Traditional multi-page apps/sites are inherently better suited to serving content, especially textual content. SPAs only ever make sense for true "apps", which a blog is not.
Yes. And some quantity of designers and implementers have overgeneralized from "this is useful on high interactivity sites" to "this is wonderful; we should do it everywhere". Also, the frameworks they use, not wishing to reinvent the wheel, probably only offer one way to do things now - that being SPA. Because they don't want to maintain 2 ways of doing things either.
The other "benefit" of SPA, as well as the slow load times, is the execrable performance on older computers. If they work at all, that is; they have a bad habit of being dependent on having the latest and greatest version of the browser - or worse yet, require something that not all up to date browsers support at all.
One other thing to consider is that, from the corporate dev's point of view, the site they're working on could become high-interactivity at any time. Being "agile" means being ready for new requirements, so it's easy to see why teams might be tempted to just go SPA from the start, even if the product is, at least for now, just a blog.
I agree mobile is key, and I have the opposite set of preferences. I like reading more on desktop and I preferred the old layout.
I am increasingly frustrated by the "mobile first" design paradigm taking over the web. Because it often means "mobile first ... desktop can just use mobile too". The desktop version of substack is not different at all from the mobile version.
I almost only read on mobile and I don't really care about the design change. Both SSC and ACX work and look fine on mobile (as least as of 2016). What does annoy me is the impermanence of the comments on Substack. If I leave the page open on the comments and come back the next day it takes forever to load and it almost never jumps to the place where I left off. This was not a problem with WordPress.
Yes this is my **major** complaint about substack vs WordPress. The load time is longer and that small friction shows up a lot. Impermanent collapsed/ expanded comments is another issue but the bigger one for me in comment load time.
We need some kind of intermediate server that just generates the HTML of the comment section (or a subthread) in the background and people can load that pre-rendered cached version easily.
Does Heroku-node let you do that? I've basically already written from scratch all the Javascript [1] to create a comment page from the Substack JSON so I just need a place to deploy it.
*EDIT* It could technically be pre-built on your local machine in the background. I'm concerned about burning a lot of cycles locally that are never used because that's part of what the Substack UI does badly but I could come up with a decent compromise.
I can see the results of just 2.5 games at a time, but the single column, lots of space, large logos/characters probably makes it easy to scroll on mobile.
Look at the old version. I can see 8 games at once in 2 columns:
So, forgetting for a minute that I wasn't going to be able to see your point because im on my phone, I clicked through to the current MLB site and thought "that's odd, I can see 7 games without even scrolling down." Then I remembered and turned on Desktop View. "Odd, now I can see even more games." This was followed by another forehead slap and me realizing to turn my phone from portrait to landscape. "There it is!" It really is terrible...
Hearing this from someone other than myself feels validating -- although I would phrase it less as "increasingly frustrated" and more as "psychotic rage that can only be ameliorated by taking the last five years of web designers, loading them on a spaceship, and lauching it into the sun." Incoming rant warning.
It's not just the SSC->Substack thing. Web design in general has gotten actively hostile to readability, and I blame a lot of that on mobile. Part of it is the form-factor difference. On a PC, vertical space is priceless and horizontal is cheap -- yet narrow designs with huge fonts, wide line spacing, and extra-tall images have become so prevalent that I feel like I need to override the CSS on every other site I touch. And there are sticky headers everywhere. Whoever invented that abortion of a misfeature is in dire need of a thermite suppository. *There are miles of empty space along the side of the page, use *that* instead of trying to forcibly occupy ~10% of my actual usable space.* Jackass.
All of those -- and there's a much longer list than that -- are things that only make even *bad* sense on mobile, with its flipped aspect ratio. Mobile-only is the modern Web's version of "this site designed for Internet Explorer". Everyone who matters is on a phone; everyone else can go take a long walk off a short pier.
The second problem is that everything wants to be an app, including Substack. So you get sites reinventing, overriding, or otherwise disrupting standard browser features with javascript garbage, constructed in such a way that the now-in-name-only 'website' breaks immediately if you turn it off. Substack *crawls* by comparison to SSC. Thankfully, it only partially breaks in the absence of JS -- comments disappear -- but like many sites, it makes a point of nagging you to turn the shithose back on. No. I turned javascript off for a reason. Whining about it is strong evidence that *you are the reason*. If I trusted you not to defecate on my experience the moment I took your suggestion, it wouldn't be off.
....so yeah. It's not that SSC was a paragon of good design, and it's not about the minimalist look; the look-and-feel isn't the issue. I preferred SSC because Substack is a web application emulating a phone app emulating a browser for the bog-simple, 1990s task of *displaying a post and any responses to it.*
Things like NoScript, uBlock Origin, Stylus, and uMatrix are increasingly non-optional. Modern web design is an act of war on the user. If one would still participate, the browser is a trench that requires defending.
Yes, all true. My speculation is that if we compare Web programming to Detroit we are in the 1970s chrome & tailfins era -- everyone competes to deliver some new random funky feature yearly -- New in the '74 Olds! Power doorlocks! -- and assorted chromed flash, and we are awaiting the equivalent of Toyota, the economical option that just runs and runs, to arrive on the scene and cause a great reset.
It's strange that blog comments, a thing that was perfected in the 90s by CmdTaco fresh out of college, to run on computers with 10^-5 the power of modern machines, are still clunky and occasionally broken here despite the enormous wages I am sure Substack pays its JS programmers. In their defense, it wouldn't surprise me if they aren't given much liberty to focus on plain reliability and speed, because someone in the front office wants them to drop everything and add some chrome and a tailfin or two.
> blog comments, a thing that was perfected in the 90s by CmdTaco fresh out of college, to run on computers with 10^-5 the power of modern machines, are still clunky and occasionally broken here despite the enormous wages I am sure Substack pays its JS programmers
Those enormous wages are part of the problem, because they contribute to the alienation of the software-developing class from the technological proletariat.
Most of my work involves connecting to various servers remotely, and the speed of the box that my keyboard and screen are physically connected to doesn't matter at all, so a six-year-old laptop that wasn't top the line even when it was new serves me just fine.
But some websites take more than a minute of clunky loading animations to show up on that machine, and I used to wonder how anyone could put up with that. Until I saw a colleague open it on her beefy workstation and it rendered basically instantly.
In the screenshot, Vladimir compared using Substack to wading through water. Now imagine Substack employees having to wade through that same water to get anything done. They wouldn't be able to move fast enough, but still break things. And if highly-paid employees are slowed down, their enormous wages are wasted money for the company. So they get speedboats, and barely even notice that the streets are flooded and children are drowning.
Anyways, I don't care much about the layout, because I just read the RSS feed, which is its own kind of alienation, I guess.
I'm not convinced the difference in processing power has that much to do with it. Something, maybe, but "don't actively break the user's UI" is not a concept that should have much to do with site performance.
Motherfucking Website and Better Motherfucking Website should be required reading for all web designers.
I've experienced too many times where lack of performance becomes no performance. The UI was written with the implicit (and incorrect) expectation that an operation always completes in a certain amount of time -- which it does, on developer boxes!
Run on a normal machine, those assumptions stop working, a call fails somewhere and the error condition breaks the app until it's restarted.
I love this comment. Is there some kind of Toyota cult we can join to help summon the Eternal One? I will pull or not pull Andon Cords all day if it will get us closer to clean website design.
> And there are sticky headers everywhere. Whoever invented that abortion of a misfeature is in dire need of a thermite suppository.
It gets much worse. A lot of sites use a header that vanishes when you scroll down and comes back when you scroll up! This is an abomination; it makes it nearly impossible to position particular text at the top of the viewing window, which is what I want to do.
Ugh, yes. The original perpetrator of *that* offensive pessimization should be forced to play a version of transparent Newcombs' where both boxes contain tigers.
Thankfully I don't see that on Substack, but it's possible my array of browser armor is stopping it. When that fails, I have a bookmarklet set up to nuke sticky headers manually. I would link it, but I can't figure out how to do links in ACX comments, and Google is not enlightening me.
I just did a little experiment on google-chrome: if you edit the chrome binary to change the last two occurrences of the words "sticky" and "fixed" to "stciky" and "fxied", the reappearing and floating toolbars on google, substack, and a few others that I've tried so far, all disappear, while still appearing at their proper positions at the top of the actual page.
I'm sure this will break things somewhere, but I'm not sure I care enough; or rather, I prefer it enough that I'm quite happy to maintain a second separate installation and profile for such things.
I work in web development and perhaps I would find it more rage inducing if my job didn't require me to work with the people you want to launch into the Sun.
I can rant for days about all the shitty ideas web designers come up with. I had to pick my battles though, and substack does at least avoid the one thing I will fight to the death about: modals (or website sanctioned popups for those who don't know the term).
For some reason designers love modals on desktop. But because things had to work on mobile, that modal just becomes a full screen modal. Which makes it look no different than navigating to another page. Sometimes designers even want full screen modals on desktop (WTF!).
Modals of course break all web navigation expectations, because if you click the 'back' button on the browser you aren't going 'back' from the modal, you are going 'back' from the base page. I tell them this, and EVERY TIME they say 'oh well it will be fine they can just click the X to close the modal'. And EVERY TIME once it gets into the hands of users they use the back button like they expect to on nearly all web pages, and then a bug gets opened to "fix the back button on this page".
So then I have to write some shitty javascript to subvert and break basic web navigation, so that when this stupid modal is open and a user clicks the back button it doesn't go back in the browser it just closes the dumb modal.
To top it all off I'm pretty sure that the modal paradigm is so popular because its the only way for designers to actually design a clean page from scratch rather than dealing with the accumulated crap from sidebars, headers, and notifications.
Modals are one of the longer-list items that didn't make it into the rant. They go in the same bucket with sticky headers as unforgivable atrocities. I nuke them on sight. And javascript subverting web navigation is one of the reasons I keep it off by default.
Both are functionally attempts at a hostile takeover of my browser. I think that might be the root conflict in the war -- it's a war for control of my interface in the same sense that modern advertising is a war for control of my attention. Individual actors can improve the situation slightly by providing an API (which substack, I note, does not), but that is of limited use when the API is provider-specific. It potentially allows others to write a non-terrible client, but few are willing to do that if it will only work against a single site.
I blame Apple for modals. I seem to recall they were a (then strange) feature of Apple's early OS UI, different from everybody else, and they were freaking proud of it. I guess the logic was that these ephemeral messages could be displayed, dealt with, and dismissed without messing up the Zen garden elegance of their windowing system. Or maybe they just loved popping open little windows because they could. Who knows? My first exposure to a windowing OS was X, and X didn't encourage these little mushroom-like monstrosities -- that's why you kept a console window open, to collect in one place anything the kernel had to say to you outside of an application. So they still annoy me. Having them within an application (or, worse, website) is even dumber. It tells me you can't be bothered to design a UI which allows the user some routine and intuitive way of knowing your (the application's) status at a glance. . It's like traffic engineers designing stop lights that spring out of the street from randomly placed manholes right into the middle of traffic, whenever conditions suggest some kind of traffic flow control is advisable.
At least Substack isn't at the point of https://how-i-experience-web-today.com/ (yet?), but SSC > ACX on Substack because of web bloat. Trying to load ACX on a weak LTE or Wi-Fi connection is awful. Less "modern web" (/JS) bloat on SSC makes the experience much faster, and much better.
The problem isn't what SSC does better. It's what Substack (and most of the "modern web") is doing worse.
That makes a lot of sense. I prefer to respond as well as read, and I never voluntarily produce text on mobile, unless it's as short as a text message or tweet, e.g. "running late, sorry".
So anything that improves mobile at a cost of dis-improving desktop is pure dis-improvement to me.
I find that the substack page lags and glitches terribly when reading the long open threads on mobile. Which is a very difficult thing to do with rendering plain text on a modern device, and implies they are doing something insanely inefficient.
I have the same problem. Long comment threads are basically unreadable and I therefore ignore the open threads, despite previously being a frequent open thread user on the old SSC.
There's so much javascript junk on substack (or any site designed after 2010) that I had to disable javascript to be able to read it on my (very old) phone. I didn't install the app because $£#% apps.
The fact that I can even see some lag on a text+img website on my gaming PC tells me that buying a more expensive phone will not solve the problem.
I can't really add much to this, but it gives me an opportunity to link to one of James Mickens' old USENIX posts, so that is a positive contribution to the world.
Read it through Feedly, or on an iOS device via Safari reader. I have no problems with it on mobile.
The big problem w/ substack is that it provides lousy mechanisms for completion it’s, or anyone who, for whatever reason, wants to read stuff within a particular date range. A semi-endless scrolling feed, anchored at “now”, and with no quick way to jump large distances or to a particular date, is not an “Archive” regardless of what substack marketing thinks!
I have experienced this myself over 15 years of blogging. And honestly - people always just like the band's old shit, the old sound. It's not more complicated than that.
I suspect that's right for the cosmetic aspects. But there are functional differences too, each of which seems small, but all of which are relevant. (Links to posts, links to comments, browsable archive, fully-loaded comments rather than appear-as-you-scroll, etc.)
None of those things seem small. Links to comments especially are important.
A related issue is that when you have a page open for a long time (like to read through lots of long comments) and then page reloads or the browser closes and opens again, it loses your place because the comments are not all expanded.
I guess the problem is that it’s a blogging platform that is designed as though it were a social media platform. For social media, content is treated as largely ephemeral, so that people aren’t looking for a full chronological history of it, but it is instead delivered by algorithm.
Right, I think that's a big part of the problem: Substack's design is modeled on sites like Facebook whose objective is to be maximally addictive, not for thoughts of enduring intellectual value, nor to make it easy to navigate complex information spaces. Systems that do a better job of these (IMHO more worthwhile) objectives include IDEs, paper books, Wikipedia, Jupyter, legal briefs (just check out the table of contents of any Supreme Court amicus brief!), and Excel.
The archive page still sucks. Should give me columns, so I can just fullscreen on 4k and get max information density.
Not that I have ever seen that anywhere ever, but it's just how you'd organize a list with hundreds to thousands of entires, if you actually care about helping people find things as fast as possible.
Interesting! For those who aren't going to click the link: eBay users revolted when the company changed the background of some pages from yellow (which, seriously?) to white. So eBay changed the color back to yellow, and then slowly faded the yellow over a period of months. This time, no one complained.
Yeah, I feel WordPress nostalgia. Reminds me of when the MR-centered blogosphere was a lively, rude Conversation, when Robin Hanson linked to PUA blogs, when Mencius Moldbug still wore a cape.
Yeah, I think it's mostly this, along with response bias on the survey. Everyone who has ever redesigned a website knows that touching a single pixel is going to elicit a howling response from a very vocal minority. And that minority is going to be extremely eager to register their unhappiness in a survey.
Personally, I think the Substack site looks better. But my main take is...shrug.
Sure, I wouldn't expect 100% of people to prefer the new site if going in blind. The truth is, both designs are pretty vanilla. How strong is your preference?
Fwiw, the old site pretty clearly violates a bunch of basic typesetting and usability principles. The contrast between font and background is often too low. The use of white space and line spacing is inconsistent. It mixes serif and san-serif fonts in a way that I find a bit jarring, although clearly that is delving into the territory of personal preference. The column seems too wide for ideal readability. More than anything else, it's really dated looking, with the embossed font on the blog roll, etc.
I'm not suggesting this proves the new site is better, but Scott is certainly right when he says that the old one is slapdash and amateurish and the new one has benefited from some basic professional attention to detail. Ultimately, they are both columns of text on a page, the differences just aren't all that great, and it is very common for people to reflexively oppose changes in web design for reasons that are somewhat mysterious to me.
Eh, that's a bit simplistic. I agree, design trends are often driven by the boredom of designers seeking novelty, with little or no regard for the value to the consumers of design. On the other hand, web standards and capabilities have evolved radically over the past twenty years and continue to do so. Devices used to access the internet have evolved radically. Also, users themselves have evolved as they become accustomed to different technologies and UI affordances. Whatever its charms, Geocities doesn't cut it today.
Also, aesthetics do matter on the margin. The embossed fonts on the old SSC blog roll surely aren't really hurting anyone, but they are dated and kind of ugly and changing them also doesn't hurt anyone.
So, sure, change for the sake of change alone shouldn't be done heedlessly or on a whim, but the entire concept of something being outdated is hardly invalid.
I think the original claim is that it doesn't make sense to criticize something as aesthetically "outdated" if it functions as well as something newer. Do you have a better response to "what's wrong with the fonts being dated" than "but they *are* dated"?
>"[on the old site]... The contrast between font and background is often too low."
Comparing the screenshots of the old versions and the current version as displayed, I think the opposite, for the main article text.
I found a word which is present in this article, and the first screenshot ("thought"), made cropped bitmaps and compared them zoomed in. The bars on the 't's and the top of the 'g' are blurry in the substack article font, while all strokes are evenly dark on the SSC article font.
This is the case if you compare other letter pairs individually - on ASX the lower case 'e' is especially bad, the central line being almost invisible; it's practically a 'c'!
Overall this makes the text look a bit faint, even though it's larger and darker in places.
To be clear, I can read both without too much trouble, but the old site text is definitely easier to read.
The main article text is black on white in both cases, which is the highest possible contrast.
I was talking about the blue text on gray in the blog roll, the white text on light blue in the header, and the gray text on gray in the byline. I haven't run them through a contrast checker, but I have no doubt they will fail standard accessibility guidelines, probably by a lot.
I'm not sure exactly what you are looking at, but you seem to be zooming in on some of the font anti-aliasing. This isn't really a valid thing to do -- most fonts are vectors that are designed to scale up and down infinitely. Converting them to bitmaps destroys this ability, and so of course they will look terrible when blown up.
The readability of a body of text is not determined solely by contrast between background and a solid block of the text colour. My point is that the font affects this, and therefore the font used for body text is very important to the overall impression you get of a primarily text-based website.
I don't particularly care about the header; in everyday use its legibility is practically irrelevant. But since you bring it up I think both the SSC headers are perfectly readable, and more recognisable as others have said.
The reason I was enlarging images to look at the pixels was to study why the font substack uses is less legible at the size it's displayed at on a monitor. I think this is entirely valid, it reveals that some of the strokes are anti-aliased into a light grey smear. The fact that a font could be rendered nicely at a much higher resolution is irrelevant - what matters is what is used when you read the article.
I thought I'd make a demonstration image, which I've put on a free image hosting website: https://ibb.co/GQTcpfg
If you look at this, you can see that for the ACX font, while there are some dark areas, many parts of the letters are much lighter. The entire central stroke of the 'e' is two pixels tall, both a rather light grey. The lighter greys are where it's effectively trying to draw a very thin line, so presumably there the glyph has a line precisely placed along the edge of two rows. Rather unfortunate. At a different point size, that may not be an issue, but perhaps some other stroke would be as bad.
Presumably the SSC font was designed to avoid this problem at this size.
I guess I should say that I couldn't rule out that Scott's computer was just using a better rendering algorithm, but I don't think it's that - the comment text here looks fine, so it's just that font.
I feel like a really big part of it is just the colors. I like the pleasing dark blue banner and darker gray sides of the old website. Now everything is too pale.
It is certainly nice to anchor to a color. I still internally map SlateStarCodex to that particular shade of blue, and Unsong to yellow. It feels pleasant and familiar to see the color where it should be.
I think what I'm feeling is lack of customization and identity -- Gwern's site still feels like its own thing, even though it's black and white.
Agreed on this. As nice as substack is in a lot of ways (like being readable, having good RSS feeds, etc), it feels sterile. Especially because most newsletters default to the black-on-white or the white-on-black. The exception being Free Black Thought that uses white-on-brown.
Edges and frames, yes. It's more comfortable to walk through Paris, with its multi-story buildings crowding the street than it is to walk on the sidewalk next to a US strip-mall with an acre of parking lot between you and the nearest building.
Strange juxtaposition. Anyway I'm more disheartened by the acres of vacant compacted dirt with odd sprouts of brittle yellow grass beyond a warped and rusty chain link fence across the street from the strip mall than I am by the parking lot itself, shattered and pale though it may be. I've been in LA enough, it's a sprawling nightmare of strip after strip of urban decay gradually dissolving in the possession of a passive, lazy, incompetent population. But yeah a prefer a nice cozy log cabin over the LA version of freedom. I do love my freedoms though.
I did a test and printed this post to a pdf, and viewed it in two page view. Was much nicer to read, since I had more info available in single glance. Nicer to skim.
Their design isn't worse than your old layout. If they pushed SSC-style layout to every website, everyone would hate it.
It's like the Hugo and Jekyll custom academic themes everyone is capitulating to (yay capitulation!). The first time you see them, they look nice. Once you see them for the 10,000th time, they look incredibly boring and tiresome and you wish people were still designing their own websites.
If I ran a salad restaurant, I probably would get a lot of salad fans. If I switched to hamburgers, a lot of my fans would be angry and like the restaurant less. This isn't because of nostalgia, it's just because a salad restaurant probably attracts lots of salad-lovers.
I don't know if relevant as quite subjective, but I would usually not comment/vote and prefer substack to the old website. Maybe silent majority likes it more?
That said I do mainly read posts through the email newsletter so...
Also I don't really read anything else on substack so no way for me to be sick of it
Design-wise I think it's not much different, but on the old site I was never able to comment for some reason. I tried a few times but the comments just never appeared and at some point I gave up trying to figure out why. The Substack reset let me express my views for the first time.
There are a number of people, myself included, who believe that the user created, original and homegrown, flashy and weird and geocities-esque websites of the web 1.0 era were strictly better than the modern flat and minimalist, muted colors, tons of javascript, everything-looks-the-same websites of the web 2.0 era. I'll take https://squeedge.neocities.org/ over other websites any day of the week.
Also, in terms of functionality, the thread system of SSC was objectively better than here. It notified you about new posts in a easy and comprehensive manor, you could jump straight to them, and there were a few other nice QOL things.
Agreed - I think "good design" is actually worse than "bad design". It's like a house that's too clean - we usually think "clean" is better than "dirty", but if a house is too clean, it feels offputtingly sterile, and all the things you want to *use* are hidden away in cupboards and closets and drawers.
A smidgen of mess helps make a house feel like a home; a smidgen of "bad design" helps a blog or forum feel like a community.
"Natural organically-evolved cities tend to be densely-packed mixtures of dark alleys, tiny shops, and overcrowded streets. Modern scientific rationalists came up with a better idea: an evenly-spaced rectangular grid of identical giant Brutalist apartment buildings separated by wide boulevards, with everything separated into carefully-zoned districts. Yet for some reason, whenever these new rational cities were built, people hated them and did everything they could to move out into more organic suburbs. And again, for some reason the urban planners got promoted, became famous, and spread their destructive techniques around the world."
