world of 2020 as if they were sci-fi dreams", it's a claim that _the thing we're breaking ground on_ will work.
---
The post seems to be asking "how 'absurd' is so absurd we can stop discussing it, or stop thinking about it, as a serious going thing?"
And on that, my rough heuristics would be
- On a timeline under 20 years, the cutoff for "too absurd to be worth discussing" might be somewhere around the 1:10,000 mark. This can be tweaked based on the cost involved.
- On a timeline over 20 years, the sky's the limit, but on the other hand, nothing can really be said with any certainty anyway, so it's _less_ useful to talk about.
In the example of "The Line", I proposed that it might be something like a 1:1,000 chance of success. This might be considered non-absurd, except for the fact that it's such a high-cost endeavor, in terms of human time, money, ecological impact, and the upside so mundane, that it _should_ be laughed out of serious discussion until its proponents have come back with a much stronger case for why it should be possible and valuable.
As another example - earlier this century, there was some buzz about having possibly discovered a reactionless drive (https://en.wikipedia.org/wiki/EmDrive). This was very very unlikely, that we'd discovered something that violated laws we believed inviolable, rather than being measurement error. let's say 1:10m. But, the potential upside was also very high - and the cost of taking it seriously was relatively cheap - some days or months of laboratory time. The EmDrive may have been absurd, but perhaps not _so_ absurd as to preclude serious discussions about it.
"Even Alexandros agrees it probably won’t work."
Yes, but he doesn't agree that it's absurd, which is a different threshold. (His exact words are "Like, I'm not saying Neom will be a huge success or whatever.")
He's accusing you of cognitive self-dealing (roughly, of straw-manning where you could have steel-manned), and I don't think your discussion above clears you of that charge.
It's one thing to say Neom's design is absurd, but quite another to say "murdering the people who previously lived in the area" in reference to a single man who, according to the Saudis (and AFAIK not denied by anyone), was armed and shooting at police. You don't need sophisticated philosophy to see how misleading this framing is.
I like Neom project. I hate living in cities although I currently live in the city myself. It is just because like majority of people I am not rich and for economic reasons, transportation issues, proximity to work, shops etc. I have to live in the city although I really feel that it is like a life in prison.
Neom, the Line is a fantastic compromise. From one side you will be close to nature, even if it is only a view over the desert. Maybe they can put some irrigation and greenhouses to grow food for the city that will make the view slightly greener. On the other hand it would be a city infrastructure. The space between two walls would be like a street with shops, bars, offices and residential quarters. It is a fantastic project.
Now back to the reality, of course it seems absurd to think that it could be built with merely 500 billion dollars when the real costs are at least 10 times of that. Maybe if they could built the city only 17 km long but even that is unlikely due to corruption and general unprofitability of this project. But it is a different thing to admit that the project is not feasible for various reasons rather than to think that the idea itself is absurd. The US might never have free-for-all universal health care but it is still not an absurd idea.
It is like thinking that Egyptian pyramids are absurd. One estimate was that building a pyramid today would cost 5 billion dollars. It is a lot even today when we have big machines to move stones and project management theories how to organize big projects. In the past it must be unbelievably expensive and basically an absurd project. And yet, pyramids were built because some powerful people believed in the necessity of building them. Who knows if Neom could become a reality if some powerful people believe in its necessity? Or it could be some other seeming absurd project. Or in contrary, if they lose faith in this or another project then “absurd” becomes self-fullfiling prediction.
Comment on the comments: There are now at least two comments which say something like:
"<correct sounding general statement about absurdity>, as an example consider <thing I consider absurd/non-absurd but which a significant fraction of people disagree with me about>."
Come on surely you guys can figure out examples which all readers will agree with you on.
It looks bad that a group that happily discusses immortality, The Singularity, Roko's Basilisk, a universe turned into paperclips, etc., etc., cannot muster any sympathy for Neom.
It is possible for there to exist a statement X that is true, and yet also anyone who believes statement X is true is crazy, because craziness is judged on not based on whether the statements you believe are true, but whether the process by which you arrived at them is crazy. If in 1901, someone says that they know for a fact that in 100 there there will network of blah blah blah adding machines, you would be right to judge them as crazy, because there is no plausible non-crazy chain of evidence and reasoning that could lead them to this conclusion. If you are Einstein, who has just finished deriving the speed of light, but not yet told anyone about it, and someone comes to you with this statement about relative speeds that they know for a fact is true, concluding that they are crazy would still be reasonable, though you'd likely want to consider other explanations such as "they've been spying on me" and "they are a brilliant physicist that I somehow haven't heard of". For a more modern example, if you told me that you know for a fact that there will be a AI singularity by 2050, I would probably conclude that you are crazy, though I may also consider the explanation that you are wildly overconfident, or don't actually know what a probability is. Compared to if you told me that you believed there was a distinct possibility of AI singularity by 2050, and it was more likely than not, I would not call you crazy. Whether there does end up being an AI singularity by 2050 is irrelevant, because the process by which you arrived at the conclusion is what makes you crazy or not, not the actual conclusion. I also wouldn't hire someone to invest money on my behalf _just_ because they correctly predicted some low probability event far in advance, I would look at their process for generating predictions. If it seems like a process that is capable of consistently generating correct predictions, I will judge them a genius. If they (earnestly) tell me that it's because they are communicating with themselves from the future, or god, or "hidden energy", I'd probably judge them crazy _because they made a correct prediction_.
I think there is a difference between "I think this is true now"/"I think this happened in the past" and "I think this will happen/be true in the future."
My current theory on "The Line" is that someone mixed "Inspiration for the next important project" and "best recent video game featuring the middle east".
On the original quoted example about absurdity bias, I think that a smart and well educated person from 1900 would be able to distinguish the truths from the lies there if they were allowed to ask follow-up questions, thoroughly interrogating our time traveller on the details of the supposed future. What happens if the shade of blue is slightly off? What equations exactly govern the distortion of space as you approach the speed of light? What has changed the economics of male prostitution so much as to make this scheme economically viable? How do you convert lesbians into numbers?
You can make anything sound absurd by describing it in a sufficiently wacky way, but reality holds up under interrogation while absurdity falls apart.
(What if those were bad examples and you could make up an absurd scenario that _would_ hold up under careful scrutiny? Then it probably isn't all that absurd. You could probably convince the people of 1900 that the Austro-Hungarian Empire dominates all of Eurasia in 2012, but that's not absurd, it's just untrue.)