(The 'for some reason' is legibility.)
Quoting James Scott on Brasilia's lack of street corners in "Seeing Like a State":
"Most of those who have moved to Brasilia from other cities are amazed to discover “that it is a city without crowds.” People complain that Brasilia lacks the bustle of street life, that it has none of the busy street corners and long stretches of storefront facades that animate a sidewalk for pedestrians. For them, it is almost as if the founders of Brasilia, rather than having planned a city, have actually planned to prevent a city. The most common way they put it is to say that Brasilia “lacks street corners,”by which they mean that it lacks the complex intersections of dense neighborhoods comprising residences and public cafes and restaurants with places for leisure, work, and shopping.
While Brasilia provides well for some human needs, the functional separation of work from residence and of both from commerce and entertainment, the great voids between superquadra, and a road system devoted exclusively to motorized traffic make the disappearance of the street corner a foregone conclusion. The plan did eliminate traffic jams; it also eliminated the welcome and familiar pedestrian jams that one of Holston’s informants called “the point of social conviviality"
Indeed, I think "Seeing like a state" is more relevant here than "Whither Tartaria".
Back when every amateur blogger built their own website in Wordpress or even in raw handwritten HTML, those websites had *personality*, warts and all. But Substack is designed for scalability, which is the Internet version of what SLAS calls legibility.
Substack isn't a state, but Substack together with the other big tech platforms like Facebook and Reddit may as well be considered a state government. And they don't want quirky little street corners, they want evenly-spaced rectangular grids that are easy to administer centrally.
In addition, just to add my voice to everybody else: yes, over-use of Javascript sucks and reduces my six-core 3500MHz desktop PC to something that feels more sluggish and glitchy than the software I ran on my single-core 7.2MHz Amiga in 1985. It's amazing that "just put all the text on a single page so that searching with Ctrl-F works as it should" is apparently a complicated technology nowadays, when the simplest website built by a teenager in 1995 who just learned what HTML is, would *do that automatically without needing to do anything special to enable it*.
Yes. Here's the problem with web designers. They have to use the tools/features that look best on their resume, not the tools/features that give the best customer experience.
Blame web designers less (they're still at fault) and blame the corporations and community that releases framework after framework and fashion after fashion, and the hiring/HR who get glowy eyes when they see $SHINY_TECH on the applicant's CV. Web designers are just responding to incentives the mOvE-fAsT cult sets up.
> It notified you about new posts in a easy and comprehensive manor, you could jump straight to them, and there were a few other nice QOL things.
By "new posts" do you mean "new comments"? The "new comment" logic was a client-side JS thing which was originally a browser extension I wrote, which I could in principle re-implement for Substack. Or we could annoy the Substack devs into implementing it themselves, possibly.
yes! and that was you? Thank you very much, that aspect of the comments was super super useful and greatly appreciated. Yeah, an equivalent for substack would be fantastic and really helpful.
Yup. I'll see about a browser extension when I get some spare time. It's probably only going to work here because this is the only blog on Substack which loads all comments under every post. (They hardcode this domain in their JavaScript, fun fact.)
If you're interested, I've already created an extension (https://github.com/Pycea/ACX-tweaks) that can mark new comments, as well as a few other features like restore the old theme.
Oh nice! Glad I saw this comment before publishing my own; better that there be only a single place for it. You should post this somewhere more visible (like the next open thread, or as the top level comment on this post), and maybe Scott will signal-boost in some open thread.
Nice, thank you! I did have to use the "load pages dynamically" option to keep it from refreshing constantly, so that might be a good thing to enable by default.
I've been using your extension and like it very much, thanx. Question - I'd like it if the "click on the vertical line to collapse a thread" worked in your extension. Is that possible to do? Or maybe I need to get an updated version?
Thanks for making it! That was really helpful and it is amazing that Substack has piles of JS that lags on fast computer but is missing such basic functionality.
The 'new comment' behavior made following long-running conversations a lot more practical, I think that would be a huge improvement if it were available on Substack. Thanks for putting the original together!
> Also, in terms of functionality, the thread system of SSC was objectively better than here. It notified you about new posts in a easy and comprehensive manor, you could jump straight to them, and there were a few other nice QOL things.
Maybe there was some merit on the aesthetics of the old site but, at least on mobile, the comment section was literally unusable. Long threads could get to almost single-character widths (https://imgur.com/F7XrMK4) with no way of reading just from a certain point of the tree. On substack I can at least look at a subset of the comments, and collapse the entire thread from anywhere in it.
I can't speak for everyone but while I would read Scott's posts on mobile I didn't engage in the comments on mobile or really care to (regardless of whether the UI supported it). On the rare occasion I did, turning the phone sideways made the last most indented column readable (even if it was still too narrow).
I certainly get that Substack would be an improvement for people who primarily browse on a mobile device, although I think you could capture the SSC style in a more mobile friendly structure.
I never browsed SSC on mobile and I think the most indented column was occasionally unreadable for me.
I think more generally the distinction not desktop versus mobile, but that really bare-bones minimal HTML is great for content that doesn't require interaction. If you want more functionality for writing comments, or even if you just want to make it easier for the blog author to write posts, then you have to move away from the minimal version of the website, which makes passive reading worse.
This is a good time to mention that threading in general is not great for cases like deep discussions. It's also very frustrating when you post something, have 3 people reply to you. Are you supposed to reply to each of them individually? Reply to yourself?
That's one of the good things about 4chan-style threads: you can reply to as much posts as you want, and you're not constrained by the "tree" structure, but are free to build your graph. Graphs are often better at modeling things, but harder to represent, while trees get you close of what you need. Something like a graph with the article at the center and comments linking to each others could be an interesting experiment.
Geocities and blogs are still web2.0, web2.0 is any user-generated content. Web1.0 is the era of html pages that could only be changed by their server.
You just hate giant-boring-corporations-dominated web2.0, which is understandable.
Substack is a giant ball of single-page-appiness, with all of the glitchy microaggressions that product managers think users love that that implies.
SSC loads, and then it's done. I can scroll around the page, find-in-page (or, google!) anything, and it all works instantly and consistently. I will gladly pay the price in lack of gloss and “user-friendly enhancement” for that alone.
I like the new aesthetic, but all the crappy collapsing of threads, where I can’t even open that collapsed thread in a new tab (stuck behind javascript? So annoying), which prevents me from full text searching, and frequently just hangs or resets to the top of the very-long-page is incredibly frustrating. So much so that I just don’t bother reading the comments much anymore.
Exactly (and there are several other issues of a similar nature).
The styling is one thing, but the machinery underneath amplifies the “meh, I kinda liked the old better” into “god I hate this crap”, which is reinforced every time you visit another substack-based site.
If every elevator button that was round gave a small shock when you pressed it, people would quickly come to hate round buttons, regardless of where they found them.
Yeah, being able to open a subset in a new tab is pretty basic, to anyone who is *not* a professional UI designer, or limiting themselves to microscopic (mobile) screens.
One specific case in point to illustrate this: on Substack, at least in a desktop web browser, there's a header bar above the article showing the blog name and the user avatar and so forth. Except it's not *really* above the article, because it's a "smart" Javascript control that jerkily hides itself as you scroll down. But then if you scroll back up at all, the header peeks out again, covering whatever it was I was trying to scroll back up to see (usually the last few lines of text) unless I scroll even further to get it out of the way again.
As some of the other commenters have also noted, the font size is quite a bit larger, so the information density is lower than on SSC, and the switch from sans serif to serif doesn't work as well on the screen as it might in print. And not being able to see all nested comments at once makes following the various discussions harder.
None of these are a big problem individually, but all together it adds up to a user experience that feels like the page is fighting us. There was none of this artificial stupidity with SSC – it was just a flat page that loaded and then got out of the way.
That's it for me. Substack brings in a pile of JavaScript and won't work without it. Wordpress I can read and even comment in w3m.
(I like reading as many things in w3m as possible, since it does me the immeasurable favor of throwing away all web design and presenting everything the same way.)
It just doesn't support JavaScript. It works fine if the site supports graceful degradation. (i.e. there's a usable document and the site uses JavaScript to add functionality on top of that.)
Exactly! If some idiot would somehow make me imperator of world then after dealing with important stuff (lock up Putin and so on) I would ban making SPA.
Also, it was fun to occasionally scroll through the archive page and look for old posts I might want to read again, but if you try this on substack it loads ~5 new posts and then you have to scroll down and wait a few seconds again to continue scrolling. And each entry takes more than an inch on the page instead of taking only a single line of text.
Making a throwaway to second this. It's the Javascript for me. The moving header bar, literally everything about the page load... Sometimes on my phone I come back to a ACX tab to continue reading and everything below the scroll position is blank, so I have to do a full reload. The back button doesn't work reliably. Increasing the font size doesn't work on mobile. I'm sure there is more. I don't know why tech people insist on breaking perfectly fine websites with this stuff. Just send the HTML to my browser, it will be fine.
In addition I hate the profile pics in the comments and all the rest of the social media stuff. I actively don't want to know how many likes things have.
One other thing that ACX doesn't seem to suffer from, but most other substacks do, is the newsletter form in the middle of an article. It feels like the end of the post every time. It makes no sense, put it at the end.
Obviously the old blog had its flaws too. As I recall it wasn't responsive, so it was sort of a pain on mobile. These things are irrationally more forgivable when it's someones humble personal corner of the internet vs. a large platform.
Speaking of glitchy microaggressions... it allows people like you to have animated avatars, which create constant movement which distract from reading :)
Same. Yes you can technically scroll and ctrl+f on Substack as well, but it can take 10+ seconds to load later parts of the page (especially on mobile) whereas searching/scrolling the old page was basically instant once it loaded.
I think this is a symptom of modern web dev being horrible at every opportunity. Modern websites have tons of Javascript and they are trying to be artful and fancy. This is the trend among designers, but real people just want something simple and straightforward that works.
> If they forced everyone into the standard non-customizable layout of 2015 SSC, would that be a straight utility gain?
No, because a large part of why people like the 2015 SSC layout is because it /felt like SSC/.
Substack, like almost all technology companies, is trying to make their design work for everyone. Every author, every audience. This inevitably results in a design that expresses no personality and has no distinguishing features whatsoever. That is what flat design is, that is what flat design is for, and that is also why people hate flat design.
Right. Right now the blog looks like any other generic Substack blog. For those of us that read multiple Substack blogs, they tend to blend together. Sometimes standardization between people works well - like Facebook, it's easy to find what you're looking for on someone's profile. But that leaves it impossible to create a specific aesthetic for your blog. People might not be missing any particular element of the old design, just that it was specific to SSC.
I prefer the substack design, but liked the writing better on the old blog. I also think you nailed why in the Why Do I Suck post: you were grappling ~7 years ago with stuff I'm grappling with now, and it's well-trodden ground for you at this point that you don't need to rehash but to me it's front-of-mind.
Your old site was uniquely yours, which people like to say they value. People like to associate with unique aesthetics and your old site was definitely memorable—if every Substack used your old site's design, it would perform horribly. It's uniqueness was the main draw, even though the design was worse on its own merits.
I think if you were to look at revealed preferences in terms of engagement, Substack will actually perform better. It's clearly more designed/optimized than your old site.
If you thought there was a replication crisis in academia, try looking at the internal engagement studies that companies use to justify their design decisions to themselves. Utter pseudoscience!
I’m mot convinced that “Every blog has the same design” would actually outperform “Every blog except ACX has the same design and ACX has the old design”. I suspect that a universe where every blog has a unique and well designed theme is more appealing to users than every blog having a shared minimalistic theme. But Substack can’t offer a unique theme to every blog it hosts, so we’re stuck with the minimalism.
The problem with "Every blog has it's own design" is that there are lots of basic functions that it's easy to glitch while implementing. This, however, is a lot different from "Every blog should have the same design".
is key, I think. When I first encountered SSC, I thought "This guy writes great, but his website looks like shit. He hasn't even bothered to put a picture in the header." But after a while that "no visual aesthetic is my aesthetic" grew on me and seemed kind of punk. I connected that punkness with the writing. But then you went and sold out to Big Substack... and now the lack of visual aesthetic seems corporate instead of punk.
Related to https://astralcodexten.substack.com/p/whither-tartaria, perhaps? I find modern web design sterile, in a similar way to modern architecture, and (mildly) preferred the SSC website for some of the reasons -- it had more character.
I like the old SSC because (apparently counter-intuitively) I think *it* is the more minimalistic one.
Substack has a superficially minimal aesthetic, but that's just the outer veneer. It's like a Tesla vs a Lexus or a Macbook vs a Thinkpad. Their formers *look* sleek and minimal, but that's just on the surface.
Ditto, especially adding the other comments' inability to browse the archives. Substack looks minimal but the UX feels a little like a maze. The old blog had more information density but you could find what you were looking for so easily.
I'll also add that the old blog's design really stands out to me among Wordpress themes. I would not mind as much on theme if certain other blogs moved to Substack.
Yes, there's the Myspace over-customization problem, but if a theme is *good*, unique themes really give blogs a nice character.
One thing I hate hate hate about Substack is how hard it is to browse the archives. SSC made that really easy and really enjoyable. Substack, please just give a list of Titles and dates as links!
I was a relative latecomer to SSC and relied heavily on the archives. I think it would be difficult for me to explore the Substack content in the same way. I wouldn't have thought of this on my own but this is a great point.
A few weeks ago, I realized I can use datasecretslox' "Astral Codex Talk" subforum for a more streamlined index and then jump to the post I want on Substack via the thread for it:
(Now, I'm going to attempt to read your mind and guess you're saying, "But I don't -want- to go use a second site to read ACX archives!" It's helpful for me; not something everyone would want to do. The future of the internet is in non-sucky, non-distracting indexing!)
God, this so much. I loved browsing through the old archive. The substack one is a infinite-loading disaster that takes literally forever to scroll through. It's just a list of links! How hard can it be!
Also, in trying to reply here, for whatever reason it forced me to re-log-in, so of course I had to refresh the comment section, which takes bizarrely long, and look for this one again, because the direct comment links are terrible. Thanks pile-of-js people.
The inconvenience of using the Substack archive does seem like a considerable handicap on the staying power of ACX posts. It would require a bit of extra work, but maybe this could be addressed by adding links to each new ACX post on the old SSC archive page.
It has to be said that Substack is easier for me to read on mobile than SSC. But there was a long period in the industry, when mobile first came around, that companies produced very distinctly designed layouts for mobile and for desktop. Substack is shooting for a one-size-fits-all design -- which of course focuses on accomodating mobile, and thus fits *only* mobile.
First of all, they should drop the single-design principle. And then they should really upgrade the archive browsing functionality on the desktop version, at least to the level present on SSC.
Agreed. Scrolling it is a series of 1-5 second delays that just feels terrible. Plus, the tags are gone! So I can't even binge-read the "things I will regret writing" tag anymore!
The comment you posted of someone speaking of their preference for the old design explicitly stated that what they liked about it was its custom, non-professional feel. Professional designers hired by a company to think hard about what constitutes good design and then implement that for everyone basically definitionally can't come up with a design that feels homey and unique for individual blogs hosted by that company.
https://astralcodexten.substack.com/p/why-do-people-prefer-my-old-blogs/comment/6400024 <- Freddy deBoer mentions people liking "the band's old shit" as a constant of human nature - but it's worth asking *why* that's the case. Maybe idiosyncratic associations that people make with an old art style, whether that's a band's first album or your hastily-designed 2013 blog, are still intrinsically aesthetically pleasing to people who encountered that art at the right time, and there's no way for more professional artists/designers to actually create better aesthetics for that group of people.
Yes, but professional designers also think hard about what needs to go on their resume, which leads to what to throw into their designs. Professional designers have long ago figured out that what to put on their quarterly MBOs (Management By Objective) also guides what tools and features they throw into web pages.
Thus, what professional designers need to do to satisfy their career goals may not exactly coincide with what provides a great user experience.
I think a major problem of "professional designers" is they design for an abstraction, the hypothetical universal user persona drinking from the universal content firehose. So they completely fail to take into account what is unique and different about *this* content and provider and *this* audience. Which leads to frustration, not surprisingly, as you try to hammer a thousand different polygonal pegs into the same damn round hole.
It's sort of mystifying why this is considered good practice. It's as if Ford, BMW and Tesla all hired a separate class of "professional designers" to design their cars -- none of whom had *actually driven a car* but only studied carefully the general theory of controlling mechanical devices, supposedly applicable to everything from pencil sharpeners to aircraft carrier arresting gear -- and everybody strived to design just one ideal car that everyone was expected to buy, in one standard color.
Mature industries long ago abandoned this idea. Car companies routinely make 10,000 cars a month, hardly one of which is identical to any other, in a wide variety of sizes, feature sets, and looks, knowing that having each individual customer be able to come quite close to his peculiar ideal is the way to thrive. They pour gigadollars into figuring out ways to become *less* standardized in the salesroom while being more standardized on the factory floor for efficiency. One imagines Web programming will get there, some day. But in the meantime we have to put up with smug statements of design best practices that were rejected as naive and unrealistic decades ago by mature industries.
Well, as far as creative output is concerned, I'm pretty sure that it has to do with artists using up their best novel ideas early on, and then either doing inferior derivations, or chasing popular trends.
I think other obvious "psychological" explanation besides selection bias is that your old readers are just "used" to it, and they liked it because it was familiar and associated it with the Good Old Days of your blog. Your old website and its old design are still around, and your new readers have probably gone to it to read some of your most popular old posts. Do they prefer that design? I doubt it.
I read your old blog on my RSS reader, and I read your new one in my email inbox. I was exposed to your old layout only when revisiting old pieces, and I always sort of regretted having to do it (no offense). If the article was very long I would usually try to find some way to avoid looking directly at the WordPress layout. Sometimes I would read using a browser extension to isolate the article text (Firefox also supports this natively), or sometimes I would send the article to my Kindle. So I have no nostalgia for the old layout, and this is also the first time I have ever clicked through to look at an article on Substack's own platform.
Agree. Anyway you could survey for this, Scott? I read SCC first and prefer it. I read SCC first and prefer ACX. I read ACX first and prefer it. etc...
Agreed. Isn't this just a live fire example of selection bias and the endowment effect?
I, old blog reader, got used to the old blog (I've been endowed with the old format), and you change it on me, and I just liked it better the first way I had it. And then when you ask me how I feel, I feel motivated to tell you that.
Other dude, new reader, never really interacted with the old one, and couldn't be bothered filling out the survey about the whole old vs new thing, because I didn't feel strongly about it. I mean, it's just a blog format, right?
Having a unique design, good or bad, makes you distinctive, and after enough time readers develop strong and warm associations between the design and your with your voice. Sometimes on Substack I forget who I’m reading.
Conflate this with reddit's redesign. An overwhelming majority of people I've interacted with prefer the old design, yet the company insists that the new one is more liked.
I guess it's probably simply just that the kind of people I interact with tend to like a different kind of thing than the majority. Selecting for content does introduce a lot of preferences
reddit's redesign introduces infinite scroll which make things trickier (especially in terms of measuring which is "more liked"; that could really come from users getting manipulated into scrolling "just a little more").
Substack wastes half of the screen and feels like it was designed for mobile or email rather than desktop viewing, probably because it was. Wordpress is built for browser. Substack doesn't allow pinch & zoom within mobile browser, presumably to force me to download the App I refuse to download on principle. Wordpress just works. Substack has a flimsy minimalist aesthetic, especially in the comments section, while Wordpress has beautiful comment boxes, obviously indented reply trees, and is overall happy to use the space it has available. Substack is all white while Wordpress has grey accents, colored comments. Substack defaults to a smaller font, especially for metadata like commenter names, comment times etc., but then DOESN'T USE THE SPACE SAVED AT ALL.
At a more abstract/kvetchy level I feel that the migration of hobbies, blogs etc. from wordpress, ipboard style individual websites to bland subdomains where the creator is just another user, and not the host, is emblematic of the fall of the old internet which I miss dearly. I think it's fair to say that your readership is probably more on the MySpace end of the scale than the Facebook one.
I can pinch & zoom just fine in Mobile Safari, for whatever that's worth. But the Substack app actually doesn't support it. Or text resizing, or search in the comments, or collapsing threads(!). Its only advantage over the web interface is that it renders threads fully, without the disruptive "Expand full comment" and "Continue Thread" functions which were inexplicably added not long ago. Otherwise, the app is a featureless placeholder.
The horizontal space thing occurs because (according to experts) columns are easier to read. Apparently we have trouble eye tracking when moving from the end of a very long line to the start of another, hence why newspapers use columnar text even though they don't have to. This is now conventional wisdom in web design, hence why we get ever wider screens but text remains stubbornly in the centre surrounded by white.
It's probably right though. I can imagine that reading lines as wide as my screen would be very annoying. HTML was designed in a time when the rule was one page = one window, so you could easily resize windows to decide how much horizontal space they should take up. After tabs and popup blockers made every web page a full screen experience, browsers never adapted and there's now no standard way to express how wide you want the content to be except by zooming.
The previous blog used columns too – just wider ones. I find reading newspapers (or did find them) to be tedious because the columns were too _narrow_; too much jumping to the next line.
My explanation is that your blog caters to a subset of the population that prefers information-heavy presentation with options. I can read Chinese and love Chinese news homepages. I can scan 100 headlines in 30 seconds. My blog is like your old one, three columns of info. Someone can visit and hit a whole bunch of links off the homepage. Saves a lot of time.
Most websites trend towards mass audiences if they grow and that means a lower general intelligence. I like Yahoo Finance from around 2000. It was data heavy, very customizable. Now mostly videos, junk articles, almost never visit. It feels like its made for stupid people. Don't go on Drudge much anymore, but that site stuck with all headlines.
"My explanation is that your blog caters to a subset of the population that prefers information-heavy presentation with options."
That is true for this blog, sure. But almost equally true for Substack in general. Substack subscribers literally pay money to get information. And I see no reason to optimize for non-subscribers.
Also why would the lowest decile in IQ of Substack subscribers prefer lower info density in particular?
Not being high IQ doesn't imply you don't want to find stuff fast, as well.
I don't see how seeing more stuff at once would be more mentally taxing.
This is extremely common in UI design. Normally though what happens is after a period of time, people adapt, and then dislike the old design and not the new one. Your audience may be one where they enforce a logical consistency on their views, so repeat what they said before to prove they weren't wrong. Normal people just forget their old views and think their current views were the ones they always had. I suppose a survey of some sort could look for this correlation, so this is just speculative as to why people haven't come around.
I don't disagree that people _do_ adapt, but I think there simply is 'too much' design – for the existing users of anything.
Professionals are – basically all the time – _furious_ whenever the UI of their professional software changes. Even if the new UI is better (and it often isn't), the cost of relearning one's workflow often, drastically, swamps any meager gains from a better UI (when that is the case).