All you need to do is say that you don't see how it's possible (and give the reasons why you think it isn't possible). Saying "it's absurd" is essentially conflating opinion with fact. It's adding "something extra" that's not necessary, that's extraneous.
This is ubiquitous on the Internet, and in the long run doesn't add value.
Even "In my opinion it's absurd" is better, but still sub-optimal; you need to account for arguing from personal incredulity, and "I think it'd absurd" can run afoul of that problem.
I like to spend a minute on a back of the envelope calculation wherever relevant. If anyone is seriously proposing X, and I'm writing a blog post criticizing X, I will apply at least a BOTEC level of analysis.
Kingdom Centre in Saudi Arabia is 76.8m * 302m * 37.8m and cost US$453 million in 2002, multiplied by 1.65 to adjust for inflation through 2022. This comes to $850/m^3. I previously calculated ( https://astralcodexten.substack.com/p/model-city-monday-8122/comment/8116431 ) that the Neom Line would be $58/m^3 if they spend $1T. So it's off by an order of magnitude, but maybe still possible to do the project if they can sell the condos for more than the cost of construction. They could build 1% of the line, sell condos and office space to get their construction costs back with a profit, and rinse and repeat until they run out of buyers. Or with enough loans those steps could be somewhat parallelized. Real estate developers usually use tons of leverage and sell houses whenever they're completed rather than waiting for the entire subdivision to be completed.
I think Scott's prediction "Saudi Arabia builds a structure at least 100m x 100m x 1000m before 2040 or the Singularity, whichever comes first: <1%" was very overconfident. 100x100x1000 would cost only $8.5 billion at Kingdom Centre's $/m^3. MBS can easily spend that much on whatever he wants to build so long as it's possible with current technology. I would put the odds at 50% for the same prediction.
Why be wary of the social epistemology? Do you hope to contain all human knowledge in your brain one day? Why would you hope to be able to do epistemology by yourself? I think rationalists should embrace social knowledge and epistemology, treat humanity as one huge complex sentient machine.
When I read the bit about Lee de Forest’s fraud prosecution I was intrigued and wanted to find out more. However the link from Eliezer’s article is now dead and other sources I found suggest that while he was found not guilty, his business partners were not so fortunate.
The actual quote provided seems pretty convincing evidence for the proposition ‘some people found his claims so absurd as to be prima facie evidence of fraud’ but I have no idea where it comes from.
Does anyone know more about this?
This reminds me of the great takedown of Hyperloop that Alon Levy wrote almost a decade ago now: https://pedestrianobservations.com/2013/08/13/loopy-ideas-are-fine-if-youre-an-entrepreneur/
It looks like the CA HSR project has slightly more than doubled in price and added a few more years since then, and Musk has shown that he can in fact make money selling cars, but there really hasn't been any more support for his Hyperloop idea.
I feel like the quoted example of absurdity bias does not fit well here. The examples he quoted are clearly formulated to be absurd, mostly by being oddly specific. Take the black sphere one, for example: These might simply be floating houses or teleporter endpoints. It's quite obvious that some of these houses would also house prostitutes and therefore you'd be able to order one from there.
It's the same thing about the internet example: This is even a stretch as it is - is one of the primary use-cases for the internet really lesbian porn? Why specifically mention encoding or the 69 meme? I'm not even sure which one is referenced here. If, on the other hand, you told that persons that we found a way to make our computers/calculators really, really fast and connected them all up and suddenly that person will probably not be as skeptical. Then you might also add that people also use it for sexual content and it turned out that's something that's quite desired, depending on that persons view they probably won't even be surprised.
The presentation of Neom doesn't fall into the same category, since they're not presenting it by making it intentionally overly absurd and focusing on specific, minor details to detract from the somewhat reasonable big picture. They're completely serious. They also don't seem to bother to explain the somewhat ambitious aspects like the super fast travel and the strange shape. Giving these aspects a pass because, if you just formulate hard enough, you can make everything sound absurd is not a good argument IMO - it's just not an apples to apples comparison.
---
On whether we should accept absurdity as an argument, I think this is actually pretty simple: It's basically a (slightly smug) way of saying "I'm gonna need some evidence to believe this". If your point isn't really important for a decision, as this stated opinion probably is, I don't think it's necessary to dig deep into a topic when the fact that this is pretty absurd is obvious to the casual observer. It'd be different if your opinion did matter a lot (a judge shouldn't be as easy to accept this argument, for example, and neither should your investment advisor) or if some people you trust point out that you might be wrong with this.
I felt as though the absurdity argument against Neom was a combination of multiple arguments, which are not entirely spelled out because any of them would work and it's assumed that most readers can recreate multiple of them without help:
(1) People who are experts at this have looked at this and declared that it will be much more difficult than is claimed.
(2) There is lots of low-hanging fruit: obvious changes that could obviously improve the design. One example mentioned in the comments is turning the thing on its side: it's easier to build a 200 m tall and 500 m wide city than a 500 m tall and 200 m wide city. The reason you build up is because there's not enough land in the city center, but there's plenty of land here. Another example is that it could follow the coastline instead of being perfectly straight with a mountain range in the middle. It doesn't follow the coastline because that's where all of the palaces for the Saudi royal family go.
(3) The project follows a known failure mode. This particular failure mode is called High Modernism and has been well described in Seeing Like A State by James C. Scott (1998).
(4) The project involves a common human experience, where our intuition should be trustworthy. In this case, Neom wants to be a city. We can ask: is this a city that I would want to live in? I'm guessing that the answer to this for Scott is no. This is a less solid argument because different people might think different things.
(5) There are other underlying problems that could ruin the project does not address. In Neom's case, the biggest problem is finding people to live there. They're not going to come from Saudi Arabia itself: the country is barely above replacement fertility rate (2.28 children / woman) and falling, and other cities in the country are also expanding. So they're going to have to attract foreigners. If it's a status project, they probably want elite foreigners, not Syrian or Yemeni refugees. But a lot of global elites do not want to live under the Saudi's authoritarian and Islamic government. Even for those who are willing, Neom has to compete with other elite-seeking Middle Eastern cities like Dubai, Riyadh, Abu Dhabi, and Jeddah.
I personally think that Scott is underconfident that something will be built (it should be >1%, maybe 50%) and overconfident that a significant number of people will end up living in whatever is built (it should be <75%, maybe 25% ... although maybe 50,000 is small enough that Scott's estimate is reasonable). Building a structure that is at least 100m x 100m x 1000m is technically possible with enough money. The Saudis have enough money. So something will likely be built. But persuading people to come will be more challenging.
1. When someone makes a facially absurd claim it is reasonable to dismiss it out of hand until they explain in detail why it is not absurd.
2. When a large amount of money backed by smart successful people is invested in something that seems facially absurd you should carefully consider two possibilities:
a. It is not as absurd as you thought. For example, I thought the idea of Elon Musk successfully and profitably creating an electric car company from scratch was absurd. I was wrong.
b. The stated and actual purpose of the investment are different. For example, in third world countries many large infrastructure projects have stated economic goals that are facially absurd. They fail at those goals. But they do an excellent job of meeting their actual goals to funnel money into the pockets of well connected people.
I do not think NEOM is a graft - the people paying for it are the best connected people. It is interesting to consider what a "failed" or scaled down NEOM project might actually accomplish.
For example, could NEOM eventually morph into a planned city (like Brasilia) that operates as an almost extra-territorial enclave within Saudi Arabia complete with world class universities that recruit globally, offer full rides, are decided non-woke, and admit based on merit alone? Could those be the core of a Silicon Arabia? Could NEOM have, in effect, a separate peace with Israel so it can bring in Israeli investment and expertise? Now that would be interesting.
One of the first steps in making sure you're applying the absurdity heuristic appropriately is just to make sure to steelman the position you think is absurd. E.g. "I know that monkeys can’t turn into humans, this is so absurd that I don’t even have to think about the question any further” sounds a lot less absurd when you understand that monkeys turning into humans isn't what evolutionary biologists are claiming is going on.
Monkeys magically transmuting into humans would be absurd and our anti-evolution guy would be right to dismiss it as absurd if that were the claim.
I think the missing piece here is that beliefs come with levels of certainty. The question of how far to go down the chain is about the degree of certainty that you want to achieve.
The claim that Neom is absurd has one level of confidence attached to it, but, as you point out sometimes absurd seeming things happen. Ok, so you work out the construction costs and that makes it even more unlikely it will happen, you supplement with data about the fact that Saudi construction costs aren't much cheaper than elsewhere and you can increase that level of confidence.
Now that isn't to say you can always increase confidence by digging deeper. That's why you stop when you get to assumptions like: but what if the universe is all a dream or God intervenes and violates the laws of physics. Because digging into those questions doesn't actually (or shouldn't) result in substantially increased confidence about your answer.
Yes, at the end of the day you can't get below your own priors but you can replace a judgement like:. silly sounding things tend not to work, which might only give you a relatively weak degree of certainty with judgements that increase your degree of certainty.
(Not that since probability distributions assume logical omniscience you have to supplement them a bit with some kind of judgement about the likelihood of there being some further deductive argument that's more compelling ...so I've played a bit loose with the term certainty which isn't quite just moving to a more extreme probability what it really is is something more like: evidence that further accessible information or consideration of argument will not lead to a refinement in your probability judgement...or at least that you don't expect the effort to be worth it).
Maybe a better way to put this is the following:. why isn't the obvious answer that you are balancing the rewards and benefits of further inquiry the right one?
I mean, you have certain judgements as to how likely further effort will be to refine your judgement and some payoff function regarding the degree of certainty. Surely, if you had to bet your life on the success or failure of the project it would make sense to expend more effort in research.
At each level you can't do anything but rely on your judgement (what would it mean to do otherwise) but part of that judgement is an evaluation of how likely further effort is to produce a more reliable answer. Do we need anything more than that mechanism?
The quote in the post by Eliezer says "This doesn't mean that anything can happen. Of all the events in the 20th century that would have been "absurd" by the standards of the 19th century, not a single one - to the best of our knowledge - violated the law of conservation of energy, which was known in 1850. Reality is not up for grabs; it works by rules even more precise than the ones we believe in instinctively."
I agree with his overall point that our current understanding of physics is such that some things seem incredibly likely to be universal laws. I have a bit of a nitpick, however, with his precise example of conservation of energy. It doesn't really affect the overall argument but I figured some people here might find it interesting nevertheless.
In classical mechanics, energy is a number which can be easily computed in principle and is conserved. In quantum mechanics, not all states can be associated with a definite energy. However, there *are* certain pure-energy states, and all states can be written as weighted combinations of these pure-energy states. The analogue of energy conservation in quantum mechanics is that the (technically, squared magnitude of the) relative weights of these different pure-energy states do not change over time. In particular, the average energy (computed via these weights) is conserved, and for large systems we can identify this with the classical energy, thus letting us recover the classical notion of energy conservation. This all works perfectly well when we observe isolated quantum systems.
The tricky part is when we also consider wavefunction collapse. Let's say we accept the Many Worlds interpretation of QM. Then, the observable reality around us is just one of a truly enormous number of wavefunction branches. Quantum mechanics only applies exactly to the *totality* of the branches, not separately to each wavefunction branch. Thus, the above argument for the conservation of the distribution over energies applies when you consider *all* of the many worlds but not necessarily when you only consider *our* world, which we inhabit. If we were capable of knowing the true distribution over energies of just our world, it would probably constantly keep fluctuating due to frequent collapse events. All we can say is that when averaged with those of the other worlds, the distribution over energies will be conserved. And yet, if we cannot observe the other worlds, is this definition of energy conservation even useful? One can avoid the question by saying that we only care about idealized, isolated physical systems, not the whole universe treated quantum mechanically. And yet, the latter is what we were trying to seek insight into in the first place...
At the other end of the scale, there are also problems with energy conservation in general relativity. I don't know much about this so I'll just link to an article on it by John Baez: https://math.ucr.edu/home/baez/physics/Relativity/GR/energy_gr.html
In Kierkegaard, there are a few separate epistemological stages. This post is a really nice gateway into his whole thing!
Immediacy: You are completely naive and unreflective. You don't yet know the difference between being and seeming. Only babies and very special adults are like this. If you're reading a blog post like this, you're not in immediacy.
Aesthetic: You've discovered reflection; that your ideas about things are different from the things themselves. You've found that reflecting critically on your own and others' ideas brings you closer to the truth, so you do that a lot. Eventually, you're in a prison of increasingly-subtle self-referential reflections, unable to get out of your own head. This is when you question all of your assumptions. He talks about Descartes' "doubt everything". Analysis paralysis, basically, and ultimately, despair.