An anecdote that I consider particularly strong evidence for this is that _designers_ themselves were finally subjected to this same treatment somewhat recently when Adobe significantly changed the UI of their tools. A designer I know admitted that they found the new UI to be a big pain! I pointed out that they finally knew how the rest of felt!
I want the current "design/style" with the snappiness and ergonomics of the old one. Substack does not load enough comments (hides too many things by default), and it does too many background requests. Oh, and it's also noticeably slow.
I assume there was something people liked about the older, more personalised design. Perhaps on an objective level the Substack Theme is just better, it obviously has a minimalist elegance, but on a subjective level the old SSC theme probably felt more personalised? That is, crude and primitive web design has a charm that quality lacks?
I think it's (1) the excessive minimalism of solid white margins, plus (2) the shortened line length. Both hit squarely on the professional-typographer vs sophisticated-reader mismatch. The typographers are very worried about people getting distracted by stuff in the margins, and have eye-tracking data from first-time unfamiliar readers to back it up. They *don't* have longer-term data, which mostly points in the opposite direction. And similarly, there's a widespread piece of ideology which says that shorter line-lengths are better, which I think originates in research done on less-sophisticated readers with a different saccade pattern.
The professional typographers have been right about a great many things for centuries, long before the data eventually vindicated them. I say consensuses among typographers are way more reliable than data. Every controversy between typographers and other people I’ve investigated was just the typographers having thought about the problem more carefully and people not bothering to read them.
I don’t know about this “longer-term data” that supposedly “mostly points in the opposite direction” which you haven’t linked to but I am extremely skeptical. And talk of “sophisticated readers” usually comes from the same kind of sentiment that leads people to think that tiny font sizes and bitmap fonts are so great, that antialiasing is so bad, etc., which, no, is not sophistication but ignorance.
I've been a bit interested in typography for awhile but I think it's pretty rare for most designs to follow any of its best practices. I think the original commenter is wrong that the divide is "professional-typographer vs sophisticated-reader". I think it's more 'web design fads' versus typographer-and-readers myself.
Yes, readers overall consistently dislike long line lengths. And that’s a good reason to avoid them no matter their impact on readability. Publishers figured that part out since they have an incentive to sell their material. But there’s a subset of readers who consider themselves more sophisticated and disagree.
Web design often gets important things wrong. I don’t know why and don’t think it’s just fads. People think they need to make text color lighter to avoid eye strain; instead they should make the page background darker and leave text at #000. Text is too light, too small and contrast is too low.
I think we might be writing past each other. I don't disagree that _too_ long line lengths are bad or disliked. I'm claiming that some line lengths are also too _short_.
I think web design gets important things wrong because fads are necessary for 'full designer employment'.
I also think that 'fresh' designs are useful for sites/apps (and their companies or organizations) but that's because they're a (relatively) costly signal to _new_ customers/users.
If web designers stuck to the existing solutions for "important" things, most sites would look much more similar and there would be much less work for designers overall.
There's already a plethora of tools for 'optimal' color palettes for 'pages' and charts/graphs/visualizations – and it's not just web designers that are mostly ignoring them. There's lots of bad (more) 'traditional' publishing/design products/artifacts too, e.g. academic papers.
I think the 'typographic evidence' actually supports the old blog design being better – too short of a line length is bad too. And the 'typographic landscape' is pretty varied too. _Some_ academic articles are published in multiple columns. I think a good bit of mathematics papers tho are often a single column. SSC/ACX sometimes 'reads like' the latter more than the former, so the 'optimal line length' for this blog might be different than others.
Multiple columns are mostly a consequence of a poor choice of paper size because large margins are problematic (aesthetically and for zoom in a PDF document). The recommendation of 66 characters is too low both for print and for screen. Professional typographers have been using lines longer than 66 characters for a long time however.
On Slate Star Codex I see lines with 130 characters. That is too long. On Substack, I see 80 characters, which is slightly too short.
When you are done reading one line, your eye travels near the beginning of the next line. For reading speed, there are two problems if the lines are too long: you may locate the wrong line or you may not reach sufficiently near the beginning of the next line. The first case is one reason you want a greater line height when lines are longer. In the second case, this is only a small loss of efficiency because you can still process the words in the part of the line that your eye has reached before it jumps to the beginning and you will not need to spend as much time on them afterwards.
I think Substack's line length is fine – but I also didn't have much of a problem with SSC. I think the 'optimal' length is a little closer to Substack, but I would (for screens like mine) prefer a larger font and a 'geometrically' longer line length than either.
About the customization, whether better or worse by some consensus metric, Substack makes it more generic. There was a certain feeling of being "at home" in the old one, reaching a known place in a corner of the web. Substack feels somewhat like a newspaper.
Substack keep it generic, plain and boring to allow any/all writers to use it and for subscribers of multiple substacks to have a consistent experience. They don’t want your brand overriding their brand.
& I think people prefer something more unique and visually inspiring on a one off, but not when they have to constantly switch visuals when switching from Substack to Substack.
Nice article, in which data scientists apparently rediscover classical time series analysis (or perhaps, a classically trained statistician explains to data scientists what looking inside the box allows one to do).
One thing I immediately noticed from a product-design perspective when you moved to Substack was that it is *much* harder to read through the archives on Substack. SSC had links to the next/previous article at the bottom of each article, whereas Substack doesn't seem to have this and instead intermittently shows you a 'greatest hits' or something, which I'm sure has tested very positively for engagement and delight or whatever metrics are currently trendy, but is really annoying if you want to just 'read the next page please'. On SSC itself I thought it was a bit naff having to click through 100 meetup/openthread posts to get to the next dose of insight porn, but I much prefer that to not having the option at all!
Yeah, just last week I opened SSC and had this nostalgic feeling. I had been pointed to SSC just a few months before NYT happened. It was, uhm, the first blog I ever read, and I read a lot of it ... so for me the old layout is just linked to the adventure and the fascination of discovering those texts in the first place. I also somehow like the colour.
More broadly, I think by definition it had a more 'personal' touch and at least some people who like your writings might just like that personal touch by association. Substack to me is very fine in general. It's just the same for every newsletter ... no 'oh, there's Scott's place' feeling.
One possibility: When you ask people what looks better, they think about the sites as things to look at. They assign a positive value to ornamental aspects. But when you are actually reading something, ornament is a distraction. What you enjoy when you pay attention to it and what actually helps you achieve your goals are not necessarily the same thing.
I think the key here is personalization. Substack's design is much better for a generic, unified presentation content *platform*. When you are hosting many writers and you want all of their blogs to look roughly the same, what Substack has done is clearly the correct answer.
However, your blog was not a platform. It was just yours, and that allowed it to be idiosyncratic and optimized for you and your style and content. More precisely, your blog isn't just a blog, its a centerpoint of a community and subculture. Most bloggers are not that. I like Matt Ygglesias, but I don't need a special theme to know i'm in his territory, because his territory is just his content. Your blog is/was a whole subculture, and as a sub-culture, it warrants an aesthetic differentiation that merely being a writer does not.
People always complain about design changes. Even if you do a well-constructed experiment and prove that people prefer the new design, most of the reactions you get will be complaints.
The upshot is that you need to weigh the utility of the improved design to your average user against the pain of dealing with the complaints about it.
I think people are very reasonable in complaining about most design changes!
I think part of the problem is that "average user" is often a _new_ user, and the best design for them is often very different than the best for a regular, let alone 'power', user.
And for regular/power users, the utility of _any_ design change is often negative, even if/when it is better. Having to relearn 'workflows' is a significant cost, often swamping any meager gains to be had from better design.
Your web design is your face and your domain is your name. Your name should never (have) change(d), and your face shouldn't change drastically all at once. It's also unfortunate to have the same face as so many other writers. SSC's design probably looked the same as some other sites out there, but it was uniquely recognizable among the sites I frequented. Same goes for just about every blog from that era. With Substack, blogs are just content using Substack's surname and wearing Substack's one-size-fits-all face.
The fact that my peripheral vision constantly reminded me that I was reading SSC seems is analogous to how a person's face constantly reminds me I'm talking to them.
The substack page is horrifically slow on mobile. I’ll scroll and it will take several seconds to render the newly visible context. And this is on a pretty decent phone.
Modern websites (including Substack, though it's offenses in this regard are minor) are not optimized for enjoyability or readability, but rather engagement. Basically, an alignment problem - we want to reward people who create enjoyable websites, but the actual incentives reward engagement (or whatever other things are driving revenue).
I believe part of the problem is the email format. On some phones, you cannot read in portrait, you have to switch to landscape. In addition, which include screenshots of data set points, surveys, et al. Must be clicked and and they view too large to simply look at.
I don't know if this is a major problem, but it slightly irksome to me.
Could be a broader form of selection bias, wherein Substack's layout is preferable to most people but there's some weird correlation between "people who like your kind of content" and "people who prefer older-style, non-minimalist blogs".
Part of UI design is fashion. Many UI designers follow fashions in UI design, and fashions can change without them actually being an improvement.
Many of us just want to read, including the comments. Substack has more wasted space (larger font, higher line spacing, shorter line wrapping around the comments) and requires additional clicks to expand things, requiring moves away from the scroll bar.
Also, it used to be easier to see which comments you hadn't read yet. That lack makes the lengthy comments much more difficult to read through more than once.
Plus the old site had more contrast. Many of us like readability above all. "Modern" design doesn't always enhance that, especially when just made fancier.
Is it my imagination or are there more comments per post on the Substack version? Substack makes it too easy to say something -- consequently it's harder to "wade" through the sheer amount of content and perhaps the comments are less well-thought out. For example, I never once commented on a single post before the blog migrated to Substack but now I do from time to time. I know it's a dull instrument but what I wouldn't give for some sort of comment upvote system...
Maybe the problem is that substack was designed by really experienced designers, who are living in a bubble, where they produce what *they* like, receiving acclaim and the occasional award from their industry peers, and insulating themselves from any users so crass and uneducated as not to share the professional opinion. (Which IMNSHO, would be most users.)
Alternatively, the problem is that substack was designed for one-to-many interaction, with only token feedback, much like newspaper comment pages. With newspaper comments, some percent of the commenters are ranting one-topic *ssh*les, spammers, and other annoyances. Moreover some perhaps even larger percent are political opponents of some readers, and so considered by those readers as even worse than the less political problem posters. So the harder it is for readers to read other people's comments, the better for the newspaper. And if it's also hard for the original poster to respond to comments, well that's perfectly fine; if my local paper is any sample, newsies eagerly ignore anyone who tries to interact with them.
Though if I were to be specific, it's not the layout that annoys me on substack - it's the whole look and feel, including but not limited to the tools for commenters. We had better thirty years ago, before html was a thing.
"Before HTML was a thing"... do you mean Usenet or something like that? Or did you perhaps mean Javascript, not HTML? Before HTML, there was no Web at all.
I was about to add that you'd also have to go back more than 30 years, but that's wrong as it turns out; HTML turns 29 this year.
I think the mere *fact* of customization explains why I prefer the old SSC design, more so than any of the actual design choices you made.
When I read a website like SSC that has its own look, I feel like I'm *located* somewhere, like I'm in a specific physical place. And there's an intimacy to it, as though someone has invited me into their home.
When I read any Substack, I don't feel like I'm "in a place" much at all. And if I am, that place is just "Substack." Likewise with Facebook and other standardized venues for content.
If every blog looked like SSC, then my reaction to the SSC design would change from "I am physically inside Scott Alexander's private salon, this is what the chairs and curtains look like" to "I am reading a blog."
Mildly interesting sidenote: tumblr (uniquely?) serves every blog in 2 versions, the Myspace-style customizable view and the Facebook-style standardized view. Consider the difference between
The latter feels more place-like intimate to me, as I've described, even though the extent of my customization goes little further than "making the background a particular shade of desaturated purple." As long as nothing else looks the same way, that's enough.
I'm pretty sure I answered "about the same" in the survey, and I stand by that. My general feeling is that the old blog was kind of ugly, but not offensively so, and the new one is nice, but not in a way that particularly stands out, so overall they're both just pretty okay. Consequently I suspect any preference for the old is nostalgia.
"Is it selection bias? My previous readership is, by definition, people who liked my old blog, so of course they like my old blog more than some new one? I’m including this because I know someone will bring it up in the comments if I don’t."
Substack is pretty good I think. The app is better. But …. I often see replies to my posts that just show the reply. The only option then is to “show thread” which goes back to the main thread, as far as I can see, and it’s not always clear what the poster was referring to.
It seems to me mostly nostalgia. Vladimir's displayed comment seems to me to read like "It was worse, in a way I was used to." (e.g. "The blog name was more quirky, harder to remember... And hey, that felt memorable!" is literally saying it being bad made it good).
Honestly this all seems like overthinking it (which is not a complaint, I'm very fond of your overthinking).
Mostly I think people just hate it when things change. I mean, what's the last time that you updated an app and went "ooh, I LOVE how everything looks slightly different now"? (And this is classically even more pronounced for autistic people...)
Now, I absolutely count myself in that camp! I like the old layout too. Change is the worst, and the old one was comfortably familiar.
My big complaint is that the comment architecture is so primitive. On SSC I could go back to a post I was interested in the next day and easily find the new comments by searching for ~new~. On ACX there is nothing at all like this, so I usually wait to read a thread until it is a few days old, and then never go back.
I never read this on mobile. The thing I don't like AFAICT is the current blog has a blank white background, identical to every other substack, without any customization or charm. It makes the text feel almost oppressive, having nothing to hem it in at all.
I'm not sure if this comment is earnest, but both the old site and Substack allow you to collapse comment threads. On the old site, it was very helpfully marked "hide". On Substack, I think it wasn't possible at first, but this was one of the most-requested features here, and I think Scott's large audience motivated Substack to incorporate it. Now, if you click one of the vertical lines to the left of a comment, then it will collapse that comment and the entire thread following it.
Just… don't do it if you're deeply nested, as it doesn't adjust the scroll location, so if you're 5 screen lengths into the thread when you hide it, now you're 5 screen lengths past the top-level comment that followed.
Yeah, I really hate that! That, and the fact that the link in the e-mail I get about comments replying to mine is only about 70% reliable at actually finding the relevant comment in the link, and otherwise drops me off at a random point in the comment thread so I have to try to find the relevant comment.
Funny enough, I think that it _sometimes_ working is probably a (minor) technical accomplishment – I suspect they must be doing some crazy something something to show you a specific comment. That it doesn't _always_ work feels like evidence that it's not trivial, whatever they're doing.
Linking to a specific item within a HTML page, was a solved problem in the previous century. On the old SSC site, just right-click on the timestamp of any comment, choose "copy link address" and then see what happens when you open that link in a different browser window/tab. You can also send that link in an e-mail and then people can open it from their e-mail client.
Of course, if instead of building a website by using standard HTML features in their intended manner, you go and poorly reinvent all kinds of wheels using unnecessarily complicated Javascript techniques, you will break all those well-established standard browser facilities. So then you need to poorly reinvent them too. And so on..
Yeah, it was rolled out in a really hard to understand way. (At first, it allowed you to collapse all the replies to a single comment, but not a comment and all its replies, which was *extremely* annoying when I wanted to skip some subthreads on a thread, but not all of them.)
For me I think it's just nastalgia, but I do prefer SSC. When you first joined substack, I installed some sort of "make ACX look more like SSC" add-on, which does a pretty crappy half-job, but I still get annoyed when I read ACX on a new device without it and it just looks like substack.
my personal answer is “I’m very used to the old layout & I’m not used to the new one yet.” I do mostly read on mobile, but I still preferred the older version tbh.
This is not complicated. Your SSC layout had STYLE, and it was your personal style. Presumably your readers like your personal style, otherwise they would not be your readers (I do like it). Substack style on the other hand is minimalist, out of necessity, because it has to cater to a diverse audience, so it has no personality. EDIT: also default Substack font for articles is imho awful, though it is better in comments.
This maybe overlaps with Whither Tartaria. I weakly preferred SSC's old layout over Substack. The reason I did was specifically is that it was, in a cute way, kind of ugly and quickly thrown-together, and it made me feel like I'd found a brilliant blog that is niche and obscure, because a website as ugly and quickly thrown-together can only be niche and obscure, and thus not only do I enjoy reading a brilliant blog, but get to pat myself on the back for finding it, quite literally, "in the rough". This is pretty resistant to reality too, e.g. my enjoyment of SSC's layout and my feeling of it being something niche and underground didn't really diminish when I realized that NYT picking on a blog probably means it's not niche or underground.
At the same time, when I come back to thoughts of how warm and fuzzy the old SSC was, I do not remember that we couldn't collapse comment threads. So I guess, for me the explanation that it is nostalgia for the "band's old shit", as Freddy deBoer puts it, is correct.
Wild guess: humans are neural nets, and the high quality of the old blog caused its layout to trigger a reward function. We like the layout because we read good things there, so our brain assumes the layout must be good. All reasoning after that is our attempt to provide a logical justification for an illogical association.
If the is true, a similar or more extreme pattern should be visible in new readers who spend time on substack without seeing the old blog.
The black sans-serif text on white background feels sterile and very visually immediate. The ACX logo is way off in the corner, and there are no visual borders. The navigational structure of the site is now "list of articles + a few sidelined sections." This makes your site feel like an unformatted powerpoint where each slide is a wall of text, or maybe like an email client on default visual settings. It's just one white text surface with no layers or edges or footholds. This design makes you inaccessible.
I get that this layout is mobile-optimized, but if you are thus required to cede aesthetic vision to this umpteenth degree... why have a website at all? You could just send emails for people to read in their email apps.
In contrast, the light-blue and faint-gray color scheme with the fonts you had chosen made your articles seem like they had come from a strange book. And, there were multiple sections that contained diverse types of items - some of which I never fully understood! I liked that I had to puzzle some of it out. The old site was like an illuminated manuscript. It had layers, and pages like a book has pages, and errata, and appendices. It felt like you were writing it and simultaneously writing in its margins. Its design invited people into you.
I don't find myself wanting to comment here as much as before. Some of that is my own change in habits, but Substack's design sure made it easier to disengage.
> You could just send emails for people to read in their email apps.
This is almost exclusively how I consume ACX, to be honest. Nowadays, I only come on the actual site to read comments whereas I often actually opened the old blog in a browse, because it felt "just as nice" as my fairly custom email experience. Substack looks like every other blog aggregation site and even conjures up thoughts of medium.com etc. Why would I want to look at this site when there's really nothing to look *at*.
Somewhat related to the first thing you posted. If you're a fan of a sports team, and you first fell in love with that sports team in and old, dingy arena. And then they move into a brand new state of the art arena.
I bet a lot of fans would say they preferred the old arena. Especially if a number of other teams make a similar move to similar arenas around the same time.
We associate the old design with Scott's writing (and, to borrow from the other post, when you didn't suck :-) ). Now, it's not unique. It's the same of a bunch of other writers, which we have varying levels of warmth toward.
I think Substack's is better because of text's max width & visual minimalism.
The biggest problem IMO is lack of organization through. Kinda hard to find specific texts. If at least there was a separation between cyclical stuff like Open Threads and text content...
I didn't read your old blog at the time, but I've read tons of it now. I prefer the new layout. The only way I don't prefer it is that it looks "generic" and like every other Substack. With the old one, I know who I'm reading. But you're pretty distinctive so I don't mix you up with, say, Matt Yglesias, the way I might mix him up with Noah Smith. Also, Substack is way easier to read on a phone IMO.
I think substack are trying to relentlessly focus on their core proposition rather than bells and whistles. They are very much trying to let the content be the king on mobile and desktop and as a blank canvas they aren't trying to be personalised in a nice way like SSC was. I think they will eventually allow you to customise it, but it's not core to their mission of getting writers paid.
I really don't think you did a bad job especially after the designer tweaked everything, it reminds me of gamer sites which isn't necessarily a bad thing. In the end despite how important designers think they are, how it looks barely matters at all.
• Not all comments are visible when loading the page. (Or weren't when I last checked without ACX Tweaks. Now they do seem to load immediately. When they didn't, this had various consequences beyond that you have to click multiple times: you couldn't easily search in all comments, or save where you've left off by clicking the permalink of a comment.)
• When viewing the thread below a comment, the blog post isn't visible.
• No way to see which comments are new.
• No formatting in the comments.
• No way to jump to the parent of a comment.
• Inexact timestamps.
• When following a link to a comment (or just loading a page with a comment's url) it doesn't jump to the right place. (In Firefox, you can then go to the address bar and press Enter to jump to the right place. In Chromium, that reloads the page, and jumps to a wrong place again.)
• The header that is hidden when you scroll down and shown when you scroll up is annoying. (If you scroll down just a bit more than you wanted, you can't undo that by scrolling up just a bit.)
• The entire archive can't be loaded at once. On the Archive page, I have to scroll to the bottom repeatedly to get to an old post.
• SSC had nice aesthetics, ACX has none.
The ACX Tweaks browser extension (https://github.com/Pycea/ACX-tweaks/) fixes some of these problems, but not all, it doesn't work perfectly, and most people presumably don't use it.
1) I have near crippling, drop out of grad school/lost a job/have problems with tasks of daily living ADHD when my stimulant wear off. Substack's pro design has S_H_I_T implementation. Every hang up for .75 seconds. Every 1 second animation instead of just loading the page. Ever "lol I'm going to reload for no reason" event while trying to scroll back to a comment from a non-new-tabbable read more button. They are just another aggravating. Just shove a pencil in my eye.
2)Same reason a family owned Italian joint can feel like family while Olive Garden feels like a strip mall. Teams of experts aren't an assurance you'll get good art.
The old SSC layout gave me the vibes of the old bulletin boards, "internet before the Eternal September", and "this exceptionally interesting but little-known and therefore sort of pleasantly elitist thing". While uniform and minimalist layouts are likely more readable and even more aesthetically pleasing in some way, they are also a sign of times when big and often disliked corporations want to constrain people's individuality and self-expression into homogenous boxes.
1) I like knowing where I am on the internet. While simple, SSC felt like its own place. No substack feels like its own place.
2) The smaller font on SSC felt more congruent with the detailed style of writing. More content on-page helped me to keep more of the reasoning chain in-view.
3) The sidebars added loads to the culture! The local ads! The local blogs!
4) It wasn't actively bad. People weren't complaining about the old site (I expect). I think things would've been different if you'd really broken the old site somehow.
I guess what I'm saying is that the old site had a unique character (in being visually different from other blogs and having the site UI be filled with things of local-interest), it wasn't actively broken, and (I think) the smaller font was a better fit for the detailed content with long chains of reasoning.
>2) The smaller font on SSC felt more congruent with the detailed style of writing. More content on-page helped me to keep more of the reasoning chain in-view.