Ethical: You can't sit around all day reflecting. You've got to do something sometime. What you do should be good for everybody and follow some kind of absolutely valid universal rules. You become conscious of yourself as having eternal validity, through your relationship to these universal rules. This is Scott's evolution denier; he doesn't need to question any further, he knows what he needs to know. But, since you can't actually live up to the rules (you're not perfect, and anyway life will trap you in impossible situations), you're still in despair.
(Here I'll pause to note that the fundamental tension described in this post, between naïveté and infinite regress, is very identifiably the same as the tension in Kierkegaard between the Aesthetic and the Ethical. How do you negotiate between these two? The answer is...)
Faith: Not as easy to describe. Still conscious of your eternal validity, still aware that you're sometimes wrong about stuff, you nevertheless choose, act and grasp hold of your life. You make a "leap of faith" which is more like a dance move than like an Indiana Jones puzzle. You know you're wrong to do it, but you do anyway. And somehow, you win! Or, if you don't, you don't notice much. For a Christian, this is something like thinking, "God put me here, now, and He gave me the knowledge I have, and set things up so I'm faced with this question; if I genuinely do my best, even if it ends up that I'm totally wrong, my sin is atoned for by Christ." Or as Kierkegaard says, you come to rest transparently on the ground of your being.
So, I don't know if other ways of looking at the world have answers to the questions raised in the post, but Christianity does, in my opinion, and Kierkegaard is the philosopher who articulated it best.
(Disclaimer: everybody is always wrong about Kierkegaard, I am not an authority, Kierkegaard would punch me if anybody believed what I said just because I said it. He would punch you if you believed what he said just because he said it. What he wanted least of all was disciples, which makes my fanboying pretty awkward. If this is interesting to you at all, read Kierkegaard yourself!)
Wait, this doesn't seem to make any sense to me. You're not deciding a binary here... not really. A good (rationalist) way to approach this would be saying "This is absurd, and (therefore) very unlikely to happen/be true" Pointing out that historical prognostications sometimes sound like they're unlikely to be true doesn't mean we shouldn't still realize that some theories are dramatically less likely to be true than others.
An absurd retrodiction: a dude with $200M (from the sale of Paypal, from which he was kicked out for "lacking a cohesive business model") would disrupt two overregulated capital-intensive industries all over the world within a decade and a half.
One line of reasoning I didn't see in the comments is using base rate. In such a case, it's not obvious what would be the relevant reference class, but I feel that "mega-projects that planned to spend similar amounts of money" and "building a city of that size from scratch, by centralized fiat" are good contenders. From what I've read by Bent Flyvbjerg, the odds are not good that this will be delivered on specs, on budget, on time.
Back to the question of the absurdity argument, I think that "if this project worked, it would be so far from its reference class that it's absurd to imagine it could succeed (in the near future, barring any major breakthrough)" is a reasonably applicable and successful heuristics.
Re. "Maybe, every so often, do a deep dive into fact-checking something, even if you’re absolutely sure it’s true. Maybe if everybody does this, then someone will (by coincidence) catch the false absurdities, and then the social epistemology thing can work." :
YES. Taking random samples of life is a tragically under-used technique. It's especially useful if it becomes a social norm.
Neom is something that you come up with if you're playing a video game where the goal is to pack as many people in as closely as possible, while damaging the environment as little as possible, while not respecting any constaints of reality. It's like a parody of the average consultant's mindest.
I realize it's Saudi Arabia, but presumably people still want to go outdoors in Saudi Arabia. I suppose you can take a few steps away from The Line and look back at the wall... and then walk back inside...
It's absurd that you didn't title this one "Reductio ad Absurdum". :)
Scott, I read your post as saying, Saudi Arabia thinks they can do this but I think it's really really unlikely.
Seems like an accurate assessment of the situation, and not a claim that no way no how maybe somehow Saudi Arabia has figured out something fantastic that will have ten million people living in a straight line
I think there's a practical exercise to quickly examine these sorts of biases; I call it a "pre-mortem" exercise. The exercise is simple: write two short memos, each 0.5-1 page in length. One is titled "why it worked" and the other is titled "why it failed". Circle back up and compare. If you all agree, those are prioritized strengths and weaknesses. If you disagree, you've unearthed differences in your assumptions! Or, in this case, differences in your absurdity lenses.
This exercise was largely inspired by the 'Safire Memo', the backup speech for if the Apollo 11 landing had gone wrong. Some credit also goes to the observation from the book Strategy Beyond the Hockey Stick, which observes that in big organizations that fell from the heights (Nokia, Kodak, etc) even when leaders individually knew that the new technology would win the organization still failed to bet that way.
Missed the chance to title this "The Neom-Central Fallacy, or, The Most Absurd Argument In The World".
More seriously, I think any sort of absurdity heuristic or bias becomes less relevant as the...complexity waterline gets raised. Without going into __any__ specific examples *cough*, the world just seems a lot more complicated today than [arbitrary past cutoff]. I guess some of this is an inevitable part of growing up - one can only hold so much information in the brain, and the rate of information-addition to the world far surpasses that. I think the slope of that latter line matters, though.
Even though in many ways it's easier than ever to Deep Dive: Much More Than You Wanted To Know, if the rate of increasing Unknowns and Unknown-Unknowns outpaces the rate of generating Knowns, then the total space of non-knowledge never shrinks. A whole lot of this can be safely written off via rational ignorance. But when even immediately practical and relevant things start to become Absurd, well, there's a problem. That's not knowledge that one can safely ignore; life's table stakes keep getting bigger. And there's so much bad-faith social epistemology now (classic update method from antiquity! insert evosci pablum here) that I'm kinda surprised to see it recommended as a potential fix. That seems to rely on one's ability to avoid a whole bunch of *other* biases...
Anyway, this feels like a post that coulda been avoided merely through slightly different word choice or a quick edit. "Define terms please" is next to spellcheck for godliness. The general free-for-all confused melee in the comments seems to indicate that while a springboard post to discussing Absurdity Bias was a good idea, Neom isn't a particularly good category member to base it off. Possibly doesn't belong in category at all.
Neom isn't a bad idea because it sounds impossibly expensive and difficult to build with today's technology. Any proposal like that might happen to get an Elon Musk to make it cheaper and change the whole equation faster than anyone could have imagined.
Neom is a bad idea because no matter how advanced our technology is, even if we're literally building Dyson spheres and are immortal and can build giant skyscrapers the height of Mount Everest in a day, a linear city is always going to be worse than a city that is literally any other shape.