It may be that this says something about my short term memory more than anything about web design, but this definitely strikes me as true. It also makes it difficult to read quickly because I'm constantly jumping to the next line or scrolling, it makes the thoughts/arguments in the article feel more disconnected.
Almost all web redesigns (well most brand redesigns too) suck by removing personality and density in favor of cleanliness. It's the same reason I hate modern architecture. Actually clunky small font websites like old reddit, ssc, and goodreads are good. (Speaking for desktop)
Actually I want to rant about the goodreads redesign since that's new. Lots of people have been complaining about this for ages but I hate the revamp. Yes it is less "ugly" but that ugliness removes a bunch of functionality or requires more clickthroughs to see info like isbn or page number.
Consider your blog. The new app removes the little sigil at the top and the colors. It's a stretch but these could be considered functionality by serving the purpose of ssc-territory/tribe signaling. The ads were all rationalist ads further encasing the bubble.
Personally, my issue with Substack (and why I never tried using it myself) is just that it's not customizable - every blog looks exactly the same, and that's no fun. Even a bunch of blue squares that you slapped together in a few days is nice because it's *yours*.
I have no way of knowing how many other people experience the same thing I do, but here it goes. Substack presents a tiny column of text with right around double the text's width in whitespace on the left and right side. The old style wrapped properly for me.
Slate Star Codex's homepage feels rich, full of content, like a library of old books full of mystery.
Astral Codex Ten doesn't. It sort of diminishes how you feel while reading certain posts.
Perhaps you might ask people: "do you think I should go back to the old layout, if Substack let me?" and see what they say. I'd say no, despite what I have just written.
>Slate Star Codex's homepage feels rich, full of content, like a library of old books full of mystery.
I think it is important to distinguish this from "nostalgia" which a lot of people are pointing to. At least for me, this was an active part of the appeal of reading SSC from the moment I first got linked there. Getting linked to ACX would have felt very different if I were coming in blind. I'm sure arguments could be made for that ultimately being good *or* bad. Just kind of a "vibe" thing (Substack doesn't have one, of course).
In the early days on Substack, the biggest thing for me was the inability to collapse comment threads. They've fixed that, but there's still a lot of problems with comment threads, some of which I've seen other people mention. It's hard to click through to a specific comment, there's some slow-ness in the loading of comments, you can't have both the main post and the comments visible (unless it's an open thread), and so on. My guess is that a lot of this has to do with how Substack stores the comments. (As I'm writing this, I just noticed that Substack has an option to view comments "Chronological" or "New First" - I don't quite know what the difference is, but if they are sortable, then that might require some sort of data structure that makes the above problems inevitable.)
The other big thing I miss here is the whole archive feature - links to "next post" and "previous post", and links to each month of the archive, and a browsable front page of the whole archive. I think a lot of this is a self-conscious change from the world of blogs, that were seen as chronological writing, to the world of social media, which are seen as ephemeral rather than things to come back to. (It's an interestingly different pair of conceptions of what "Web 2.0" meant!)
All the visual touches and the blogroll and stuff are fine, but it's really these structural things that make the relevant difference. And it seems that Substack expects most authors to prefer things like sortable comments and ephemeral posts, rather than an archive, and that's why their thing, that is more professional, is worse for what your readers want.
It looks like "Chronological" means "Old First" (which is how SSC forced comments to display). Substack should make that terminology more obvious, since "New First" is a kind of chronological too.
Mostly the colours, honestly. Substack is a uniform, eye-searing white. Slate Star Codex had a grey frame, a bit of blue on the top, a bit of light grey on the sides. The overall effect was easier on the eyes. (But also, I don't hate the Substack layout and design. I just prefer the Slate Star Codex one.)
The latter I mercifully had yet to notice, because I've not had the urge to browse this site's archive, because if I had, I would certainly have been very annoyed at yet another instance of the far too prevalent 'lazy-loading content' paradigm. I flippantly want to say "god created scrollbars for a reason" - but basically that would just be a tongue-in-cheek way of saying that I *like* that a scrollbar tells me how much more content to expect at the bottom of the page, and this is completely nerfed by the lazy-loading content paradigm. Prefer clear pagination if we do need to chunk it for size.
>Mostly the colours, honestly. Substack is a uniform, eye-searing white
Yes, the wall of white is a little painful on my eyes. I don't think I would prefer a full on Dark Mode, but the muted colors of SSC were more pleasant, at least for me. Some people have mentioned the that the smaller, Serif font on SSC resulted in denser blocks of text that made it easier to consume some of the more challenging writing, which strikes me as spot on. I think this is the biggest thing for me, although there are a number of other good points brought up throughout the comments.
As someone above mentioned, it's about margins. Our eyes want margins. Color is nice too. MR still has its classic green. There's no Marginal Revolution without margins and green.
I prefer the compactness, generally, and increased information on the non-single-item pages of the old format. The layout within a single post/article is fine enough (though not better). This isn't limited to SSC.
And the comments section. Of all the areas they could really improve functionality, the comments section has to be the most wasted opportunity. Nothing that is "improved" actually improves the user experience or promotes more engagement, generally, and it's absolutely terrible for highly active comments areas.
I suspect there's a signaling mechanism at play--the expense in building and maintaining a distinctive, functional blog theme acts as a signalling proxy for the quality of your blog. By contrast, appearing as a generic substack or medium blog--while functionally identical or superior to your custom theme--signals nothing other than "I signed up for an account with one of these services."
Everyone hates the new minimalist designs when you ask them about it, but they engage more with them and they're way easier to adapt to different browsing environments.
People engage a lot with Twitter but I avoid that website because it's fucking awful in every way, design wise. "Engagement" doesn't mean "good," it doesn't mean "memorable," it doesn't mean "improving your life in any way." It means you're wireheading people to click the "Next" button. No more, no less.
That's where they're falling down by contrast with the old blog and it's *badly* needed. It's cutely retro that we're back to using *x* to indicate italics like the early days but it's no longer the Century of the Fruitbat, you know.
The trouble with Substack's layout models is that they are pushing very hard for writers/authors/content creators. 'Here's how you can start your own Substack! Here's how to price subscriptions!' Meanwhile, if you want to know how to use Substack to post comments, it's "Oh, here's how you subscribe to a Substack. You want to know about comment formating? Hey, here's how you start your own Substack!"
And sure, that's their business model: get people producing content that draws subscriptions that they take a slice of for their operating costs and (fingers crossed) profits. But it makes it difficult at best, and actively lousy at worst, for anyone who's a reader but not interested in creating their own Substack. Your job is to passively consume and pay for the privilege.
By contrast, even though WordPress can be a steaming pile of horseshite, the old blog was simple and did what it said on the tin. Even a fool like me could learn simple HTML tags for formating, links and so on. It didn't need bells and whistles, it needed "Here is Scott's post. Here is how we comment and have discussions" and that is what we got. Flashy but superficial design puttering-about wouldn't have contributed to that, and wouldn't do anything. Yeah, great, your paid theme is ultra-high-minimalist post-ironic Zen repurposed via Bauhaus and so achingly chic it has won design awards five years from now, but I can't find the 'sort by new' button.
I prefer the old layout, probably nostalgia. A few quick notes:
-There's a big difference between reading articles in your inbox and on a website. Substack feels perfect for email but very sparse for a website. There's no links, the commenting sucks, it feels very "read and leave" not "read and stay".
-I've found myself clicking over to other substacks a lot more on the website. Sometimes this is cool, I especially like it when I see one writer I like commenting on another writer's article (shoutout to FdB) because it brings back that old blogosphere vibe everyone is missing.
-Again, reading from the inbox, it feels very, I dunno, "special" or "professional". The inbox is where real things happen. But the website feels like "content". Like, everyone's consolidated around some standard practices which are very optimized but they "smell" like content.
From a design standpoint substack is fine. A little bit bland
From a technical implementation standpoint substack is AGGRESSIVELY bad.
Some of the open threads put my poor old desktop to its knees. It's just text on a white background, why does that require layers upon layers of shitty javascript?
FYI - from someone who does research - the way you frame that question is designed to get a specific response. You might frame the question differently: Is ACX or Substack easier to read?Easier to Navigate? From there you might actually get more pertinent results. People don't like change in their visual environments. Think of all the hand-wringing about Google's new logo, Netflix's new design, Medium's new logo, that California School's new logo, the new...[new design of something] and there is a lot of scorn for it, but it's all gone in moments. In a few months, people can't even remember that Google had a serif font and don't even care. It might be that people don't like the attachment to what SUBSTACK represents - that ACX felt more personal to them because it was designed in 2 days. It doesn't mean you have some secret visual taste because people prefer it - you definitely don't. It's what it meant to them. I don't personally care, but I deal with this all the time. A colleague calls it the, "You moved my cheese" problem.
Piles of JavaScript, horrifically bad performance (it displays text and occasional image, and it brings my computer down), unneeded elements.
http://prize.hutter1.net/ may be going too far (I would add margins at least), but SSC was for me in nearly optimal position. Substack broke several things and added nothing better at all and added plenty of JS infestation beyond what was useful or needed.
And removed useful things.
For example there is no way to see just new comments.
Dysfunctional threading - sometimes just scrolling is enough to hog my computer (16GB RAM and displaying formatted text is beyond it!).
Unwanted extra stuff: "subscribe now", subscribing to access it, subscriber-only posts.
Some of that - like trying to spam my email for no good reason, subscribing-related fluff is a direct cost of monetization.
Part of that is team of designers breaking things that worked well, partially because they actually think that it makes thing better, partially to justify their employment.
> Whenever anyone asks, people say they hate the modern minimalist designs and wish they could go back to the complicated colorful ones. But for some reason nobody ever does. Is this just the Internet version of the same general phenomenon?
I designed my personal website exactly how I like it. But it gets about 1/10000000000000000000000000000000000000 traffic of FB.
I don't think you can separate "Why I suck now" from the site thing very much. You have this overall perception among some that things have changed, and that they've changed for the worse. This may or may not be the case, but if it is:
1. You would have some people who were overall just unsatisfied, mad about the writing/topics/takes or whatever, and taking that out on both you and the site
2. Same as above, but didn't want to talk about the writing (in rationalism you can't be mean!) so they talk about the site
3. Same as 1, but they don't even realize it's writing-related; it's just a general sense of discontent
So those 3 are all possible, almost all certainly happening somewhat based on what you've written. But you also have a site that swings hard-nerd and pretty spectrum in it's readership; you know who hates change in comfortable, favorite properties? That segment.
All that to say, so long as you are comfortable with what you are writing I wouldn't worry too much - it would have probably had to have been exactly the same to not get a lot of complaints. And if it was actually the same, it probably still would have got a fair amount of complaints, just aimed at different things.
(postscript afterthought: Another thing that happened is you were gone for a long time, and the comments section broke, permanently, in the way that it will never be what it was again. You can take everything people might believe about your writing and apply it to the comments, as well)
I assume some of the more objectively better things (faster to load comments, easier to read and scroll comments) are just giving people an associated warm fuzzy about the other aspects (minor layout, color, etc.)
There's some kind of horrible stop-the-world pause on Substack that freezes the page for several seconds at its most pathological. It doesn't happen reliably and I have no reproduction steps other than scrolling around.
I get the impression that it's something synchronous in the JS that halts execution for a while, and that it causes small freezes as well as big ones. For example, while typing this comment, input has frozen for half a second multiple times, then caught up.
I don't care what the design looks like as long as the text is vaguely readable and the visual noise is kept down, but a short blog post causing the page to freeze is ridiculous. I can go to Project Gutenberg and scroll through the entirety of The Count of Monte Cristo (which would take more than 1000 pages to print) without any slowdown.
No offense, but Substack is considerably easier on my eyes :) I for one really appreciate the minimalism, I get easily distracted, so having the content surrounded by tons of links in sidebars and topbars wasn’t really my thing. Substack mostly just keeps out of the way.
OTOH, I’m not really an SSC veteran — I read a few posts in the years before ACX, but didn’t become a regular until after the transition. So I don’t really have any nostalgia for the halcyon days of SSC to go on.
Also, I can imagine how people who *can* afford to get distracted / read faster than I do might miss that link-laden context in the UI. But as for straightforward having an aesthetic preference for ACX — that doesn’t really compute. But hey, that’s how aesthetic preferences work, I guess.
My mobile experience has been that my phone browser's "show simplified view" option works more reliably on the old site but it's less necessary on the new one, which is probably a reason in favor of both WordPress and minimalism.
I don't really think there's a mystery here. Your tech friends had something like a decade and a half to optimize a set of features and layouts exactly to your readerships preferences, and substack has ... not. Two things strike me. First, the tech people who were making changes actually talked to your users directly. A lot. There were no barriers between them, and in many cases they were the same people. Substack, to the extent that they get feedback at all is probably mediated by third parties, mostly you. Second, even if substack knew how to make the perfect set of features for you, there's no reason for them to do that unless it would benefit enough of the other bloggers and therefore substack. I think most of it comes down to whoever it was that designed your old blog thought of your readers as the users, but substack thinks of the authors as the most important users. It shouldn't be a surprise that your readers prefer the former on net.
My theory is that the previous site had a better fit between form and content. Scott's pieces are long and involved; they require investment and effort on the part of the reader. The font, line-spacing and layout made the text slightly more information-dense and very slightly more difficult to read, which accords with the higher-than-usual difficulty of the content. Like having an ornate frame on a highly-detailed painting.
The Substack font, line-spacing and layout are easy to read. They appeal universally, and offer very little friction. So seeing the same high-effort writing presented in this low-friction way is jarring. Like printing a treatise by Rousseau in an Animorphs book.
Oh, it's lots of things. The old blog had more color, the old blog had more features, the old blog had more personality - this one is pale, poorly featured, and bland. Would much prefer SSC.
I prefer if each blog that I read looks different, so that I'm aware in the back of my mind who I'm reading. The context change when I go to a different blog is reflected in the different surroundings. The various associations/expectations that I have with the blog also get attached to the layout, and I get used to particular ways of interacting with the site.
So any change to site layout has short-term negatives, as I lose that familiar context and it's a bit jarring. This particular new layout is very similar to lots of other sites, so it can't grow as much into a familiar distinctive context.
Substack is far too white. SSC had colors and sidebars. Substack is glaring and barren.
I think that most people prefer websites with white backgrounds, although certain groups of people, like coders, prefer darker backgrounds. I wonder if ACX is sufficiently tech-adjacent to have mostly people who prefer dark backgrounds, or at least backgrounds with some color. It would be an interesting thing to include in a future survey.
And for a literary description of the effect whiteness can have on the human psyche, we turn to Herman Melville:
What the white [website] was to [Scott Alexander], has been hinted; what, at times, he was to me, as yet remains unsaid.
Aside from those more obvious considerations touching [Substack], which could not but occasionally awaken in any man’s soul some alarm, there was another thought, or rather vague, nameless horror concerning him, which at times by its intensity completely overpowered all the rest; and yet so mystical and well nigh ineffable was it, that I almost despair of putting it in a comprehensible form. It was the whiteness of the [website] that above all things appalled me. But how can I hope to explain myself here; and yet, in some dim, random way, explain myself I must, else all these chapters might be naught.
Though in many natural objects, whiteness refiningly enhances beauty, as if imparting some special virtue of its own, as in marbles, japonicas, and pearls; and though various nations have in some way recognised a certain royal preeminence in this hue; ... and though, besides, all this, whiteness has been even made significant of gladness, for among the Romans a white stone marked a joyful day; and though in other mortal sympathies and symbolizings, this same hue is made the emblem of many touching, noble things—the innocence of brides, the benignity of age; ... though in the Vision of St. John, white robes are given to the redeemed, and the four-and-twenty elders stand clothed in white before the great white throne, and the Holy One that sitteth there white like wool; yet for all these accumulated associations, with whatever is sweet, and honorable, and sublime, there yet lurks an elusive something in the innermost idea of this hue, which strikes more of panic to the soul than that redness which affrights in blood.
This elusive quality it is, which causes the thought of whiteness, when divorced from more kindly associations, and coupled with any object terrible in itself, to heighten that terror to the furthest bounds. Witness the white bear of the poles, and the white shark of the tropics; what but their smooth, flaky whiteness makes them the transcendent horrors they are? That ghastly whiteness it is which imparts such an abhorrent mildness, even more loathsome than terrific, to the dumb gloating of their aspect. So that not the fierce-fanged tiger in his heraldic coat can so stagger courage as the white-shrouded bear or shark.
...
Bethink thee of the albatross, whence come those clouds of spiritual wonderment and pale dread, in which that white phantom sails in all imaginations? Not Coleridge first threw that spell; but God’s great, unflattering laureate, Nature.
I assert, then, that in the wondrous bodily whiteness of the bird chiefly lurks the secret of the spell; a truth the more evinced in this, that by a solecism of terms there are birds called grey albatrosses; and these I have frequently seen, but never with such emotions as when I beheld the Antarctic fowl.
I remember the first albatross I ever saw. It was during a prolonged gale, in waters hard upon the Antarctic seas. From my forenoon watch below, I ascended to the overclouded deck; and there, dashed upon the main hatches, I saw a regal, feathery thing of unspotted whiteness, and with a hooked, Roman bill sublime. At intervals, it arched forth its vast archangel wings, as if to embrace some holy ark. Wondrous flutterings and throbbings shook it. Though bodily unharmed, it uttered cries, as some king’s ghost in supernatural distress. Through its inexpressible, strange eyes, methought I peeped to secrets which took hold of God. As Abraham before the angels, I bowed myself; the white thing was so white, its wings so wide, and in those for ever exiled waters, I had lost the miserable warping memories of traditions and of towns. Long I gazed at that prodigy of plumage. I cannot tell, can only hint, the things that darted through me then.
...
Nor, in some things, does the common, hereditary experience of all mankind fail to bear witness to the supernaturalism of this hue. It cannot well be doubted, that the one visible quality in the aspect of the dead which most appals the gazer, is the marble pallor lingering there; as if indeed that pallor were as much like the badge of consternation in the other world, as of mortal trepidation here. And from that pallor of the dead, we borrow the expressive hue of the shroud in which we wrap them. Nor even in our superstitions do we fail to throw the same snowy mantle round our phantoms; all ghosts rising in a milk-white fog—Yea, while these terrors seize us, let us add, that even the king of terrors, when personified by the evangelist, rides on his pallid horse.
Therefore, in his other moods, symbolize whatever grand or gracious thing he will by whiteness, no man can deny that in its profoundest idealized significance it calls up a peculiar apparition to the soul.
...
But not yet have we solved the incantation of this whiteness, and learned why it appeals with such power to the soul; and more strange and far more portentous—why, as we have seen, it is at once the most meaning symbol of spiritual things, nay, the very veil of the Christian’s Deity; and yet should be as it is, the intensifying agent in things the most appalling to mankind.
Is it that by its indefiniteness it shadows forth the heartless voids and immensities of the universe, and thus stabs us from behind with the thought of annihilation, when beholding the white depths of the milky way? Or is it, that as in essence whiteness is not so much a colour as the visible absence of colour; and at the same time the concrete of all colours; is it for these reasons that there is such a dumb blankness, full of meaning, in a wide landscape of snows—a colourless, all-colour of atheism from which we shrink? And when we consider that other theory of the natural philosophers, that all other earthly hues—every stately or lovely emblazoning—the sweet tinges of sunset skies and woods; yea, and the gilded velvets of butterflies, and the butterfly cheeks of young girls; all these are but subtile deceits, not actually inherent in substances, but only laid on from without; so that all deified Nature absolutely paints like the harlot, whose allurements cover nothing but the charnel-house within; and when we proceed further, and consider that the mystical cosmetic which produces every one of her hues, the great principle of light, for ever remains white or colorless in itself, and if operating without medium upon matter, would touch all objects, even tulips and roses, with its own blank tinge—pondering all this, the palsied universe lies before us a leper; and like wilful travellers in Lapland, who refuse to wear coloured and colouring glasses upon their eyes, so the wretched infidel gazes himself blind at the monumental white shroud that wraps all the prospect around him. And of all these things the Albino [website] was the symbol. Wonder ye then at the fiery hunt?
Why do people like your old design? Was it objectively good? It was certainly more complex and had a signature, and was not cookie cutter, and indicated an actual human cared about it. But these things are not inherently antithetical to a "clean" or "sparse" design.
As you've asked, why did Substack—in other words, everything digital—end up boring and awful? Because designers deliberately engaged in an anti-aesthetic, and anti-intellectual ideological mass preference falsification through fear and intimidation. They consciously led the entire web going to shit, and exactly zero designers (more accurately, approximately a single one) fought back against it.
For anyone wondering... yes, I did cover the question of responsive design and mobile. And the short answer is no, these developments in screen viewport variability absolutely did not necessitate us getting boring layouts. That is the fault of ideological designers who hate you.
I think it could have to do with the uniqueness of the SCC website, and the way that this effects information retention and engagement.
I remember noticing when I started using a kindle that things I read would sort of blur together and I would find it harder to coherently remember an individual book. Whereas a paper book has a distinct colour, size, shape, mass, smell etc associated with it.
I find a similar thing happens with substack, where there is nothing distinct between any of the blogs to latch onto. This makes it harder to kickstart the process of information sorting, and there are less sensory clues for memories to form onto.
SSC blog had its own character. Substack ones all look the same, and the design is made as lowest common denominator for a million of blogs, all alike. It's not optimized for SSC reader/writer, it's optimized for whatever criteria Substack has. Just like a lot of supermarket food tasted like crap because it's optimized by price, durability, transportability, etc. but not taste. I wouldn't say Substack design "tastes like crap", but definitely quite a bit bland.
It's not a huge deal though - I wouldn't stop reading the blog because of it.
1. Modern UX design with its obsession with whitespace and disdain for borders is a plague upon humanity.
2. The old layout had marginally more characters per line, which I think strikes a better balance there.
3. The two points above combine to make it so that on the new site, 2/3 of my monitor width is pure whitespace, while on the old site, more than half the width is being used, and even the unused parts are better since they aren't bright white lights.
I think the whole thing is just path dependency - if it had always been the other way around, no one would have minded - and also, the only people to say anything are those with strong feelings about it being bad. You solicited the comments in a way that maximizes negative feedback.
If I needed to take a stab, I'd say it's personality. The old blog with its mediocre-matched colors wasn't the prettiest thing, but it was its own thing. It was also a bit reminiscent of the 'old internet' with a lot of distinct, personal sites, which definitely helped with flair. Your substack, on the other hand, looks like any other substack. In fact, it pretty much looks like any other modern text website - if I take my glasses off, l could not tell whether it's medium or substack or even New York Times (I'm exaggerating a bit, but I hope you get the point).