Irrelevant nitpicking ahead: EY's examples in the Stranger Than History quote are bad.
In 1901 there were maybe physicists like Plank who had an inkling to what Einstein will propose only 4 years later because it was based on theories and experiments already available like the Michelson–Morley experiment of 1887 suggesting the Lorentz invariance of the speed of light. Plank was definitely not shocked by relativity.
Also in 1901 the world was crisscrossed by telegraph lines, telephones networks were growing and there were plenty of newfangled sophisticated mechanical calculators like the Millionaire of 1892 and electric tabulating machines using punch cards made since 1890 by Hollerith (his company will eventually by a founding part of IBM). Telling someone that in a 100 years there will be a network of electric calculators connected by wires would have not been a shocker and I'm pretty sure it was predicted by some at the time. And they did had all kinds of porn fiction in 1901, including lesbian.
It's not just about absurdity; it's about Bayesian probability. If you collect everything we know about physics and technology, and plug it into the Bayes' Rule, then we will discover that painting yourself a specific color is incredibly unlikely to reverse gravity. Of course, we could be wrong; but we would be foolish to make long-term plans or investments based on such a thin hope.
The tricky part here is that if we told someone from 1901 about the Internet, and he plugged in everything he knew about science and technology of the time into Bayes' Rule, he'd be forced to conclude that the Internet is a scam. He would be factually incorrect, but his conclusion would be fully justified. But, if he stuck with his rigorous thought processes (and somehow lived about 100 more years), he might be one of the people who got to *develop* the Internet.
There is no magic oracle that you can consult to get 100% accurate answers about any question. The best you can do is balance false positives and false negatives; and this means that you would occasionally miss the one cool revolutionary idea that will change everything. In exchange, you gain the kind of steady progress that gets you from steam engines to the Internet in only 100 years. It's a worthwhile tradeoff.
This post is IMHO completely unnecessary (and, sadly, showcases why "the substack quality has fallen compared to SSC"). Any modicum of "common sense" (horror!) would give you the same conclusions in 20x less words. This is simply not interesting and not why I read ACX (less and less these days, for reasons above).
I honestly don't understand the concern here. I mean, I do understand the object-level concern, I just don't see how the problem doesn't simply generalize to "P(A|B) = [P(A)*P(B|A)]/P(B), all the rest is commentary".
I mean, the only way I can conceive the problems posed in the post as problems is when I forcibly reduce the available options to simple binaries. Like, is "absurd" supposed to mean "literally impossible" - or just "having a very low probability"? Is "questioning your assumptions" a yes/no option where you either require them to have a strict logical proof, or require no proof at all - or is it more like having probabilities assigned to them, and calculating the need to revisit them from the current context? Isn't the evolution guy's problem the fact that he already decided he "knows" something, instead of admitting and processing new evidence?
In each of those cases, reducing to a binary appears to be the obvious failure mode, and eliminating that failure mode an instant improvement. (Now, in each of those cases, there's also no simple way forward after that. But I, for one, find reassuring the conclusion that there's no simple trick and we should just all follow the usual epistemic best practices.)
(I don't understand the problem with social evidence either, the vast majority of each of our individual knowledge was provided by other people, one way or another, there's no escaping that. Not to mention it would be extremely impractical to escape. Humanity runs on trusting others to do things one cannot do oneself. It got us far, certainly farther than each of us trusting only one's individual experience would have gotten us. It doesn't feel like a cop-out to me, maybe because it seems like a description of what's already happening, rather than a prescription of what should happen.)
I think I'd defend "use absurdity as a heuristic" in basically all cases. Taking Yudkowsky's examples:
Special relativity is completely absurd, and people should dismiss it out of hand until presented with a stack of evidence. No-one does this, because its so strangely beautiful you end up wanting it to be true.
If you explained the internet in more detail, most reasonably intelligent people in 1901 would probably accept that the invention itself would happen if it was possible. They'd accept the possibility wasn't absurd if they has a period-appropriate grasp of electricity. Anyone who was being intellectually honest would accept the porn thing as well.
The answer to the hidden example is that everyone thinks their own moral opinions (possibly all opinions?) are correct. If they thought they was wrong in some way, they'd change them to whatever they thought was right. "Moral consensus kind of drifts around haphazardly" is an idea that's clearly true and not absurd in the slightest, but people reject it because they want morality to be objective.
More generally, if you hear a very complicated argument for something ridiculous, you should still dismiss it because its more likely that the argument is wrong in some way than that the ridiculous thing is true. Neom is a bloody stupid idea because it's meant to be a desert city that's one block wide and the length of Ireland. If you show me a really good argument for why this is a good idea, then the right thing for me to do (even if I find the argument utterly convincing) is to reject the argument nonetheless and fall back on it still being a bloody stupid idea and you being a sophist.
There's a difference between it's impossible and it's impractical. A difference between being on the edge of known physics and being central to everyday experience.
There's very little which is uncertain about this project from an engineering and physics perspective. The 2 main areas I'd be worried about would be wind loading (absolute and differential) on a surface that large, and the environmental impact. The first can be mitigated by over-specifying the design for that force (and we already know how to build moment frames). I presume the latter will involve a lot of sand piling up in annoying locations which can be addressed with standard ground moving equipment. If nothing else, it will be a great way to learn more about environmental science.
What makes the project absurd isn't the physics/engineering which, granted, a lot of people are unfamiliar with. It's the parts that people are familiar with - everyday living in urban environments. And approximately everybody seems to think this isn't a place they would want to live.
Everything connected together with doesn't have to be terrible. The Las Vegas strip has a lot of major attractions connected together with a monorail. Lots of varied businesses connected indoors can work very well - the Minneapolis Skyway system is fantastic. But for individual large structures the advantages are limited. Skyscrapers exist because the demand for an area means the cost of land is higher than the cost of building upwards. Malls exist because there's an advantage of having lots of otherwise-unrelated stores in close proximity creating a network effect. And Amazon Warehouses exist because there's a need to be able to store a wide selection of stuff in an area where it can be individually boxed up as needed. What's left? Airports, because the minimum cost of decent runway is best amortized over a lot of flights which are themselves best served by a lot of jetways?
This is in contrast to other wacky projects where the objections are related to engineering factors like cost rather than experiential. Approximately nobody is objecting to Elon Musk's tunnel building projects based on the idea that "tunnels can't work" or "nobody wants to take tunnels". A lot of people are familiar with subways, highway tunnels, etc. The objections are on the engineering side about the ability to reduce the cost of development for the given service level/capacity and safety required.
If this was simply an infrastructure which pre-deployed subway, water, sewage and other utilities in the desert under a "if you build it, they will come" model it would still probably be a gigantic waste of money, but at the very least it would be preparing for a future which might possibly occur and meet the needs of people.
(I got an email about the comment I'm replying to here, but I can't see the comment on the actual page. I wonder if it'll show up once I reply to it? EDIT: comment system is bugged; this comment was made as a reply to a reply to my top-level comment, here: https://astralcodexten.substack.com/p/absurdity-bias-neom-edition/comment/8185050)
Wow, you're totally right. Chapman is in a very Kierkegaardian place. His solution to the whole problem, "non-systematic situated meaning-making", he draws from Heidegger, who was inspired by Kierkegaard. So it looks like he's getting some Kierkegaardian insights at second hand.
Hegel had noticed the problem of nebulosity, that no system is actually true and right, and had responded to it by making a system of systems; this giant uber-system which would account for and rationalize all possible systems, and explain how and why various systems arise in various times throughout history. Hegelian thought had completely taken control of Europe, and much of Kierkegaard's writing was rebelling against it. So Kierkegaard talks a lot about the deficiencies of this kind of naive systematizing.
As someone who is so excited about prediction markets and, by extension, the wisdom of crowds, I am a little surprised at the degree to which you dismiss the "social epistemology" dimension. I'm not saying that you can't attempt to address the problem with your individual brain, but surely for most issues it is preferable (and much more practical) to rely on social epistemology, considering how many assumptions we are faced with every day. Is the Earth round? Will my car explode while I am driving down the freeway? Do vaccines contain microchips? I don't have the expertise, nor the time to establish the precise level of consideration to give all of the these questions.
"Some people try to dodge the question and say that all rationality is basically a social process. Maybe on my own, I will naturally stop at whatever level seems self-evident to me. Then other people might challenge me, and I can reassess. But I hate this answer. It seems to be preemptively giving up and hoping other people are less lazy than you are.”
I don’t think this is about being lazy. I think it’s about inferential range. Slack in inferential range has a clear tradeoff, which you well describe in Studies on Slack.
https://slatestarcodex.com/2020/05/12/studies-on-slack/
I think this really is a social process where the group benefits from different individuals having different amounts of slack.
The easiest absurdity dismissal is that Neom is presented as this giant thing all at once. No city (maybe Brasilia?) ever just got planned and millions of people moved there. Now throw in a totally untested architectural model for this city, this style of living. It would not be absurd to build one section, in the most desirable location (near the sea) and see if it works, see if people like it. Then build another section and so on. It may not get to the length of Ireland. They may have learnings as they go and as residents give input. Newer sections may be designed differently. But if the economics or the functioning of this thing only works as a giant whole, it seems doomed.
The goal here is probably to minimize instances where we lock in a wrong position and refuse to update further without wasting too much time on analysis, I would guess. This is the position I would try to adhere to, at least, because like other humans I can't optimize for only the truth without serious loss to everything else I consider worthwhile.
I think the trick for things outside my fields of expertise is splitting up the heuristics, rather than serious study:
-Is it in my field of expertise, or is it something that looks like it is but is adjacent? I see a lot of experts make this mistake, where they comment on something in a related field, and end up wildly off base because they mistook it for their own field.
-Do experts in the field say its impossible? Being fringe is one thing, being impossible is another. Not infallible, but good enough particularly for hard sciences. I'm not going to necessarily understand *why* something is impossible without a background in the field, so it's impractical to learn it myself.
-If the claim appears particularly absurd: Do I understand the claim correctly? Some claims are as insane as they appear, but this is worth a double check.
I think all of that has to happen past the initial gutcheck of absurdity, though, so it seems more like a series of backup heuristics. I'd also be concerned about the opposite; blandly accepting a statement because it seems reasonable. I have a feeling the "reasonableness heuristic" is the more deadly of the two, but I have no evidence to back that up.
My point is that what you wrote didn't pass muster as something a NEOM proponent would read and say "yeah, that's a good point. I still think we will succeed, but if we fail, it will be because of these things Scott highlighted".
In other words, my concern was strawmanning.
I was even able to play "rubber duck NEOM proponent" for you, and show you that the argument about the trains needing multiple stations was wrong, by using the example of the boring company: https://twitter.com/alexandrosM/status/1554888465870794753?s=20&t=XQ8iGM2nkdh6cWeF2D_s0Q
That example didn't have to exist for the argument to be wrong, but it was convenient that it did exist.
Another IMO not-very-good argument was to call the project impossible. Jumping from "this is very hard, I don't think the saudis can do it" to "this is impossible". Now, someone might say that this is a nitpick.
And yet, I think the difference between "very very hard" and "impossible" is everything we've ever done as a species that has been worth a damn. A sensemaking error that causes one to confuse the two is a sensemaking error that makes all the difference between a thriving civilization and one that stagnates.
I believe impossibility claims should be handled with care, ideally built on things about as certain as the laws of physics, because they're powerful tools of epistemic warfare
"But how do you, pondering a question on your own, know when to stop because a line of argument strikes you as absurd, vs. to stick around and gather more facts and see whether your first impressions were accurate?"
Here's the thing: you can just say "looks super hard, not convinced they can do it". And if pressed to make a decision whether to buy discount NEOM property, you can just say no. You don't need to jump to claiming impossibility to do that.
I even described a path towards constructing a stronger argument, by focusing on the implementation path rather than the endpoint.
And so, in a way, this piece makes the same mistake as the one on NEOM. You go in depth about the absurdity heuristic, and the result reads to me like you're not actually engaging with the point at hand: using weak arguments to support a position, regardless of the validity of the position itself. Had you just said "seems pretty hard to me, I'd bet against it" I would have been in complete agreement.
It really feels like much of this would be resolved much better with a conversation.
Speaking of which, I'd be pretty happy if you read my analysis of your ivermectin piece, ideally letting me know where I'm misunderstanding you and/or correcting the elements you agree are in error. I'm learning a lot while doing it and I still think there's a lot of good to be done by engaging in open-ended discourse about the topic at hand.
https://doyourownresearch.substack.com/?sort=search&search=potemkin
Does anyone know the post where Scott argues that the multiverse is a good scientific theory?