That being said, there are also a few trade-offs. Mobile experience is much better on substack. Comments, on the other hand, load in chunk-wise and, as one of your quoted comments points out, does not work too well. This progressive loading is great for SEO and page speed (since your page is not ten times larger due to 500 comments), but it's not to great for reading large threads. It's quite obvious that they are a second-class citizen and are especially not optimized for the kind of in-depth discussions this readership generates.
Then, of course, there's also some nostalgia thrown in and some resistance to change, but I think the above is enough to explain a lot already.
I was someone who came to your old blog relatively late, only a year or two before the switch. I don't have strong opinions one way or the other, but I did notice that the comment formatting on the the comments of your old blog was kind of broken for at least 6 months before the switch. I tried multiple browsers and it didn't fix it. No one else seemed to be complaining about it (and I asked on the subreddit but never got a response), so no one else seemed to have that issue? But it made the comments completely useless for me, so I like not having that issue on substack.
The old one had a retro, unique, weird aesthetic. New one is more generic. In a vacuum, new one would probably be better, but the old one stands out amidst the sea of identical substacks
"Meanwhile, Substack is run by tech industry veterans who probably hired a team of really experienced designers and thought really hard about every aspect of their product. It doesn’t make any sense at all for me to do a better job than them. So what’s going on?"
What's going on is that web designers are (A) aggressively bad, having developed their own culture of what websites "ought" to look like, which I suspect is driven to constant churn by the need to seem novel common to most artists, (B) adhere to a lot of stupid design principles from Apple which say you can never be too minimalist, (C) believe people don't like text, and (D) take Twitter and Facebook as models of website organization.
I would ask 3 designers. Some people hate change and are very vocal about it. Even if something is better, they want it the old way. As a usability engineer and designer, I test and ask people for their opinions by testing them. See how fast they use something. Design should stay in the background to the meaning. Get people the content they want most quickly and clearly and you have excellent design. People won't notice great design. Myself, I prefer text that I can turn to audio which allows multi-tasking. I do read on a small chromebook but I listen on mobile.
What do you like? Content is king they say. When I'm coming to read words, visuals are to support the message. That's just me. I'm a designer and people are so busy that my design work stands in the background to what people came for.
Who says this? Designers? Product Managers? I like to *enjoy* the content I'm consuming, I don't necessarily want the visual equivalent of Meal Squares while doing my web browsing (and for me specifically the sea of white isn't just neutral-boring, it is actively unpleasant). I get that some people *prefer* the flat white look, I'm not disputing that, but I'm taken aback by your confusion that some people would prefer something else.
I feel like the prevalence of themes and dark modes supports me here, do you have some reason to think differently?
Statistically, there are people who read dark mode better than black on white but most people can read black on white better. Myself, I prefer written content in audiobook format for multi-tasking. Several people who commented say they read only on their smartphones while other read only on their laptops.
The fastest way of reading is with a line length similar to newspapers. For me, while I don't mind certain vignettes while consuming content, I mostly prefer to stay focused on the topic. Otherwise, I have the natural environment as background. That is enough to make it unpleasant to try to stay focused.
While there are time when I need some visual "advertising" in order to gain my agreement and attention, with Astral Codex Ten, the concepts are interesting enough and my own internal vignettes are enough to entertain me and fully interest me.
There are certainly people who would prefer something else, but I suggest to them they may not be interested in the content and if they aren't just switch away from the article they are reading and move onto something that entertains or interests them.
I see you've added a pink background to your last comment. That can soften a message, as pink is shown to affect human anger response. So, I say, adding pink supports your message without distracting from the words which are your content. So in this I agree.
SSC felt more like a niche little community (the rationalist targeted ads were always cute, made it feel like I was in the 1920s reading an underground newspaper), whereas the design of ACX makes it feel a lot more standard.
I was immediately reminded of this video (https://www.youtube.com/watch?v=0OB1g8CUdbA) I came across recently, featuring a DIY gadget built on Raspberry Pi that lets you plug in legacy hardware through ethernet connection and gives you a physical dial to find old web pages on Archive.org Wayback machine. The author checks out how various sites have changed over the years and makes notes of overall shifts in trends.
Now, I am broadly of the opinion that many old web pages tend to look ugly, but I don't think anyone will deny that they at least used to look more personable! However, my sense of ugliness relates to types of garishness (such as animated GIFs and scrolling text), and impractical design choices. In fact, I kinda dislike the modern "flat" designs what with their pure white backgrounds and all. The old SSC site comes from an era where it still felt somewhat personable but has the modern degree of practicality and elegance. Indeed, it's in various ways more practical, such as page width that I find more suited to desktop monitors, and clearer quote nesting in the comments, whereas Substack, like many modern sites, feels as though it's been designed for mobile first and foremost.
I don't like text on an endless white background. Even just a light gray background to frame the text makes things feel nicer and less sterile. A blog should feel cozy without feeling cluttered. Not having a mass of links and banners on individual post and instead confining them to the overview page is much better though.
I like clean and simple, but your substack is visually indistinguishable from many others. Even a very simple color scheme would help make this place *feel* like its your blog. At the moment, it's almost as if your posts are your just several amongst many on a common site (like thought catologue or medium).
Am I so out of touch? No, it's the maximizers with granular data who are wrong.
>It doesn’t make any sense at all for me to do a better job than them. So what’s going on?
ACX is a little under 2 orders of magnitude more profitable than SSC, right? Feels like the starting assumption shouldn't be that the designers didn't do their job right.
Oh, substack is well designed for making money. That is, after all, probably the only thing its creators care about.
That doesn't mean they provide a good user experience - just one not quite bad enough for users to stop paying, given the lack of competition. (Medium is worse.)
With substack, I can read Scott's articles - generally by email, because the site sucks. If I work hard at it, I can comment/read other people's comments, but without the same kind of interaction I enjoyed at SSC. (I don't even bother to go to the site for more than a couple or articles a month.) For interactive commenting, I'm better off on the bulletin board site set up by some of the SSC refugees; it lacks Scott's content, except to the extent that people respond there to posts he makes here. But the software is designed for conversation. And it's old enough to lack any number of modern misfeatures.
If Scott's primary aim in blogging is to make pots of money, he may well be going about it the right way. That was *not* his aim back at SSC, but maybe it is now.
Or OTOH, maybe you've spent too much time studying economics, and now consider money, value, goodness, and utility as synonyms, all best measured in $.
It is absurd to try and compare the competence of different actors based on their products without consideration to their differing goals and constraints. Even pretending that the metrics used here are useful ones, user experience is neither what Substack now nor Scott previously were optimizing for. It's a nice-to-have, but not something the website is first and foremost built to deliver.
Sure, we can talk about design elements and their resulting tradeoffs until the cows come home. But the question "why is Substack's design *worse*" is built on a faulty premise, and most of the purported explanations are swinging at air.
>just one not quite bad enough for users to stop paying
"Substack is run by tech industry veterans who probably hired a team of really experienced designers"
That is exactly why it's bad.
In an industry that generates so much profit that it literally does not matter whether you do a good job or a merely passable job, "experienced" does not mean "competent". It means "has internalized years of cargo cult practices".
There is no competitive pressure to make the best design - at best, there is pressure to copy the designs of the biggest companies, and those companies themselves are already big enough that people are just going to use their products regardless of how bad they look and how user-unfriendly they've become.
Nobody is going to unilaterally stop using Youtube, or Twitter, or their iPhone because of a little extra whitespace here and there, or because buttons lost their bevel, or because the website loads and runs a little bit slower due to all the extra bloated javascript running in the background. Everyone will bitch about it, but keep using the product, and so anyone who is trying to do A/B testing in these places will only get pure noise out of any measurement they're trying to make.
Fast forward a couple decades and that's how you get to the situation we're in right now, not just in web design but generally in all computer software. Everything is slower, buggier, and less usable than it should be, sometimes by multiple orders of magnitude. As a fun example, 3d video games are running enough math to compute and draw an entire three-dimensional world with tens of millions of triangles and complex interacting physics, and they're doing it SIXTY TIMES EVERY SECOND. (at least! More than twice that if you're using a 144Hz monitor). That is, they're doing it once every ~16.67 miliseconds. (6.95ms at 144 frames per second). Consider that fact, next time you open some boring 2d software on your computer and it takes a couple *seconds* to load a dozen flat buttons and images, and then you click on a menu and it inexplicably hitches for a few *hundred milliseconds*. Consider what kind of code could be written that manages to waste on the order of a billion cpu cycles, to do something we were already doing in the 1970s with computers that were at least 10000 times slower.
Software developers will often be quick to come up with excuses as to why it's actually reasonable, that everything is more complex now, and have you thought of X and Y, but there is no explanation you can come up with that explains a discrepancy this massive.
It's fully reasonable to think "these are professionals, who are being employed by very successful companies, and they are intelligent people who work very hard, so they can't possibly be doing a bad job - this must simply be the best humans can do", but it is wrong.
They are intelligent people, who work very hard, and the companies who employ them are very successful. It just doesn't matter if they do a bad job.
>In an industry that generates so much profit that it literally does not matter whether you do a good job or a merely passable job, "experienced" does not mean "competent".
>there is no explanation you can come up with that explains a discrepancy this massive
(epistemic status: i've been doing webdev on easy mode for the past three years, so my intuitions on what's exactly going on under the hood with fancy modern web frameworks may be off.)
From my experience maintaining terrible websites: the reason sites load so much slower isn't rendering, it's network latency. Unlike a game, which usually has all its assets already on your machine and packaged to load as efficiently as possible (and where a ton of work is done up front in loading screens that are frequently much longer than a webpage load), a webpage needs to wait for its requests for each asset to make its way over to the hosts, and for those assets to come back over the network, usually in an unpredictable order. Network traffic is WAY slower than code! You can't do network traffic on a graphics card! (This is why Google Stadia was an insane idea that failed almost immediately.) When something loads slowly, it's almost always because it's waiting on a network request to finish, not for render code to run. Open any webpage, hit F12, and go to the Network tab, and you'll see exactly how many milliseconds of load time each request accounts for.
...and on doing that, the first thing I notice when trying to reproduce complaints about comments taking a while to load is... am I crazy? It looks like it's *dynamically rendering and unloading every comment every time it scrolls in and out of view*, for some reason. And part of that render pipeline appears to be "making a new request for the user's icon image". *What?* You never want to put network requests in the render step! That can't be what's happening, can it? There's gotta be a more efficient way to load all those default icons up top. I can get not wanting to load long comment threads immediately, since (for the typical website) most users never scroll down to them, but this is overkill. With how much reading-the-comments happens on ACX, this dynamic on-demand comment loading might even be costing them more bandwidth.
That said, in the broad strokes, the rest of this comment is on-point. "Everything is more complex now" *is* the answer, but it really really really doesn't need to be. Tons of network traffic on the modern web is devoted to tracking scripts and ad service and anti-adblock arms race nonsense trying to obfuscate the site's operation as much as possible. Why do something the simple way when you could do it a fifty-times-more-complicated way that'll make it marginally harder for ad blockers to tell when you're serving an ad? And why let a static webpage do something when you can instead dynamically load every little thing and tell your engagement metrics about each and every detail along the way? A big chunk of "advancement" in the webdev space has been dedicated to facilitating this kind of anti-user design. Insane obfuscatory practices have become industry standard.
But yeah, back to the Substack question- they don't seem to be doing nearly as much crap as most modern sites. It doesn't seem to have any third-party tracking pixels or ad junk, presumably due to their revenue model. Most stuff's limited to their API and CDN, and there's not a whole lot of code running on the client end. It's always been pretty fast for me, though I can reproduce the issue with comment loading times when there's a lot of them, yeah. My best guess without looking at the source code is that the "floatiness" of the comment scrolling is a result of getting these icon loads stuck in the dynamic render pipeline somehow- like the same icon is a subtly different URL every time it scrolls into view and that makes the browser try to reload it, or something. And that'll vary with the user's network speed, so you have some users reporting slow floaty comments and some having no trouble at all.
This is a standard whinge about software "bloat" dressed up with some numbers to make something that is not very mysterious at all seem like a deep riddle.
There is a kernel of truth here: incentives are playing a role. But the notion that the entire software industry is doing a terrible job simply because they're rolling in such fat profits that they've ceased caring about quality is just absurd.
Games are fast because the ability to render high-quality graphics without lag is what matters to the success of games. The entire history of the computer gaming industry going back to the 80s and presumably much earlier is one of engineers engaging in extremely clever hacks to squeeze every possible ounce of performance out of the available hardware. There is a reason that graphics GPUs are used for machine learning.
Meanwhile, the major constraint on the development of desktop or web-based software is engineering resources. I have been a software product manager for about 25 years now, and I would estimate that a typical launched software product contains about 10 - 30% of the features its creators want it to have. 70 - 90% of features end up cut because the resources to build them simply don't exist.
This is a number I am very much pulling out of my ass, and I will freely admit that there is a massive gray area here. In one sense, I can imagine almost infinite features for a product ("Then in phase 2, we will add an AGI to this to-do list tracker"), so of course most features don't get built. But that's not what I mean at all. I'm referring to fairly obvious extensions or enhancements to a product that will deepen the feature set, make it more pleasant to work with, or otherwise offer value to its users. I'm referring to things that the software "should" do but doesn't.
And then there is the bug backlog for the product, which in general is going to be a sucking black hole of unaddressed problems. For example, here are just the publicly visible bugs that have been recently filed against various Google products:
Click any "recent issues" link to see just what has been filed in the past few weeks. Some of these bugs are quite bad! My guess is that Google would prefer that these bugs not exist, but fixing them would mean not devoting engineering resources to other tasks, and so a choice has to be made. And this is Google! They have all the money and all the engineers. I gaze upon my own bug backlog and despair.
Of course the maturity of the software in question matters. Microsoft Excel has been around for literally decades, has millions (billions?) of users, and ungodly amounts of resources behind it. It is a mature product. Of course, Microsoft is still pouring tons of resources into Excel and I'm sure engineering time is still a constraint (the AGI add-on isn't so absurd in this case), but it's fair to say that Excel is more than 10% feature complete.
I'm not denying that software bloat is a real thing. Software gets bloated. Many of the cut features are bad ideas that the world is better off without. But the reason software is bloated is because getting rid of bloat is really hard, meaning expensive in engineering time, and doing so would involve direct tradeoffs with all the other things engineers could be working on, such as bug fixes, new features, etc.
So, yes, the incentives matter. But those incentives are driven by user needs, user priorities, and ultimately by what users value. In video games, lag is death -- unresponsive games are frustrating and awful. For web software, what users often want more than anything else is to pay nothing for good-enough functionality and performance. That is absolutely their right. But the outcome should not be a surprise.
You've basically nailed this. For a 3D action video game, 60fps >99% of the time is the #1 developer design goal. For a blog platform, I'm pretty sure "keep latency under 200ms" (let alone 16ms) is probably not even in the top 10. Competing goals are: high uptime, billing reliability and reporting, analytics, reliable email delivery, easy to make stylistic changes without crashing the site, etc.
I mean, sure every project would also like their offering to be fast but even with tens of millions of dollars of development effort allocated to a project you still need to make tradeoffs and nobody really believes they'll lose users just because this new feature caused page load times to take an extra 50ms. Obviously, accumulated latency can add up and become pathological but then they probably have a month or two where they focus on speeding up page load times, it's probably not a core part of their DNA to keep a well funded team continuously on top of this that can veto everyone else.
Adding money is also not a panacea. Though we've made progress enabling developers to work independently without stomping on each other the last few decades, software still has a ways to go before you can just double the size of a development team and get twice as much rate of improvement. (Without even being cynical, it was the norm for a long time that adding developers slowed the rate of improvement)
Tesla is extraordinarily well resourced, and nobody would believe they aren't motivated to add improvements as fast as humanly possible, but I'm still regularly surprised (but not really) that some obvious, pure software improvements are only making it into over-the-air software updates today. I have no doubt that developers obviously thought of some of these features years ago (or users have been suggesting them for years) and the span of time between "idea" and "appears in the next release" is frustratingly long because software is still not a solved problem.
Yup. People really have no idea how vast is the chasm between what exists in a software product manager's head and what gets built, simply for lack of resources. And part of the reason that software engineers are so much more productive today than they used to be (which they are) is that they can use an extensive set of libraries and services that take care of much of the infrastructure of coding. But using libraries and services means exactly that you're going to give up on some of the flexibility and optimization you might get by coding things from scratch. At least in theory, if you were so motivated to invest in optimization.
Someone wrote and then deleted a response to my comment that read, more or less, "lol it's just a blog, if it's slow then the engineers are incompetent." What's interesting is just how wrong this notion of "just a blog" is. Wordpress is the most widely used blogging software in the world, and about 43% of the world's websites are built on top of Wordpress. Anyone who has ever used it knows that it is an insanely complicated content management system, which is about what you would expect from a 20-year-old piece of software that supports 43% of the world's websites.
Software is expensive to write and engineering resources are highly constrained. Stuff gets optimized if it needs to be optimized. Otherwise it doesn't. There's no mystery here at all.
This is really interesting, and complements what I was saying perfectly: it is actually extremely costly in terms of engineering resources to eliminate latency, and the there is no way to justify the opportunity cost when there are so many other things to work on and so little pull from users to address these issues.
Not only that, but user preferences militate against one another in this case. Nothing stopping lower latency from existing, except you may be unable to do more than one thing at once or be on the network.
Not true, largely because your premise — that inside monopolies, individuals act with impunity — is incorrect.
While tech co's might be 'successful no matter what design is picked', this doesn't mean that the individuals making said decisions don't have a large incentive to find the best possible design.
They do: promotion.
^The incentive of 'promotion' generalizes all the way up to the CEO, in the form of performance-based grants + the (variable) value of their equity.
Further, we've never been better positioned to differentiate good from bad design. For any given metric, Substack has a large enough 'n' — and sophisticated enough instrumentation — that they can see which design optimizes said metric.
Of course, there are common failure modes associated with this reality. They generally smell like: incrementalism, near-termism, and prioritizing revenue ahead of user-perceived value.
I think there's a trend in UI design to create things that are visually striking, minimalistic, and interactive. Sounds great! But I'd probably have used the words flashy, lacking, and cumbersome. As an example, we know that Google is not lacking for funds for good design, but here are a few changes Google has made over the years that frustrate me still to this day, maybe over a decade after some of the changes were implemented:
- Google adopted menus requiring at least two mouse clicks to do anything, while also making it impossible to cycle many links/buttons using the tab key. This gets rid of descriptive text links and compresses everything into a sleek interactive icon with snappy clickly menus. Looks great, but less usable interface.
- Gmail by default greatly expanded the margins/padding between emails, lowering the number of emails you can see on a page. It looks "clean" and "zen", but has about half the number of emails visible at a time. (Fortunately this can be turned off.)
- The new Google icons are uniform boxy rainbow line things, making all their icons so visually similar that they might as well all have the same icon at a glance. But they are vivid as anything and super recognizable for brand association.
- Google search used to have longer form descriptions for each search item, and links to a "Cached" page, "Similar pages" search, and some other features. Now they have an extremely short snippet and a hamburger menu that creates a pretty ugly popup monad that doesn't seem to have very useful information available (this is in Beta for me, not sure what others will see).
In general this oversimplifying is the trend for modern design. I imagine most "power users" that rely on a tool for better productivity don't like these sorts of changes, but I also understand that clean visuals and snappy graphics draw people in and often make it easier for new users to learn.
On Google Search — when testing this out just now, I discovered that the "cached page" and "similar page" are still there in the new beta menu, all the way at the bottom. I thought they were gone for good!
0. IDK, overall I just like the old one better. I don't always understand my aesthetics well enough to fully (or even mostly) verbalize why I like them, and this is one of those times. You say you have no taste, but I also have no taste, so maybe you disproportionately attract people who have no taste — and so we all like things with color and art and backgrounds, and dislike blank white featurelessness.
1. I dislike the general homogenization of the internet. Mainly the way in which everything is now on the same couple websites, and I don't really stumble across small hidden gems anymore. But less importantly (and possibly just out of nostalgia for that time), there's the aesthetics — I dislike the way everyone from you to the National Bureau of Economic Research abandoned your old unique styles and switched to the same flat minimalistic design. Even if I did think the new design was aesthetically better, then still — I don't want everyone to have the "best" design anymore than I always want to eat the "best" meal. I like variety!
2. If I hit <space> to scroll a page down, then <shift+space> to scroll a page up (or vice-versa), then I don't end up in the same place I started. I end up a few lines higher. I am possibly the only person in the world mildly annoyed by this, but you can empathize with me if you're in this xkcd: https://xkcd.com/1271/. (In the analogy, Substack is the last row.)
3. Relatedly: I don't like floating headers. I find them mildly annoying. (I'd like JavaScript stuff everywhere if it led to more variety, but it seems like it doesn't.)
(The last two issues can be fixed with the "Kill Sticky" browser extension, but I have to click the button every time I go to the site.)
I do like the title images on articles. Even though they're sometimes too small for me to make out what the picture is.
In general, the more modern Web design tends to look slicker, cooler, and more elegant. Meanwhile, the old hand-coded Web design is easier to use. If you want to sell a website to investors, you need to impress them, and thus you go with the modern design. If you want to satisfy users, this is the worst decision you could make -- but users don't pay the bills, VCs do, so who cares about the users ?
Not directly related, but somehow connected. This 9 years old post discusses one very interesting difference between web design trends in Western countries vs. Japan. Here minimalism has become the norm, while over there sites continue be extremely information dense. The author makes some tentative explanations for the difference and why it remains, so worth considering:
The selection bias of most of your current readers having enjoyed your site enough to come back consistently enough to become a 'fan' is probably one part. Another is that the sleek, monetized, minimalistic design that many companies target for broadest appeal is unappealing to many seeking out substantial content due to its association with blogspam and lower quality content. While the opposite association was true of your old website (ties into first point).
For me personally, I like reading the posts better on Substack. I think that is what they’ve focused on — they sell themselves as the newsletter app (transient information). What I miss on the old site is the navigation / organization. Hopefully they will add more “blogging” (persistent information) features as they grow to encompass more blogs.
It’s much harder to find older posts on Substack, and it’s lacking the links to other blogs. This makes the blog feel devoid of context — cut off from the community of blogs. It also makes the posts feel transient and devoid of context — when I’m reading it I know it will be hard to refer back to it later, and I can’t navigate to other related posts via categories/tags.
I think the ability to add tags to posts would help a lot, as would the ability to add some persistent/pinned posts besides the about page.
"In every art form, complicated colorful designs transition to “modern” minimalist designs over time. Whenever anyone asks, people say they hate the modern minimalist designs and wish they could go back to the complicated colorful ones."
This is the opposite of the problem with SSC v. ACX. The problem is the same as the problem of old.reddit vs. new reddit. Old reddit is cleaner, more readable, and with less buttons popping out at you. Ditto the most recent youtube update. What you have is significantly reduced accessibility and usability for competent users, presumably so they can better optimize ad sales and direct very unskilled users to the right videos.
To me it's simple: the substack design just has way to much whitespace. When I view it in full screen on a wide screen monitor, about 2/3 of the screen is just pure white. I'm a dark-mode everything kind of guy, so the pure white light just burns my eyes. Your old blog had smaller margins and they were a tasteful grey- much easier to read. The text also looks like it's spaced farther apart than your old blog, so more white light there. Then it gets worse in the comment section, as comments nest inward to create even more empty white space.
Flakey theory: designers and visual-artists in general really love designs of pure white. I see it in a lot of trendy restaurants and bars, too, and on fashion models outfits. But I wish they'd understand that the white color on a SCREEN is different from the white on a wall. It's emitting light, not reflecting it, so it feels very different.
I've tried some, but I find it breaks too many websites (or just makes them look really weird) to be worth it. I haven't tried Sauron though, maybe I'll give it a shot.
"Meanwhile, Substack is run by tech industry veterans who probably hired a team of really experienced designers and thought really hard about every aspect of their product. It doesn’t make any sense at all for me to do a better job than them. So what’s going on?"
I'm not sure you're super familiar with how the tech industry works. We are pervasively incompetent at absolutely everything. Assuming that because tech people did it, that they have expertise, and therefore worked hard and made good choices, is absurd on a level I honestly struggle to describe.
Yeah, there may be some form of Gell-Mann amnesia at work here. Scott is always great at describing in loving detail how flawed, and sometimes downright dysfunctional, his own field of medicine and psychiatry is -- e.g. how most doctors just follow the herd in deciding which medicines to prescribe, because they lack the scientific literacy to read a research paper and draw their own conclusions.
And then when talking about a field he's not personally familiar with, such as software engineering, he apparently assumes we're all flawlessly competent professional experts. Sorry Scott, but dysfunction and follow-the-herd behavior are just the normal human condition, I'm afraid.
Groups ... or even individuals often do things that scratch their itch, target personal goals, satisfy departmental goals, or even corporation goals at the expense of customer service—whatever form that may be.
Let's say that Scott & Michael each lead platform design teams within XYZ.com, the latest blog-hosting site. Scott & Michael each have teams competing to design a better blogging platform. Just because Scott's team delivers a better design, internal politics may select Michael's team's design contrary to the actual corporate goal of delivering the best designed platform to the customer.
Likewise with architecture. The well-connected elite public building architects aren't out to design a building which pleases the actual public, they're out to design the building which gets noticed in the elite architectural journals, which will lead to architectural society awards, which will lead to more elite white-washing, and more elite public works projects ... ala "The King's New Clothes."
So Substack probably didn't deliver the very best platform it has in it's quiver ... it delivered the best internal political solution—for better or worse.
A better version of the selection effect: People fell in love with SSC/Scott's blogging between 2013-2016 and their formative, most passionate years of that relationship are associated with a particular visual stimuli. The new visual stimulus is associated with the later, mellower years. Naturally meditation on the former induces certain warm and fuzzy's.
It's because the font in the old blog's comments is serif and the new blog's comments are sans serif.
(Also every infinite scroll piece of shit design that I hate but has proliferated across the web despite how fucking annoying it is. But mostly the serif/sans serif.)
This is an incredibly common problem in design. If you do a user study with people who are used to something, they'll complain about just about any style change, no matter what it is. However, when you do a user study with users who haven't used either, you'll find they hate or are unable to use the old thing.
Big tech companies have learned that the solution to this problem is to introduce changes *extremely* slowly so users have time to adjust subconsciously. For example, change the shade of a color slowly over several iterations so no one notices, or move a single button at a time.
While it is possible your old design actually was better, if you want an answer to that you must run a study that removes the "no-change bias" (made this up, not sure if there is an actual name for it). Easiest way to do that is to hire a testing service that can recruit people who have seen neither, but that is expensive.
Most of your users will eventually adjust to the new design, but it could take many years before they accept and integrate it. Some will be permanently stuck in the past due to things outside of your control, like them finding help in your early blog that was very impactful in their life, causing them to attach to that old design the same way someone can end up with a permanent attachment to a particular song or scent.
>Is it something something mobile? I put no effort into optimizing my old design for mobile phones, so maybe that adds another layer of complexity. But I think at some point some web designer friend made a version that worked for mobile, so this can’t be too hard.
Yeah, the issue is that Substack's designed for portrait (mobile), so a 16:9 PC has half the screen as eye-searing, useless whitespace (SSC doesn't fill 16:9 either, but it uses more of it and the borders are at least *not white*).
At a wild guess, I'd say SSC/ACX's readership slews more toward PC than Substack's audience in general.
EDIT: Upon rechecking, SSC uses essentially the same amount of horizontal space for actual text, but it uses a smaller font for the main posts so it still fits a lot more per line. And, again, it uses the sides for something relatively dim rather than eye-searing white.
>Yeah, the issue is that Substack's designed for portrait (mobile), so a 16:9 PC has half the screen as eye-searing, useless whitespace.
I'm surprised by the number of "obviously it's just people being resistant to change" comments given the pervasiveness of "dark mode" as an option these days. It isn't a universal preference, but it seems pretty clear that there is decent chunk of the population that doesn't like looking at their screen and seeing a sea of white staring back at them (and this is just one of many reasons people might have preferred the old site besides nostalgia).
I think some of this is just learning curve stuff. Some people, particularly those who've made an investment and climbed a learning curve, will tend to prefer higher information density and quick availability of related information, useful functions, etc. For widest appeal, though, lowest common denominator will usually prevail. Think vi/emacs or a modern IDE vs Notepad.
A well-designed blog is more than just graphic design. Experts/nerds, in my experience, will favor higher information density, while noobs will prefer not to be overwhelmed.
I should add, for my part, I have no idea what things look like on substack vs SSC, because now I pretty much always read posts via email, whereas I previously pretty much always read them via an RSS reader.
Professional web developer here. I don't love Substack's design, but I like it MUCH better than the old ACX layout. Substack lacks personality, but it's very readable. The old site was both bland and messy. I actually applied my own CSS to it to hide a lot of the elements in the old one because I found it too noisy.
Speaking as a software engineer, you're not a professional software engineer. I don't mean that as a criticism! I mean you're playing a different game. "Correct" design for you means that you and your readers all like how we read each others' words. "Correct" design in a company means hitting deliverables, which for engineers means not reinventing the wheel and instead relying on other peoples' work. Oh and the product is designed by committee. Oh, and half that committee will turn over during the project. Oh, and sometimes the external libraries just doesn't work or don't work for your use. Oh, you might have taken on too many use cases in order to please your board. Oh, serving ads sometimes degrades your own product. And so on.
The more minimal the product, the harder it is to load it down with wrong opinions. Maybe your opinions aren't as precisely calibrated as a professional's, but you don't need that many significant digits. It's words. On a page. And the constellation of ideas that those words explore.
If someone is asking the wrong question, it doesn't matter how good their answer is. Companies aren't asking your questions, they're asking their questions. Sometimes those are close enough that you can squint and pretend they're the same. Sometimes.
On the SSC page I see your constellation of ideas on the left and a substantial piece of your content in the center, and some housekeeping on the right. On the ACT page I see a whole lot of whitespace, and your headline and your first two lines. Going out on a limb, many of us like how you explore ideas in a compounding, introspective way. One of those pages is designed to do that. One of those pages shows clickbait AND NOTHING ELSE. There's no sense that we're exploring something larger, gathering tools to understand the world. There's just the screaming void that begs a question whose answer is probably "no".
Generally speaking, I think Substack's design is better than SSC's, but - part of the reason I like SSC's design so much is the association with your writing, and I like reading your writing there because it just feels right. When I read SSC it brings me back to when I first found you and read some of your best posts. Substack doesn't ever make me go "oh, I'm now reading Scott Alexander" (though the writing style still does). I guess that's the power of differentiation, even if on its face there's an option that's "objectively" better.
For me, the biggest difference is that this doesn't *feel* different from any other substack. On the old site, it would open and I instantly would just know I was on SSC. Now, all substacks kind of blend together. The other day, I started reading one of your posts until I realized "Wait, why is SA talking about the electoral implications of a senate bill? Oh, wait, which blog is this?"
There are a bunch of other frictions too. The site is a bit slow. The login behavior often feels weird. The big "Subscribe now" button right above this comment box is an eye sore. But none of those are as big a deal as the fact that it has a different *feel*.
You should get data on new readers and which percentage of those prefer the new vs. old layout. (Even if they haven't experienced both in detail, they can still have opinions.) I predict new readers will prefer the new layout.
Frankly the fact that only 36% of old readers prefer the old layout seems really low to me, and a sign that the old layout has less sticking power/nostalgia than I would expect.
Basically my prior is much higher for people making decisions based on what they have more experience with and much lower for people evaluating the designs "objectively" in any sense.
Someone may have said it already, but the main reason why I prefer SSC’s layout is that Scott’s old posts are often worth re-reading, and the Substack archive page makes that harder, one has to scroll down for ages to get to even the not very old posts. That’s why, even though I discovered ACX first, I still have a fairly strong preference for the old design.
Every website is always getting worse, which I attribute to optimizing for mobile. (I never read anything on mobile devices.) But then other people on this thread say that SSC was better specifically because of mobile, so shrug emoji.
For me it's almost entirely nostalgia. There was something warm associated with the old design. I've spent years and years reading to it. I've read so much to it it's become engrained.
I loved the old name. Slate Star Codex felt arcane and potent in a way i feel didn't translate to AstralCodexTen.
I dont mind the substack layout nearly as much as the godawful site performance. Even on my gaming PC, let alone on my (admittedly rather cheap) phone, I have noticable delay when I scroll over the comment section. And I am 100% sure the problem is caused by abnormal CPU load, not bad internet.
Very simple thing: I like being able to scroll through multiple posts on a single page. There is a continuity to your blog that is totally lost in Substack.
When you do a series of posts on a similar topic, in SSC I can easily view all those posts in chronological order as well as see what other things were on your mind around the time you were posting them. On Substack I have to navigate back to the homepage, then navigate to the archives page, and instead of being able to ctrl+f the title of the post, I have to scroll down, wait for it to load the next set of posts, scroll down some more, wait for it to load, etc. etc. etc. until finally I find the title of the post and I can click into the next one. Godforbid I ever want to read through a series you wrote over a year ago.
I believe it has to do with the authenticity of SSC. It possessed a certain rawness, call it an early web2 aesthetic. SSC had your finger prints on it and it felt substantially more intimate. It’s cause for nostalgia now because we’ve all been drowning in a sea of refined homogeneity. Refined design (ACX) and raw self expression (SSC) connect with users in two entirely different capacities.
One big thing for me is that SSC has a nice text list of all blog posts, sorted by date and easily searchable. Substack had a dynamic list that loads more as you scroll down, meaning it's hard to find a particular older post if I want to refer to it. I usually just end up using Google if I want to find a particular ACX post.
For me, this is about feeling like you're in 'Slate Star Codex land' vs 'Substack land'. The former is far preferable, as the latter is unfamiliar, but also large and anonymous (I don't feel like I know anyone here; I don't know the 'culture', etc.).
For Facebook vs MySpace, the common layout makes Facebook feel like we are all a bunch of people interacting in Facebook-land, whereas MySpace felt like a bunch of different sites that were loosely connected together. If the goal is interaction with other blogs, maybe Substack is preferable, but for a cohesive readership and community built around the blog, I think a distinct identity is better.
On Substack all content looks the same as all other content on Substack and, indeed, all content elsewhere in the internet. Contrast with the SSC design where you as a repeat reader know right away where you are and what you're doing there. It's homey.
For me one of the major differences is that the old design had some color, and then filled the unused parts of the screen with a nice dull grey. Substack just has white, which is like staring into a lightbulb.
I think this is mostly a 'music was at its peak precisely when I was a teenager' thing. SSC was what what a lot of blogs looked like when your readers were in their formative years – in fact your blog may have got them into reading them, so that's where their idea of what a blog 'should' look like formed.
I'd also probably say that a disproportion of your readers, including myself, read on laptops, while most of the public does things on mobile. Old blog design was laptop-optimised, modern design is mobile-optimised.
The typeface here is gross and the text is (way, way) too large. The comment section code is also pure satanism. That's pretty much all it is for me, anyway.
I dislike the fact all substack pages look the same. I associated the blue with SSC, green with Marginal Revolution, etc. It allowed for some different mental context switching. Now, all substack bloggers blend together.
For me, the biggest thing is how SLOW substack is. I don't want to wait 5 seconds to load an article, and then wait 5 seconds to load comments. When I see a new substack, I don't want to click "let me read it first" then wait 2-3 seconds to find the 'top' tab.
Same reason I prefer hackernews over new.reddit, even though hackernews is purposefully made ugly. It's just quicker. I don't need fluff, I just want it to be responsive.
Your old blog was charming because of your amateurishness, not in spite of it. The fact that you designed it personally means that it had some gestaltic connection to your personality and community. The user feels this connection without being able to explain it. This is the charm of the old web.
I imagine Substack is designed by one or more teams of product managers, designers and engineers who apply established principles and careful experimentation to maximize user satisfaction and engagement. Is it any wonder that we don't vibe to it?
To me Substack feels like a corporate lounge. Sleek, clean and coldly elegant, but not exactly homely.
I don't think that Substack could do it any different. Authenticity is a primitive. You cannot engineer it. Although some have tried (perhaps the history of punk is relevant here?).
> This may be a little too cute, but I can’t help but think of Whither Tartaria? In every art form, complicated colorful designs transition to “modern” minimalist designs over time. Whenever anyone asks, people say they hate the modern minimalist designs and wish they could go back to the complicated colorful ones. But for some reason nobody ever does.
Isn't this what happened to Reddit? For the record, I joined Reddit when the new design was already in place, and I am not able to use the old one as I find it confusing.
I've discovered your blog after you moved to substack. Since then, I've read basically everything you've ever written. As someone who started with your substack stuff, I'm honestly indifferent. I guess I slightly prefer this, but it is a lot more sterile.
One thing I vastly prefer about the old layout is the tag system. Having to trawl through your posts manually to find your fictional short stories really sucks.
Hell. I like your image of the original SSC design better than the 'improved' one. Sidebars are evil, your rendition only had 1. Every time I read ACX my eyes are stabbed by these massive gutters of white on either side of the content. It's wasted space, it's aesthetically ugly, it's hard on the eyes (dark mode at least helps with this one singular complaint), the only purpose it seems to serve is to make the page longer by squishing the text into a newspaper column, and newspaper column text is hard to read.
Obviously this is all my own opinion and I seem to be in the minority if the prevalence of space-wasting empty white boxes in modern applications and sites is a halfway reasonable indicator. But for my part all these 'experienced design veterans' can go stuff themselves someplace ugly. Give me back my information dense interfaces. (though yes, you can keep all the equally eye-gouging flashy distractions)
To me it seems crazy to spend lots of time web surfing on a mobile device because the input mechanisms so far inferior to a desktop that it slows down everything you do and requires all the designers to dumb down their interfaces. If I'm away from home, it's almost always because I'm doing something that is NOT surfing the web. So all these websites are being optimized for a use case that should barely even exist imo.
I have no preference. I spent lots of time over the last seven years reading SSC as ebooks, text files, or Reader mode renderings, or in a heavily customized browser. The comments were akin to marginalia, not a conversation. So I don't really care about how it looks (and reading the comments here, there clearly wasn't a consistent user experience). However, reading ACX in the Substack app is the first time I was motivated to register and post instead of lurking, even while both the old and new comment systems seem worse to me than a threaded NNTP client. Substack reduced the height of the first step on the engagement curve and nudged me to take part. Is this a good thing when there are already hundreds of great contributors here?
I hate the modern bullshit. ALL the modern bullshit. All the stuff that slows things down and eats up screenspace. If I wanted to exaggerate somewhat, I'd say for me the best internet are the "print versions" of websites, and that I wish I could go back to browsing in Lynx. You get the idea.
Like the bar with the name of the blog popping up when I scroll up even a little. Who needs this?
I meant the desktop website. But in your situation I'd just use the "back" button, unless I hadn't originally opened the post from the homepage.
Come to think of it, I imagine if I looked into it, I would find a substantial amount of my peeves with modern websites to come from influence by mobile phones, similarly to consolitis in PC games.
Another theory is it's the distinction. SSC looked like SSC, astral codex 10 looks like every other substack.
I don't know if it's just different things are interesting or if it's a conditioned place preference kind of thing. Where you prefer the look of the website that gave you a decade of mind blowing revelations.
The old design was a product of its times. Sure, You, Scott, were not a web design specialist, but you were not writing the site from scratch, you were just putting together things created by people who were. And those people worked in a different environment and optimized for different things than the modern web designers optimize for.
I think other comments have already conclusively explained all the object-level technical reasons for which modern web design, of which Substack is but one example, is extremely awful. No need to repeat them, so straight off to meta I go.
The process that led us here can be framed in several ways, largely complementary, but emphasizing different aspects of the dynamic. Among them:
1. Desktop vs. mobile. Mobile designs optimize for their small screens and crappy input devices by simplifying user controls. Hence, thin columns, simplictic UI, dynamic loading / infinite scroll, and so on, and so on, dozens of people described it in detail already.
2. Customizability vs. intuitiveness. Or, put differently, optimization for experienced/extensive users vs. optimization for newcomers. Up to a point, newcomers aren't going to care about the loading lag or inability to do a simple text search over the entire page, or other technical problems that only start bothering you once you're a regular. The best option for catering to them is a maximally simplified site that showcases everything that's supposed to draw them in, while leaving aside complex options that might confuse them, no matter how helpful they become for the regular user. Also, there's no need to optimize for speed or responsiveness, because newcomers didn't yet learn enough automatisms to make those lags slow them down. Iterate this race to the bottom over the entire internet, and you get the modern web experience.
3. User experience vs. corporate control. Once, internet services competed for users by providing them with better tools to reach what they, users, wanted to see. Nowadays, they mostly enjoy natural monopolies of scale, so they optimize for forcing their captive users to see what they, corporations, want them to see. This isn't really Substack's goal, but again, their expert designers use paradigms evolved in internet-wide race to the bottom, so here we are.
4. Sophisticated vs. casual users. Internet was once a niche of intellectuals, professionals and hobbyists. Now, it's an everyday necessity aimed at a common person. It's not just that some options are taken away from the user, it's also that a majority of the users will increasingly have no need for them and not demand them back, leaving those who do want them a minority not worth catering to.
My suspicion is that Substack is optimized for two distinct use cases, one of which is e-mail. And it's the e-mail experience that allows frequent users maximum control at the level of their mail client, while the browser interface is basically an advertisement for the core, e-mail product, aimed at the outsiders. Great for people still employing habits formed in the days of usenet (assumedly, someone who actually reads substack newsletters in the mail please confirm), but bad for those of us who gave up on e-mail as an everyday tool and wish for an optimized experience in their browsers, because we're getting a laggy, dumbed-down, hard-to-efficiently-navigate version designed for complete newcomers.
Substack does not index comments. It is impossible to Google "site:astralcodexten.substack.com JohnWittle" to find comment threads I replied to, or whatever, the way that I could on SSC.
Instead, I am forced to use Substack's notification system... which only shows me comments over the last month. Older comments seem to just disappear into the void, unless I want to manually ctrl+f every single article
Frankly this is horrible. I can't count the number of times I wanted to look up an old conversation I had, and it was easy on ssc. On substack so far if it's older than about a month, it's been impossible if I couldn't remember the exact context.
My theory behind why Substack (and other companies) do this is because if everyone made an optimal design there would be no non-textual way to tell websites apart. So by making your design a pain in the ass to use your website becomes more memorable and recognisable and you are able to build brand recognition. This also explains why customisation is not allowed: it subtracts from brand recognition.
Brand recognition is probably very helpful in extracting money from venture capitalists so it doesn't matter if you're pissing off half your users.
Designers are like economists: very handy when it comes to the small stuff but just smile and nod and then do it your own way when it comes to the big stuff. They are taught a lot of theory that changes how they see the world, but they then become very convinced their profession is always right and in the process lose some common sense.
This is really a small thing, but you put a picture next to each post title (presumably because substack forces you to), so I click through because I want to see the picture bigger and in context. But most of the time the picture is not present in the post.
(I feel like this has come up before, but I can't find it. So I'll assume it's new to ACX / SSC.)
The house is markedly different from other houses and is decorated in an idiosyncratic way that I assume pleases the owner. The owner is also mired in legal battles with other somewhat-nearby homeowners who object to the idea of the owner decorating her house in a way that she likes. And while I believe the battle is unusual, the general legal background is not. If people were allowed to decorate their own houses, design their own buildings, etc., they might have better-looking houses, buildings, etc.
Substack has no personality or identity. It’s plain and sparse and feels very generic. Scrolling through it feels like it could just as well be showing me listicles or recipes.
I prefer Substack to Slate Star Codex for reading individual articles. The text is bigger (I have to zoom SSC at 150% to be able to read it easily), it's less "busy" (just the text, no sidebars).
On the other hand, visually I prefer the SSC. It may not be the best design ever, but it's "quiet" (I don't feel "attacked" by the colors), and more importantly, it's yours. When I read a blog article on SSC, I know that I'm on SSC by the writing and the design. When I read a blog article on Substack, you have the exact same design as everyone else. It's like being at an independant coffee shop instead of at a Starbucks. Sure the coffee shop may be a bit messy and not to your taste, but it was made by human beings, it's unique, and it says things about them. It has meaning. Starbucks on the other hand are a product of a company. They have no "soul".
Another thing is that Substack uses the serif "Spectral" font (or a backup serif font). Most people prefer sans serif to serif fonts for screens. Substack may give a better impression of "authority" (serif fonts reminds people of printed stuff, newspaper), but the font on SSC is more welcoming (and easier to read). It also, in my opinion, goes better with your writing.
Probably just change aversion. It's well known that people who have spent a long time using some design will react negatively to any new design, even objectively vastly better ones. Loss of familiarity, loss of instinct-level knowledge one has built for using the site...
Try surveying only the people who started following you in the ACX era, and you'll get a very different answer I'm sure.
The old blog was less easy to scroll through, but practicality isn't everything. This one feels less you, less like reading what a specific person wrote and more like the hundreds of sanitized other blogs we stumble upon all the time. I got used to it because your ideas are interesting and they make up for it, but for me it's like the difference between a cozy living room and a modern office with white everywhere.