Back in 1969, I saw Paolo Soleri's exhibit at the Museum of Modern Art in NYC showing plans and models for what he called arcologies. These were immense buildings, as tall as skyscrapers and running across the land for miles and miles. They were an an update of Corbusier's Radiant City or Buckminster Fuller's urban structures with an attempt to think of a city in ecological terms, as forming its own world and lessening its impact on the natural world. I think the name came from architecture and ecology. The idea was that such a linear city could provide housing, controlled environment agriculture and recreation, a transportation system and all the infrastructure needed for urban life with a minimal ecological footprint.
There were a lot of big thinkers in the 1960s and a lot of fascinating new ideas, many of which we take for granted today. Arcologies are not one of them, though I noticed a reference in Neuromancer. Neom, for all its grandiosity, offers the same appeal. Saudi Arabia is not noted for its pleasant climate, abundant fresh water supplies, or agricultural production. An arcology could offer better climate control, more efficient water use and possibly a platform for urban farming. Neom is not absurd or inconceivable by that standard.
I think the main flaw is that the design does not seem to provide a pathway for organic urban growth. What is the minimum unit that could be built as a useful prototype, something more than Habitat '67? How would it work as a test bed? How would it fit with the nation's traditional urban life? How would it deal with the large number of foreign workers who are major part of Saudi Arabia's workforce? Then there's the matter of the variability of oil income, but I'm not a Saudi staff economist, so that isn't my problem.
It helps to remember that absurd comes from the mathematical term surd or irrational number. These were a real problem in mathematics for a long time, but nowadays we work with surds all the time and barely notice their absurdity.
(*) Soleri's work informed a lot of the model design for Star Wars. Compare some of his arcologies with some Star Wars battle cruisers and the like.
*Epistemic status: Stating an opinion*
I think Alexandros Marinos is actually right in the case of the Neom post. I believe it's usually most productive to approach a disagreement as such:
1. State the opposing side's view, as far up in the Graham hierarchy [1] as possible
2. State your point of view, as well as possible
3. Rebut your opponent
4. Optionally make fun of your opponent
As far as I can see, 1 was done minimally, 2. was part of the post, 3. was only minimally present (absurdity heuristic), but it was *heavily* entangled with 4 (which is probably the worst kind of entaglement possible in a disagreement: Even singling out the the sentences "I disagree with this because of the absurdity heuristic. I will now proceed to make fun of Neom." would have been more productive.
Unfortunately, it seems like entangling 3 & 4 is psychologically easy to fall into (I've too often fallen into that trap, and imagine that'd continue if I garnered more status), and gives you the most clicks, so avoiding it is *hard*. I've seen some comments suggest that the ACX posts lean more into this than the SSC posts did, this feels true to me, and I want to *gently* point to CONSTANT VIGILANCE [2].
[1]: https://external-preview.redd.it/Jk8lfGW2JnNpYGWqRRwRLEiPJyXXPOdwKlW8LYhL7ag.jpg?auto=webp&s=2b817ec0fbbc58bdb49eb429797d3ace40111462
[2]: https://slatestarcodex.com/2014/06/09/constant-vigilance/
That's right. Corbusier, not Fuller. I haven't really been following that stuff since maybe 1975. It's no wonder I'm getting things mixed up.
That kind of big systems thinking got a lot less popular in the 1980s. It's interesting that it has been making something of a comeback since then.
I hadn't heard of Neom before this post, I agree it sounds absurd.
But I think there may also be some absurd construction goals in the west.
I've been reading about green energy, and I get a similar feel for some of the zero carbon goals.
I made an estimate of how large a national wind farm would be, the size struck me as absurd. I don't think we'll build that much wind power.
I think we could build enough solar power, but the batteries to store it will be a huge challenge:
https://medium.com/@tgof137/heres-what-it-would-cost-for-the-united-states-to-go-100-solar-1b7c34e6fcef
I made an estimate of how much lead we'd need for a national lead-acid battery. It's more lead than is known to exist in the world.
I made an estimate for how much a national lithium ion battery would cost. It's over 100 trillion dollars, and it would take a large fraction of known lithium reserves.
I made an estimate of how much pumped hydro storage we'd need, instead of batteries. It's absurdly large, the national reservoir size would be equal to draining Lake Tahoe just to store 3 days of power for the country.
So far, the most common response has been:
"You're just knocking down strawman arguments, of course we wouldn't solve the problem with 1 thing".
When my point has been more like:
"Solving green energy and intermittency looks really hard and every solution looks expensive. We need new solutions, or maybe we need nuclear"
"their king is a megalomaniac"
MBS is the crown prince, not the king.
He is for most intents and purposes Saudi Arabia's dictator, though.
My favourite absurdity is how my job involves making a special type of rock do my bidding by feeding it patterns of lightning. Lots of my time and effort is spent making sure that the lightning is just so, otherwise the rock misbehaves.
Quite a few people are preoccupied by worrying that the machines with these rocks at their heart might one day go out of our control and kill us all.
I think that the virtuous rationalist path is most likely to do a large amount of research on Neom down to the construction physics level, then write a Much More Than You Wanted To Know post, then get feedback from others, add corrections as needed, and then rinse and repeat because you can never be sure that you didn't miss something.
The problem with that is that we are human, we don't have the time, energy, nor inclination to do deep dives, and reality has a surprising amount of detail (http://johnsalvatier.org/blog/2017/reality-has-a-surprising-amount-of-detail). So, we rely on heuristics based on our priors. This includes the absurdity heuristic, even though that heuristic's poor track record is a major theme of the rationalist canon (ie, how heliocentrism, evolution, special relativity, quantum mechanics, etc. seem so absurd).
I think the idea is that there is usually never a royal road or even a muddy goat path, but there is always a compass. That compass always points in the same direction, labeled "Do more research, and listen to more criticism and feedback to make sure you are not going astray". When looking at the direction the compass is going, we see difficult terrain, lots of skulls (https://slatestarcodex.com/2017/04/07/yes-we-have-noticed-the-skulls/), terrible weather, and it goes uphill both ways.
I don't actually think we should be on the most virtuous rationalist path all of the time. After all, we have a limited budget of time, energy, money, etc. However, even following it a little bit of the way is probably good.
For instance, instead of "Neom is so absurd! Ha ha ha!", it would be better to say "I think Neom is absurd because of A, B, C inefficiencies, which if we look at ordinary construction costs would cost $X, and make it hard for residents to get around because of Y. Even if the Saudis have super-secret extra-efficient construction technology (unlikely because of Z), constructing a more ordinarily-shaped city with walkability and transit in mind would still cost Alpha % less and increase mobility by a factor of Beta for reasons Gamma." At least for me, sometimes the absurdity becomes more apparent when the critic lays out the facts than when the critic points and laughs.