Substack on desktop feels like a badly adapted mobile app, it just doesn't use all the real estate a wider screen provides, while the old design did. This is a general trend in webdesign, almost everything is made with mobile in mind so websites are tall rather than wide with lots of empty space on the sides. I know some people think that's a better for aesthetics, but it just make everything feel like twitter/instagram.
There's like... actual utilization of most of the screen on the old layout. The new layout is a center column of text with literally over half the screen space completely empty.
So yeah - I think it's a mobile thing. Websites used to be designed for 4:3 and later 16:9 screens. Now they're designed for 9:16 screens and look stupid on 16:9 screens.
(and I should note I never read your old blog, so I have no nostalgia for its layout - just nostalgia for websites that were actually designed for the screens they're being displayed on)
Is there though? I just went over to slatestarcodex.com, and the central band of text seems to be the same width as on substack. Near the top of the page, there's sidebands with content, but that cuts out pretty quickly.
One of the things that's been learned in web-dev that turned out to be both true and useful was that text areas should not be super-wide. You want it to be easy for eyes to skip to the next line in a paragraph.
But the optimal width is unlikely to be exactly the same for everybody.
The original philosophy of the web was that you'd send basically plain text to the user's browser with only some minimal markup hints, and leave it up to the browser to decide on the details of how to render it. After all, the browser knows more than the server about the local device's capabilities and about a specific user's preferences.
E.g. if the website would not enforce a text width at all, then it would look ugly when viewed in fullscreen, but I could easily choose the optimal width for myself by simply resizing my browser screen.
Unfortunately, we've been steadily moving away from that, and now web designers insist on controlling what their site will look like on the user's device, down to the last pixel. Except that they don't want to actually customise it for every possible target device and every possible user, so you end up with a one-size-fits-all design.
It is still possible for users to override the choices imposed by the server, via custom CSS. Several custom stylesheets to make ACS more usable have been posted in this thread. But that's a hassle, and it needs to be done for every website separately.
Sure - and I suppose beyond the sidebars, the difference is mainly in the background color on SSC not being harsh white across the entire screen. The parts of the screen that aren't being used aren't bright and annoying.
Caveat: I only ever use Substack's UI when I'm trying to debug or fix something. I always read from my own specially-written UI [1] and on desktop.
The #1 problem with the old site was that posting a comment reloaded the page. Substack fixed that. But it's otherwise doing too much . . . *something*, I don't know what, that slows everything to a crawl.
The old site was simple but it functioned. It got stuff done. My desktop wouldn't slow to a crawl because I had 3 tabs open to it.
I would say the old layout was a bit too cramped, cluttered, and quirky. But Substack is a bit too bit, minimal, and bland. I expect there's a happy medium somewhere inbetween.
I've been a web developer for a couple of decades, and it is *very* rare for a redesign not to inspire a backlash among regular users. It's not even rare for someone to be mad at a redesign, get used to it, forget they used to hate it, and then defend the redesign when another redesign comes along.
Obviously, some of this isn't rational, and it's just discomfort with change. But to steel man that a bit, it's often the case that people get good at using really clunky sites/features, and that continuing to use clunky-but-familiar things is, at least in the short-term, less work than acclimating to something better-but-new. It can also be the case that a redesign makes the site MUCH better and more accessible for more people (like with mobile) at a relatively slight cost to existing users who don't care about that.
Primary point being: even if a redesign is obviously better, some people will always be upset. But most of them will get used to it.
You've just made me realize there's a sorta-kinda overlap with this and the "considering the good of future people who will exist" moral calculus that factors into some social and political questions. Web developers have to balance the wants of existing users with the theoretical good of additional users who have/would bounce off the site without certain changes.
> It's not even rare for someone to be mad at a redesign, get used to it, forget they used to hate it, and then defend the redesign when another redesign comes along.
That could also be explained by the second redesign being objectively even worse than the first. Just because users defend version 2 against version 3, doesn't mean they were being irrational/inconsistent when they said they preferred version 1 over version 2. E.g. if users hate stupid unnecessary Javascript crap, then every time you add another piece of stupid unnecessary Javascript crap, your users will now hate it *even more*.
As for "making the site better for mobile users at a small cost to desktop users": that is indeed exactly what a lot of people here are saying, that Substack appears to be optimised for mobile phones and now it sucks for desktop users. The only disagreement is on how "small" that cost is, and how reasonable the trade-off given that possibly the portion of desktop users among SSC readers is larger than elsewhere. A design which is optimal for reading a three-paragraph Facebook post on your phone, may not be optimal for reading Scott's posts, many of which are thousands of words long and require some uninterrupted focused attention to fully digest.
And anyway, it doesn't need to be a tradeoff! Detecting whether the user is on mobile, and serving them a different version of the site if they are, isn't rocket science. Yes, it means a bit more work for the Substack devs, since they now need to maintain two versions of their website. But the question wasn't "which design is most convenient to maintain for the Substack devs" but "what do the users prefer".
> That could also be explained by the second redesign being objectively even worse than the first. Just because users defend version 2 against version 3, doesn't mean they were being irrational/inconsistent when they said they preferred version 1 over version 2.
True, though occasionally I see them defending the very things they said were bad, not just as better, but as *good*, and with no clarification like the above about going from bad to worse. That is a logical possibility based on the vagueness of my original comment, though not something I think I've observed in practice.
To add some detail: I had an argument with a user about a mobile change awhile back and they made demonstrably false (and sometimes even mutually exclusive) complaints about it from post to post. At first I foolishly tried to engage each point, but in retrospect it was obviously just someone trying to say they didn't like it and coming up with reasons post-hoc. I think this happens a lot in some communities, at a level and frequency that might be shocking to a community like this one, which is more analytically rigorous than most. To that point, I run some sites for people who don't know the first thing about development and are not even particularly web savvy (most of them never complain about "JavaScript crap" because they have no idea what JavaScript is). I don't look down on them for this, they have different interests and priorities than I do, but I have to constantly remind myself that they're having a totally different Internet experience than I am.
As for mobile, I don't find it hard (or harder) to read on desktop. Quite the opposite. I'd describe Substack as optimized for reading in basically all formats, but that's just me.
As for managing two separate sites, I'm not sure how viable it is. I did this (and still do, out of tech debt necessity for the time being) for one of my sites and boy, is it inefficient and frustrating sometimes. Like, a lot. I definitely get why responsive design is all the rage. I'll preemptively acknowledge it can be done in a very lazy, bare-bones sort of way that's less about making mobile functional than it is about making desktop excessively plain, though I think even in that form it's a pretty reasonable fit for a platform that's ostensibly just about reading.
I would agree this needs saying, but I also tend to think it gets oversaid by designers -- who too often mistake grumpy acquiescence for acceptance. People are not forgetting they hated the change (or dislike the new), they just don't have enough energy to keep up the fight, they've got other stuff to do (particularly if it seems no one is listening), so they just acquiesce.
If you haven't read it, I highly recommend "The Design of Everyday Things" by Don Norman. He makes a very good case that *outstanding* design doesn't even *have* a learning curve. It's so intuitive new users are just as competent as old hands, and nobody is even much aware of having to learn something. For example, nobody who grew up in England and goes to China for his very first visit out of country and has to ask his hosts how to pick up a Chinese teapot versus the English teapots he's seen before. How you do it is so obvious it not only doesn't need explaining, the user is not even *aware* of figuring it out.
He also of course has a lot of very funny and sobering examples of everyday things that are *not* designed well, e.g. doors that must be conspicuously marked "PUSH" and "PULL" because the design fails to make it obvious which you should do.
Anyway, it persuaded me that in a fair number of cases -- although I would hesitate greatly to say most or all -- observations that of course the user just has to accept some period of clumsy apprenticeship and conscious study to use a modern contraption may just be a load of excuse-making for actually shoddy design.
> it's often the case that people get good at using really clunky sites/features, and that continuing to use clunky-but-familiar things is, at least in the short-term, less work than acclimating to something better-but-new.
I think this coexists with the idea that good design has no learning curve. Even if we posit that there is such a thing as perfect design, and that with it new users would be exactly as competent using it as old ones, I think that holds only when the hypothetical user has not used some version of the design before. Overcoming *habit* is another thing entirely. Reminds me of the "debate" about which way a toilet paper roll goes.
That's perfectly reasonable. My comment is that some people will always dislike the new thing because they are accustomed to the old thing, not that everyone who dislikes the new thing does so for that reason.
I personally like each blog I read having a slightly different feel. Makes it more than just words on a page. So a bit of individualized layout is a plus.
I think the touch of colour from the blue also helped a lot. Imagining the current layout with a bit of the old blue, I'd consider them basically equivalent with that change. White background, black text wears thin fast.
What I don't like in ACX design is how difficult it is to browse old archives. The SCC style, where they're all on the same page, easily searchable and overviewable and archive-diveable, is much better than ACX's infinite scroll that has you start the scrolling down all over again if you lose your place and refresh the page.
The reason old SSC's layout is appealing is the same reason a crowded, boisterous Hong Kong alleyway can be more appealing than a sterile shopping mall.
As counterintuitive as it is, the new substack layout is just a bit... sparse, white, plain. This is usually a good thing for reading, but I think there is a degree of idiosyncrasy people seek out in websites. SSC's layout (I prefer the pre-touch up image, funnily enough!) isn't a tour de force in web design, but it's... idiosyncratic. Recognizable. Perhaps unique?
Going back to the Hong Kong analogy, SSC just had way more going on in the margins: neon signs for tech startups and EA groups, merchants hawking their comments in the open thread, and a blogroll of some of the most revered literati of their small circles posted in the town square like a bulletin. I don't remember all the signs and shops and posters, but they add a bit of flavor and "placeness," as opposed to the blank white margins that greet me on this very page.
My pet peeve with Substack is that the mouse cursor doesn't turn into a text cursor when I mouse over text. I didn't really think about it before, but I must subconsciously point to text with the i-beam cursor as I read!
Ah, this sounds to be the terminology for one of the main things I was thinking about!
Very on-point.
Traditional multi-page apps/sites are inherently better suited to serving content, especially textual content. SPAs only ever make sense for true "apps", which a blog is not.
Yes. And some quantity of designers and implementers have overgeneralized from "this is useful on high interactivity sites" to "this is wonderful; we should do it everywhere". Also, the frameworks they use, not wishing to reinvent the wheel, probably only offer one way to do things now - that being SPA. Because they don't want to maintain 2 ways of doing things either.
The other "benefit" of SPA, as well as the slow load times, is the execrable performance on older computers. If they work at all, that is; they have a bad habit of being dependent on having the latest and greatest version of the browser - or worse yet, require something that not all up to date browsers support at all.
One other thing to consider is that, from the corporate dev's point of view, the site they're working on could become high-interactivity at any time. Being "agile" means being ready for new requirements, so it's easy to see why teams might be tempted to just go SPA from the start, even if the product is, at least for now, just a blog.
This is actually something that substack lets you change.
Scott turned it on once, then back off again a few hours later. I absolutely hated that behavior. I gave me FOMO.
I think the key is mobile - I prefer to read on mobile and therefore highly prefer substack’s app vs. the old blog layout
I agree mobile is key, and I have the opposite set of preferences. I like reading more on desktop and I preferred the old layout.
I am increasingly frustrated by the "mobile first" design paradigm taking over the web. Because it often means "mobile first ... desktop can just use mobile too". The desktop version of substack is not different at all from the mobile version.
I almost only read on mobile and I don't really care about the design change. Both SSC and ACX work and look fine on mobile (as least as of 2016). What does annoy me is the impermanence of the comments on Substack. If I leave the page open on the comments and come back the next day it takes forever to load and it almost never jumps to the place where I left off. This was not a problem with WordPress.
Yes this is my **major** complaint about substack vs WordPress. The load time is longer and that small friction shows up a lot. Impermanent collapsed/ expanded comments is another issue but the bigger one for me in comment load time.
We need some kind of intermediate server that just generates the HTML of the comment section (or a subthread) in the background and people can load that pre-rendered cached version easily.
Does Heroku-node let you do that? I've basically already written from scratch all the Javascript [1] to create a comment page from the Substack JSON so I just need a place to deploy it.
*EDIT* It could technically be pre-built on your local machine in the background. I'm concerned about burning a lot of cycles locally that are never used because that's part of what the Substack UI does badly but I could come up with a decent compromise.
[1] https://github.com/EdwardScizorhands/ACX-simple/blob/main/src/new-ui.js I'm not claiming it's very clean
The worst part: deeply nested comment threads. At some level of nesting, you get the "read more" button.
You can't long-press and open in new tab, because apparently it's not a real link.
But opening in the current task is a disaster, because returning to the main page, everything is messed up:
The page takes up to 20 seconds to render.
Collapsed threads are no longer collapsed.
The scroll position is forgotten, or incorrectly restored.
I do not read deep discussions anymore, because I don't want to deal with this mess.
Feel the same way. This started happening 10 or so years ago I think.
For example, look at the current MLB scoreboard on espn:
https://www.espn.com/mlb/scoreboard
I can see the results of just 2.5 games at a time, but the single column, lots of space, large logos/characters probably makes it easy to scroll on mobile.
Look at the old version. I can see 8 games at once in 2 columns:
https://web.archive.org/web/20080829083105/http://scores.espn.go.com/mlb/scoreboard
So, forgetting for a minute that I wasn't going to be able to see your point because im on my phone, I clicked through to the current MLB site and thought "that's odd, I can see 7 games without even scrolling down." Then I remembered and turned on Desktop View. "Odd, now I can see even more games." This was followed by another forehead slap and me realizing to turn my phone from portrait to landscape. "There it is!" It really is terrible...
Hearing this from someone other than myself feels validating -- although I would phrase it less as "increasingly frustrated" and more as "psychotic rage that can only be ameliorated by taking the last five years of web designers, loading them on a spaceship, and lauching it into the sun." Incoming rant warning.
It's not just the SSC->Substack thing. Web design in general has gotten actively hostile to readability, and I blame a lot of that on mobile. Part of it is the form-factor difference. On a PC, vertical space is priceless and horizontal is cheap -- yet narrow designs with huge fonts, wide line spacing, and extra-tall images have become so prevalent that I feel like I need to override the CSS on every other site I touch. And there are sticky headers everywhere. Whoever invented that abortion of a misfeature is in dire need of a thermite suppository. *There are miles of empty space along the side of the page, use *that* instead of trying to forcibly occupy ~10% of my actual usable space.* Jackass.
All of those -- and there's a much longer list than that -- are things that only make even *bad* sense on mobile, with its flipped aspect ratio. Mobile-only is the modern Web's version of "this site designed for Internet Explorer". Everyone who matters is on a phone; everyone else can go take a long walk off a short pier.
The second problem is that everything wants to be an app, including Substack. So you get sites reinventing, overriding, or otherwise disrupting standard browser features with javascript garbage, constructed in such a way that the now-in-name-only 'website' breaks immediately if you turn it off. Substack *crawls* by comparison to SSC. Thankfully, it only partially breaks in the absence of JS -- comments disappear -- but like many sites, it makes a point of nagging you to turn the shithose back on. No. I turned javascript off for a reason. Whining about it is strong evidence that *you are the reason*. If I trusted you not to defecate on my experience the moment I took your suggestion, it wouldn't be off.
....so yeah. It's not that SSC was a paragon of good design, and it's not about the minimalist look; the look-and-feel isn't the issue. I preferred SSC because Substack is a web application emulating a phone app emulating a browser for the bog-simple, 1990s task of *displaying a post and any responses to it.*
Things like NoScript, uBlock Origin, Stylus, and uMatrix are increasingly non-optional. Modern web design is an act of war on the user. If one would still participate, the browser is a trench that requires defending.
Yes, all of this also.
Yes, all true. My speculation is that if we compare Web programming to Detroit we are in the 1970s chrome & tailfins era -- everyone competes to deliver some new random funky feature yearly -- New in the '74 Olds! Power doorlocks! -- and assorted chromed flash, and we are awaiting the equivalent of Toyota, the economical option that just runs and runs, to arrive on the scene and cause a great reset.
It's strange that blog comments, a thing that was perfected in the 90s by CmdTaco fresh out of college, to run on computers with 10^-5 the power of modern machines, are still clunky and occasionally broken here despite the enormous wages I am sure Substack pays its JS programmers. In their defense, it wouldn't surprise me if they aren't given much liberty to focus on plain reliability and speed, because someone in the front office wants them to drop everything and add some chrome and a tailfin or two.
> blog comments, a thing that was perfected in the 90s by CmdTaco fresh out of college, to run on computers with 10^-5 the power of modern machines, are still clunky and occasionally broken here despite the enormous wages I am sure Substack pays its JS programmers
Those enormous wages are part of the problem, because they contribute to the alienation of the software-developing class from the technological proletariat.
Most of my work involves connecting to various servers remotely, and the speed of the box that my keyboard and screen are physically connected to doesn't matter at all, so a six-year-old laptop that wasn't top the line even when it was new serves me just fine.
But some websites take more than a minute of clunky loading animations to show up on that machine, and I used to wonder how anyone could put up with that. Until I saw a colleague open it on her beefy workstation and it rendered basically instantly.
In the screenshot, Vladimir compared using Substack to wading through water. Now imagine Substack employees having to wade through that same water to get anything done. They wouldn't be able to move fast enough, but still break things. And if highly-paid employees are slowed down, their enormous wages are wasted money for the company. So they get speedboats, and barely even notice that the streets are flooded and children are drowning.
Anyways, I don't care much about the layout, because I just read the RSS feed, which is its own kind of alienation, I guess.
I had a shitty boss who was cheap and said giving us old PCs to code on made sure our software was efficient.
He *was* cheap and shitty. But on that issue, he was right. We fixed performance issues immediately because we couldn't use our software without it.
My fantasy is that every non-AAA-Game developer is forced to develop on a decade old machine.
I'm not convinced the difference in processing power has that much to do with it. Something, maybe, but "don't actively break the user's UI" is not a concept that should have much to do with site performance.
Motherfucking Website and Better Motherfucking Website should be required reading for all web designers.
I've experienced too many times where lack of performance becomes no performance. The UI was written with the implicit (and incorrect) expectation that an operation always completes in a certain amount of time -- which it does, on developer boxes!
Run on a normal machine, those assumptions stop working, a call fails somewhere and the error condition breaks the app until it's restarted.
What do you use to read the RSS feed?
Thunderbird.
I love this comment. Is there some kind of Toyota cult we can join to help summon the Eternal One? I will pull or not pull Andon Cords all day if it will get us closer to clean website design.
> And there are sticky headers everywhere. Whoever invented that abortion of a misfeature is in dire need of a thermite suppository.
It gets much worse. A lot of sites use a header that vanishes when you scroll down and comes back when you scroll up! This is an abomination; it makes it nearly impossible to position particular text at the top of the viewing window, which is what I want to do.
A lot of sites, including this one.
Ugh, yes. The original perpetrator of *that* offensive pessimization should be forced to play a version of transparent Newcombs' where both boxes contain tigers.
Thankfully I don't see that on Substack, but it's possible my array of browser armor is stopping it. When that fails, I have a bookmarklet set up to nuke sticky headers manually. I would link it, but I can't figure out how to do links in ACX comments, and Google is not enlightening me.
Yes! End this abomination now.
I just did a little experiment on google-chrome: if you edit the chrome binary to change the last two occurrences of the words "sticky" and "fixed" to "stciky" and "fxied", the reappearing and floating toolbars on google, substack, and a few others that I've tried so far, all disappear, while still appearing at their proper positions at the top of the actual page.
I'm sure this will break things somewhere, but I'm not sure I care enough; or rather, I prefer it enough that I'm quite happy to maintain a second separate installation and profile for such things.
Virgin script blocker vs chad binary modifier.
I work in web development and perhaps I would find it more rage inducing if my job didn't require me to work with the people you want to launch into the Sun.
I can rant for days about all the shitty ideas web designers come up with. I had to pick my battles though, and substack does at least avoid the one thing I will fight to the death about: modals (or website sanctioned popups for those who don't know the term).
For some reason designers love modals on desktop. But because things had to work on mobile, that modal just becomes a full screen modal. Which makes it look no different than navigating to another page. Sometimes designers even want full screen modals on desktop (WTF!).
Modals of course break all web navigation expectations, because if you click the 'back' button on the browser you aren't going 'back' from the modal, you are going 'back' from the base page. I tell them this, and EVERY TIME they say 'oh well it will be fine they can just click the X to close the modal'. And EVERY TIME once it gets into the hands of users they use the back button like they expect to on nearly all web pages, and then a bug gets opened to "fix the back button on this page".
So then I have to write some shitty javascript to subvert and break basic web navigation, so that when this stupid modal is open and a user clicks the back button it doesn't go back in the browser it just closes the dumb modal.
To top it all off I'm pretty sure that the modal paradigm is so popular because its the only way for designers to actually design a clean page from scratch rather than dealing with the accumulated crap from sidebars, headers, and notifications.
Modals are one of the longer-list items that didn't make it into the rant. They go in the same bucket with sticky headers as unforgivable atrocities. I nuke them on sight. And javascript subverting web navigation is one of the reasons I keep it off by default.
Both are functionally attempts at a hostile takeover of my browser. I think that might be the root conflict in the war -- it's a war for control of my interface in the same sense that modern advertising is a war for control of my attention. Individual actors can improve the situation slightly by providing an API (which substack, I note, does not), but that is of limited use when the API is provider-specific. It potentially allows others to write a non-terrible client, but few are willing to do that if it will only work against a single site.
I blame Apple for modals. I seem to recall they were a (then strange) feature of Apple's early OS UI, different from everybody else, and they were freaking proud of it. I guess the logic was that these ephemeral messages could be displayed, dealt with, and dismissed without messing up the Zen garden elegance of their windowing system. Or maybe they just loved popping open little windows because they could. Who knows? My first exposure to a windowing OS was X, and X didn't encourage these little mushroom-like monstrosities -- that's why you kept a console window open, to collect in one place anything the kernel had to say to you outside of an application. So they still annoy me. Having them within an application (or, worse, website) is even dumber. It tells me you can't be bothered to design a UI which allows the user some routine and intuitive way of knowing your (the application's) status at a glance. . It's like traffic engineers designing stop lights that spring out of the street from randomly placed manholes right into the middle of traffic, whenever conditions suggest some kind of traffic flow control is advisable.
Haha agreed.
+1
At least Substack isn't at the point of https://how-i-experience-web-today.com/ (yet?), but SSC > ACX on Substack because of web bloat. Trying to load ACX on a weak LTE or Wi-Fi connection is awful. Less "modern web" (/JS) bloat on SSC makes the experience much faster, and much better.
The problem isn't what SSC does better. It's what Substack (and most of the "modern web") is doing worse.
That makes a lot of sense. I prefer to respond as well as read, and I never voluntarily produce text on mobile, unless it's as short as a text message or tweet, e.g. "running late, sorry".