As for how far we should follow the compass of Rationalist Virtue given our time/energy/money/inclination constraints, maybe here there is really not only no royal road nor goat path, but also no compass. I suppose it depends on things whether we are trying to persuade or get to the bottom of things, how important is it personally to know the truth (ie, getting Neom wrong costs nothing to anyone here in the West, at least for now), and how certain we already are (this last one can still go wrong, since closed-minded bigots are very certain of their beliefs).
The point is that we should always see absurdity as a cop-out. Maybe we can agree that it is the best we can do for now given a large multitude of constraints, but we should never accept it as a good argument in and of itself. If we stop at a certain region when making a Map of Knowledge, we should never mark that region as "Here Be Absurdity". Rather, we should mark it as "Here we gave up", or maybe include a pointer to a cache of arguments someone else made. In the latter case, we would let the reader/future mapmakers decide if the pointer points in the correct direction.
Also, I specifically disagree with "in order to dismiss it as absurd, I need to explain why the Saudi government would waste $500 billion on an obviously absurd idea. This is easy: their king is a megalomaniac, plus people are afraid to voice dissent."
Yes, MBS really is a megalomaniac, and yes people really are afraid to voice dissent, and yes, all of those are good reasons why the Saudi government would waste such a sum on an absurd idea. However, these are not by themselves reasons to dismiss Neom, and Scott himself noticed the red flags of Bulverism and bias arguments. To properly dismiss Neom, it would be better to discuss urban planning, economics, and construction physics.
Neom seems more like a concept with some strong statements, but we've seen similar grand scale plans fail before or simply never come to fruition. Does anyone remember those artificial land palm trees? They're...not doing so well and were never finished. So until ground is broken or a highly detailed multi thousand page report is issued, then we can dismiss any concept car or concept arcology building out of hand based on priors that such things never come to market or look very different if they do.
It seems like we might match our level of effort in dealing with a specific claim to some dgree with the amount of specificity being offered.
Do they have a detailed list of supplies, purchase prices, and hours of effort by various categories of builders to construct their price estimate with comparison rates of real projects? Or is it just a number pulled out of thin air? If there are some 'high level' graphs then that's pretty low specificity and lacks data to support it from prior building costs in the area, as Scott mentioned.
Now you don't want to get sucked into a rabbit hole by some crazy person who writes endlessly long lists of things and infer that somehow makes what they're saying any more or less likely.
But it is easy enough to make some guess as to how seriously you choose to take any pile of estimates and if it is too complex, to rely on experts to analyse such a complex argument. If you see a brochure with 30 pages of pretty pictures and pie charts....you can push back to say it is absurd relying on your superficial priors...but if you have a 3,000 page plan then a bit more effort will be required to analyse it for overly rosy or pessimistic projections. This is how political think tanks generally work to promote highly biased and often false ideas using detailed reports to support the foregone conclusions their sponsors wanted to promote.
Given the relative levels of technology, is Neom more or less absurd than the Great Wall of China?
I'm going to offer an anecdote that might be slightly relevant, but I sincerely hope is not. Mostly I just want to tell this story....
Sometime around 2007-2008, a friend of mine was hired as a contract welder and fabricator in (I think) UAE. He was working on the enormous stage for the wedding of the Princess. Six months. Enormous.
He had several trenchant observations about the culture there, but the semi-relevant one in this case is about "labor conditions". 'Berto was working steadily, but saw there was no way all this work could get done, so he asked up the chain for some help.
A few days later, he got some help. A hundred (100) Sri Lankan laborers showed up, and were put at his service. These guys had only the clothes on their backs, and no skills at all. They were put to work sometimes pounding in screws with hammers. They worked frantically hard. 'Berto was deeply troubled by the implications of this. After six days of this, there was the collective day off. Then on [Monday? after the day off anyway], a hundred COMPLETELY DIFFERENT Sri Lankans took their place. This went on for almost two months.
My friend prayed the whole time that each hundred laborers were not just fed to the sharks after six days.
The fact that he considered that might be possible, well it might have a bearing on construction economics in the Arab World.
"Marginal spending on medicine has zero net effect" is only absurd if you don't know what "marginal" or "net" means, right? It means "we're spending roughly the correct amount of money on medicine" (whereas marginal spending having negative net effect would mean we're spending too much, and positive that we should spend more). Or am I missing something?
I think the main thing this discussion tells us is that there are many ways to analyze a project, and people are EXTREMELY convinced that their "rational" way is the single canonical rational way...
- One dimension of this is the Saudi government and its various out-group properties. The all-consuming issue to about 90% of the commenters, regardless of specifics about the project.
- A second dimension is the cost. The all-consuming issue to about 9% of the commenters.
- What very people appear to be interested in is the city qua city aspect of this; yet to me that's the part that obvious renders it "absurd" regardless of bias or otherwise.
Consider the pictures as drawn. Not how you you want save the project by redefining, the project as described.
What you are living in is a TUNNEL 170km long. You are living in a skyscraper/shopping mall that you can never leave. Is that what people want, especially the kind of people who rhapsodize about this sort of city?
How are things delivered in this city if there are no cars? How are buildings repaired if there are no open spaces (eg roads) that act as emergency locations for cranes and building equipment? How is air pumped into the city and heat removed? (All this stuff can be done for skyscrapers because they have huge open spaces on the sides where the equipment lives and can vent, but according the video two of the sides of each "skyscraper unit" don't exist, and the other two are mirrored glass with no ugly chillers or pumps.
What are the fire escape plans?
Do animals and birds live in this space along with the trees? How does that play out in terms of animal waste and falling leaves?
etc etc etc
This stuff is considered absurd because there are no intermediate proofs of concept. Yes, you can grow one tree in a skyscraper, but no-one is growing forests in skyscrapers. Every real-world image I see of this sort of thing ACTUALLY has the trees on the outside of the building, so the problem is moot.
It's not quite the same, but the logic here is somewhat like "we know how to build gardens, therefor how hard can it be to build Biosphere 2?" And we know how that turned out.
Now if you want to say that the images in the video are all nonsense, well sure. Then we can have a different sort of argument. But the argument at least some of us (1%) are having is with the concept as presented by the video. And we call absurdity because of these eminently practical things: air, heat, life forms, repair, logistics, building evolution, ...