So anything that improves mobile at a cost of dis-improving desktop is pure dis-improvement to me.
Mobile is a big reason for design changes. Scroll is messed up on a lot of sites, even desktop articles won't load everything right away.
I find that the substack page lags and glitches terribly when reading the long open threads on mobile. Which is a very difficult thing to do with rendering plain text on a modern device, and implies they are doing something insanely inefficient.
I have the same problem. Long comment threads are basically unreadable and I therefore ignore the open threads, despite previously being a frequent open thread user on the old SSC.
Try viewing it on an old laptop or PC. It is doing a lot of crazy inefficient javascript stuff in the background for sure.
Yes, ACX is unreadable on mobile, and virtually unreadable on desktop without the "ACX Tweaks" extension.
Thanks. This is actually useful.
I was unaware of the extension. Installed. Thanks for mentioning it.
I always read in the browser on mobile and I have no idea why anyone would prefer the old design over the Substack.
+1
There's so much javascript junk on substack (or any site designed after 2010) that I had to disable javascript to be able to read it on my (very old) phone. I didn't install the app because $£#% apps.
The fact that I can even see some lag on a text+img website on my gaming PC tells me that buying a more expensive phone will not solve the problem.
Yep. Modern web sucks in general, and mobile-first ideology is the number one reason.
I can't really add much to this, but it gives me an opportunity to link to one of James Mickens' old USENIX posts, so that is a positive contribution to the world.
A sample from "To Wash it all Away" ( https://scholar.harvard.edu/files/mickens/files/towashitallaway.pdf )
Given the unbearable proliferation of Web standards, and the
comically ill-expressed semantics of those standards, browser
vendors should just give up and tell society to stop asking for
such ridiculous things. However, this opinion is unpopular,
because nobody will watch your TED talk if your sense of
optimism is grounded in reality. I frequently try to explain to
my friends why they should abandon Web pages and exchange
information using sunlight reflected from mirrors, or the
enthusiastic waving of colored flags. My friends inevitably
respond with a spiritually vacant affirmation like, “People
invented flying machines, so we can certainly make a good
browser!” Unfortunately, defining success for a flying machine
is easy (“I’M ME BUT I’M A BIRD”), whereas defining success
for a Web browser involves Cascading Style Sheets, a technology which intrinsically dooms any project to epic failure. For
the uninitiated, Cascading Style Sheets are a cryptic language
developed by the Freemasons to obscure the visual nature
of reality and encourage people to depict things using ASCII
art. Ostensibly, CSS files allow you to separate the definition of your content from the definition of how that content
looks—using CSS, you can specify the layout for your HTML
tags, as well as the fonts and the color schemes used by those
tags. Sadly, the relationship between CSS and HTML is the
same relationship that links the instructions for building your
IKEA bed, and the unassembled, spiteful wooden planks that
purportedly contain latent bed structures. CSS is not so much
a description of what your final page will look like, but rather
a loose, high-level overview of what could happen to your page,
depending on the weather, the stock market, and how long it’s
been since you last spoke to your mother. Like a naïve Dungeon
Master untouched by the sorrow of adulthood, you create
imaginative CSS classes for your <div> tags and your <span>
tags, assigning them strengths and weaknesses, and defining
the roles that they will play in the larger, uplifting narrative of
your HTML. Everything is assembled in its proper place; you
load your page in a browser and prepare yourself for a glorious
victory. However, you quickly discover that your elf tag is overweight. THE ELF CAN NEVER BE OVERWEIGHT. Even
worse, your barbarian tag does not have an oversized hammer
or axe. Without an oversized hammer or axe, YOUR BARBARIAN IS JUST AN ILLITERATE STEROID USER. And then
you look at your wizard tag, and you see that he’s not an old
white man with a flowing beard, but a young black man from
Brooklyn. FOR COMPLEX REASONS THAT ARE ROOTED
IN EUROPEAN COLONIAL NARRATIVES, YOUR WIZARD
MUST BE AN OLD WHITE MAN WITH A FLOWING BEARD,
NOT A BLACK MAN WITH HIPSTER SHOES AND A
FANTASTIC VINYL COLLECTION.
Read it through Feedly, or on an iOS device via Safari reader. I have no problems with it on mobile.
The big problem w/ substack is that it provides lousy mechanisms for completion it’s, or anyone who, for whatever reason, wants to read stuff within a particular date range. A semi-endless scrolling feed, anchored at “now”, and with no quick way to jump large distances or to a particular date, is not an “Archive” regardless of what substack marketing thinks!
I have experienced this myself over 15 years of blogging. And honestly - people always just like the band's old shit, the old sound. It's not more complicated than that.
^probably this^
yup. people just want to show they liked you 'before it was cool'
eh, I don't think I had this take for the livejournal -> wordpress shift.
Yeah, it's just nostalgia talking. Even relatively open-minded people don't like change and remember the past through rose-tinted glasses.
I suspect that's right for the cosmetic aspects. But there are functional differences too, each of which seems small, but all of which are relevant. (Links to posts, links to comments, browsable archive, fully-loaded comments rather than appear-as-you-scroll, etc.)
None of those things seem small. Links to comments especially are important.
A related issue is that when you have a page open for a long time (like to read through lots of long comments) and then page reloads or the browser closes and opens again, it loses your place because the comments are not all expanded.
Right, this annoys me too.
Or you turn night mode on/off. Twice daily.
That wouldn't be enough to change my vote, but the lack of a single-page archive is tickling me the wrong way.
If xkcd.com/archive can link to 2600 posts in a single page with no javascript, surely this isn't beyond the means of a modern blogging platform ?
I guess the problem is that it’s a blogging platform that is designed as though it were a social media platform. For social media, content is treated as largely ephemeral, so that people aren’t looking for a full chronological history of it, but it is instead delivered by algorithm.
Right, I think that's a big part of the problem: Substack's design is modeled on sites like Facebook whose objective is to be maximally addictive, not for thoughts of enduring intellectual value, nor to make it easy to navigate complex information spaces. Systems that do a better job of these (IMHO more worthwhile) objectives include IDEs, paper books, Wikipedia, Jupyter, legal briefs (just check out the table of contents of any Supreme Court amicus brief!), and Excel.
This is great for the low standards of our time.
The archive page still sucks. Should give me columns, so I can just fullscreen on 4k and get max information density.
Not that I have ever seen that anywhere ever, but it's just how you'd organize a list with hundreds to thousands of entires, if you actually care about helping people find things as fast as possible.
Agreed.
Reminds me of the story of eBay changing their background colour from yellow to white.
https://kulor.medium.com/how-ebay-secretly-changed-their-background-colour-from-yellow-to-white-ffd9718e7bb
good post. I didn't know that
Interesting! For those who aren't going to click the link: eBay users revolted when the company changed the background of some pages from yellow (which, seriously?) to white. So eBay changed the color back to yellow, and then slowly faded the yellow over a period of months. This time, no one complained.
Yeah, I feel WordPress nostalgia. Reminds me of when the MR-centered blogosphere was a lively, rude Conversation, when Robin Hanson linked to PUA blogs, when Mencius Moldbug still wore a cape.
Yeah, I think it's mostly this, along with response bias on the survey. Everyone who has ever redesigned a website knows that touching a single pixel is going to elicit a howling response from a very vocal minority. And that minority is going to be extremely eager to register their unhappiness in a survey.
Personally, I think the Substack site looks better. But my main take is...shrug.
I didn't read SSC. I came here after that magazine article business. I like SSC's design better anyway.
Sure, I wouldn't expect 100% of people to prefer the new site if going in blind. The truth is, both designs are pretty vanilla. How strong is your preference?
Fwiw, the old site pretty clearly violates a bunch of basic typesetting and usability principles. The contrast between font and background is often too low. The use of white space and line spacing is inconsistent. It mixes serif and san-serif fonts in a way that I find a bit jarring, although clearly that is delving into the territory of personal preference. The column seems too wide for ideal readability. More than anything else, it's really dated looking, with the embossed font on the blog roll, etc.
I'm not suggesting this proves the new site is better, but Scott is certainly right when he says that the old one is slapdash and amateurish and the new one has benefited from some basic professional attention to detail. Ultimately, they are both columns of text on a page, the differences just aren't all that great, and it is very common for people to reflexively oppose changes in web design for reasons that are somewhat mysterious to me.
"More than anything else, it's really dated looking, with the embossed font on the blog roll, etc."
If there is one concept that should be banished from design circles, it is the concept of "dated looking" as a negative.
Eh, that's a bit simplistic. I agree, design trends are often driven by the boredom of designers seeking novelty, with little or no regard for the value to the consumers of design. On the other hand, web standards and capabilities have evolved radically over the past twenty years and continue to do so. Devices used to access the internet have evolved radically. Also, users themselves have evolved as they become accustomed to different technologies and UI affordances. Whatever its charms, Geocities doesn't cut it today.
Also, aesthetics do matter on the margin. The embossed fonts on the old SSC blog roll surely aren't really hurting anyone, but they are dated and kind of ugly and changing them also doesn't hurt anyone.
So, sure, change for the sake of change alone shouldn't be done heedlessly or on a whim, but the entire concept of something being outdated is hardly invalid.
Things become outdated when standards change. Thus the problem is not something's age, but that it does not live up to present standards.
Also, in terms of taking advantage of the capabilites of computer monitors web design has gone massively backwards.
I think the original claim is that it doesn't make sense to criticize something as aesthetically "outdated" if it functions as well as something newer. Do you have a better response to "what's wrong with the fonts being dated" than "but they *are* dated"?
Yes, see the response I left.
Hear fucking hear!!!
>"[on the old site]... The contrast between font and background is often too low."
Comparing the screenshots of the old versions and the current version as displayed, I think the opposite, for the main article text.
I found a word which is present in this article, and the first screenshot ("thought"), made cropped bitmaps and compared them zoomed in. The bars on the 't's and the top of the 'g' are blurry in the substack article font, while all strokes are evenly dark on the SSC article font.
This is the case if you compare other letter pairs individually - on ASX the lower case 'e' is especially bad, the central line being almost invisible; it's practically a 'c'!
Overall this makes the text look a bit faint, even though it's larger and darker in places.
To be clear, I can read both without too much trouble, but the old site text is definitely easier to read.
The main article text is black on white in both cases, which is the highest possible contrast.
I was talking about the blue text on gray in the blog roll, the white text on light blue in the header, and the gray text on gray in the byline. I haven't run them through a contrast checker, but I have no doubt they will fail standard accessibility guidelines, probably by a lot.
I'm not sure exactly what you are looking at, but you seem to be zooming in on some of the font anti-aliasing. This isn't really a valid thing to do -- most fonts are vectors that are designed to scale up and down infinitely. Converting them to bitmaps destroys this ability, and so of course they will look terrible when blown up.
The readability of a body of text is not determined solely by contrast between background and a solid block of the text colour. My point is that the font affects this, and therefore the font used for body text is very important to the overall impression you get of a primarily text-based website.
I don't particularly care about the header; in everyday use its legibility is practically irrelevant. But since you bring it up I think both the SSC headers are perfectly readable, and more recognisable as others have said.
The reason I was enlarging images to look at the pixels was to study why the font substack uses is less legible at the size it's displayed at on a monitor. I think this is entirely valid, it reveals that some of the strokes are anti-aliased into a light grey smear. The fact that a font could be rendered nicely at a much higher resolution is irrelevant - what matters is what is used when you read the article.
I thought I'd make a demonstration image, which I've put on a free image hosting website: https://ibb.co/GQTcpfg
If you look at this, you can see that for the ACX font, while there are some dark areas, many parts of the letters are much lighter. The entire central stroke of the 'e' is two pixels tall, both a rather light grey. The lighter greys are where it's effectively trying to draw a very thin line, so presumably there the glyph has a line precisely placed along the edge of two rows. Rather unfortunate. At a different point size, that may not be an issue, but perhaps some other stroke would be as bad.
Presumably the SSC font was designed to avoid this problem at this size.
I guess I should say that I couldn't rule out that Scott's computer was just using a better rendering algorithm, but I don't think it's that - the comment text here looks fine, so it's just that font.
This.
Scott could control for this by polling people who never read SSC, only ACX. This is so simple, that I hope he tries it.
I feel like a really big part of it is just the colors. I like the pleasing dark blue banner and darker gray sides of the old website. Now everything is too pale.
It is certainly nice to anchor to a color. I still internally map SlateStarCodex to that particular shade of blue, and Unsong to yellow. It feels pleasant and familiar to see the color where it should be.
I think what I'm feeling is lack of customization and identity -- Gwern's site still feels like its own thing, even though it's black and white.
Agreed on this. As nice as substack is in a lot of ways (like being readable, having good RSS feeds, etc), it feels sterile. Especially because most newsletters default to the black-on-white or the white-on-black. The exception being Free Black Thought that uses white-on-brown.
The colors helped. For me the wide open white space is too... freeing. I like the coziness of edges and frames.
Edges and frames, yes. It's more comfortable to walk through Paris, with its multi-story buildings crowding the street than it is to walk on the sidewalk next to a US strip-mall with an acre of parking lot between you and the nearest building.
Strange juxtaposition. Anyway I'm more disheartened by the acres of vacant compacted dirt with odd sprouts of brittle yellow grass beyond a warped and rusty chain link fence across the street from the strip mall than I am by the parking lot itself, shattered and pale though it may be. I've been in LA enough, it's a sprawling nightmare of strip after strip of urban decay gradually dissolving in the possession of a passive, lazy, incompetent population. But yeah a prefer a nice cozy log cabin over the LA version of freedom. I do love my freedoms though.
Yeah, the white's a bit blinding. And bland.
Information density feels lower on Substack (information per square inch).
I feel like that is everything, not just substack; probably related to optimized for mobile.
Yes. Information density seems to be roughly 60% of what it was in SSC.
Less words fit into one line.
I did a test and printed this post to a pdf, and viewed it in two page view. Was much nicer to read, since I had more info available in single glance. Nicer to skim.
Their design isn't worse than your old layout. If they pushed SSC-style layout to every website, everyone would hate it.
It's like the Hugo and Jekyll custom academic themes everyone is capitulating to (yay capitulation!). The first time you see them, they look nice. Once you see them for the 10,000th time, they look incredibly boring and tiresome and you wish people were still designing their own websites.
Also, the MIRmaids are gone! (And all the other clever and whimsical stuff from the SSC blogroll.) I'm guessing I'm not the only one missing them.
Wouldn't another word for the selection bias you speak of just be nostalgia?
If I ran a salad restaurant, I probably would get a lot of salad fans. If I switched to hamburgers, a lot of my fans would be angry and like the restaurant less. This isn't because of nostalgia, it's just because a salad restaurant probably attracts lots of salad-lovers.
In this case it's still a salad restaurant, the restaurant just did a few renovations, picked a new font for the menu, and slightly changed the name.
(and coicidentally got rid of most of the seating because users prefer to eat on the go these days)
I don't know if relevant as quite subjective, but I would usually not comment/vote and prefer substack to the old website. Maybe silent majority likes it more?
That said I do mainly read posts through the email newsletter so...
Also I don't really read anything else on substack so no way for me to be sick of it
Design-wise I think it's not much different, but on the old site I was never able to comment for some reason. I tried a few times but the comments just never appeared and at some point I gave up trying to figure out why. The Substack reset let me express my views for the first time.
The most powerful websites are disproportionately blue-colored
The old SSC was blue-colored
Actually, the Power Palette research didn't replicate.
What's that? I just see blue
There are a number of people, myself included, who believe that the user created, original and homegrown, flashy and weird and geocities-esque websites of the web 1.0 era were strictly better than the modern flat and minimalist, muted colors, tons of javascript, everything-looks-the-same websites of the web 2.0 era. I'll take https://squeedge.neocities.org/ over other websites any day of the week.
Also, in terms of functionality, the thread system of SSC was objectively better than here. It notified you about new posts in a easy and comprehensive manor, you could jump straight to them, and there were a few other nice QOL things.
Wholeheartedly agree with this—being back “bad” design!!
Agreed - I think "good design" is actually worse than "bad design". It's like a house that's too clean - we usually think "clean" is better than "dirty", but if a house is too clean, it feels offputtingly sterile, and all the things you want to *use* are hidden away in cupboards and closets and drawers.
A smidgen of mess helps make a house feel like a home; a smidgen of "bad design" helps a blog or forum feel like a community.
Look up what Kim Kardashian's house looks like. It's AWFUL, and most of the web is the equivalent that.
Your first paragraph reminds me of the whole legibility thing applied to city design, as in the SSC book review https://slatestarcodex.com/2017/03/16/book-review-seeing-like-a-state/
"Natural organically-evolved cities tend to be densely-packed mixtures of dark alleys, tiny shops, and overcrowded streets. Modern scientific rationalists came up with a better idea: an evenly-spaced rectangular grid of identical giant Brutalist apartment buildings separated by wide boulevards, with everything separated into carefully-zoned districts. Yet for some reason, whenever these new rational cities were built, people hated them and did everything they could to move out into more organic suburbs. And again, for some reason the urban planners got promoted, became famous, and spread their destructive techniques around the world."
(The 'for some reason' is legibility.)
Quoting James Scott on Brasilia's lack of street corners in "Seeing Like a State":
"Most of those who have moved to Brasilia from other cities are amazed to discover “that it is a city without crowds.” People complain that Brasilia lacks the bustle of street life, that it has none of the busy street corners and long stretches of storefront facades that animate a sidewalk for pedestrians. For them, it is almost as if the founders of Brasilia, rather than having planned a city, have actually planned to prevent a city. The most common way they put it is to say that Brasilia “lacks street corners,”by which they mean that it lacks the complex intersections of dense neighborhoods comprising residences and public cafes and restaurants with places for leisure, work, and shopping.
While Brasilia provides well for some human needs, the functional separation of work from residence and of both from commerce and entertainment, the great voids between superquadra, and a road system devoted exclusively to motorized traffic make the disappearance of the street corner a foregone conclusion. The plan did eliminate traffic jams; it also eliminated the welcome and familiar pedestrian jams that one of Holston’s informants called “the point of social conviviality"
Indeed, I think "Seeing like a state" is more relevant here than "Whither Tartaria".
Back when every amateur blogger built their own website in Wordpress or even in raw handwritten HTML, those websites had *personality*, warts and all. But Substack is designed for scalability, which is the Internet version of what SLAS calls legibility.
Substack isn't a state, but Substack together with the other big tech platforms like Facebook and Reddit may as well be considered a state government. And they don't want quirky little street corners, they want evenly-spaced rectangular grids that are easy to administer centrally.
In addition, just to add my voice to everybody else: yes, over-use of Javascript sucks and reduces my six-core 3500MHz desktop PC to something that feels more sluggish and glitchy than the software I ran on my single-core 7.2MHz Amiga in 1985. It's amazing that "just put all the text on a single page so that searching with Ctrl-F works as it should" is apparently a complicated technology nowadays, when the simplest website built by a teenager in 1995 who just learned what HTML is, would *do that automatically without needing to do anything special to enable it*.
This, fuck web designers.
Yes. Here's the problem with web designers. They have to use the tools/features that look best on their resume, not the tools/features that give the best customer experience.
Blame web designers less (they're still at fault) and blame the corporations and community that releases framework after framework and fashion after fashion, and the hiring/HR who get glowy eyes when they see $SHINY_TECH on the applicant's CV. Web designers are just responding to incentives the mOvE-fAsT cult sets up.
> It notified you about new posts in a easy and comprehensive manor, you could jump straight to them, and there were a few other nice QOL things.
By "new posts" do you mean "new comments"? The "new comment" logic was a client-side JS thing which was originally a browser extension I wrote, which I could in principle re-implement for Substack. Or we could annoy the Substack devs into implementing it themselves, possibly.
yes! and that was you? Thank you very much, that aspect of the comments was super super useful and greatly appreciated. Yeah, an equivalent for substack would be fantastic and really helpful.
Yup. I'll see about a browser extension when I get some spare time. It's probably only going to work here because this is the only blog on Substack which loads all comments under every post. (They hardcode this domain in their JavaScript, fun fact.)
If you're interested, I've already created an extension (https://github.com/Pycea/ACX-tweaks) that can mark new comments, as well as a few other features like restore the old theme.
Oh nice! Glad I saw this comment before publishing my own; better that there be only a single place for it. You should post this somewhere more visible (like the next open thread, or as the top level comment on this post), and maybe Scott will signal-boost in some open thread.
Nice, thank you! I did have to use the "load pages dynamically" option to keep it from refreshing constantly, so that might be a good thing to enable by default.
I've been using your extension and like it very much, thanx. Question - I'd like it if the "click on the vertical line to collapse a thread" worked in your extension. Is that possible to do? Or maybe I need to get an updated version?
It should already work. Can you tell me what version you're on (even if you do end up updating)?
And thank you for this!
(!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!)
Thanks for making it! That was really helpful and it is amazing that Substack has piles of JS that lags on fast computer but is missing such basic functionality.
The 'new comment' behavior made following long-running conversations a lot more practical, I think that would be a huge improvement if it were available on Substack. Thanks for putting the original together!
> Also, in terms of functionality, the thread system of SSC was objectively better than here. It notified you about new posts in a easy and comprehensive manor, you could jump straight to them, and there were a few other nice QOL things.
Maybe there was some merit on the aesthetics of the old site but, at least on mobile, the comment section was literally unusable. Long threads could get to almost single-character widths (https://imgur.com/F7XrMK4) with no way of reading just from a certain point of the tree. On substack I can at least look at a subset of the comments, and collapse the entire thread from anywhere in it.
I think most people feeling nostalgic for the old site are selectively forgetting that detail of the design.
I can't speak for everyone but while I would read Scott's posts on mobile I didn't engage in the comments on mobile or really care to (regardless of whether the UI supported it). On the rare occasion I did, turning the phone sideways made the last most indented column readable (even if it was still too narrow).
I certainly get that Substack would be an improvement for people who primarily browse on a mobile device, although I think you could capture the SSC style in a more mobile friendly structure.
I never browsed SSC on mobile and I think the most indented column was occasionally unreadable for me.
I think more generally the distinction not desktop versus mobile, but that really bare-bones minimal HTML is great for content that doesn't require interaction. If you want more functionality for writing comments, or even if you just want to make it easier for the blog author to write posts, then you have to move away from the minimal version of the website, which makes passive reading worse.
This is a good time to mention that threading in general is not great for cases like deep discussions. It's also very frustrating when you post something, have 3 people reply to you. Are you supposed to reply to each of them individually? Reply to yourself?
That's one of the good things about 4chan-style threads: you can reply to as much posts as you want, and you're not constrained by the "tree" structure, but are free to build your graph. Graphs are often better at modeling things, but harder to represent, while trees get you close of what you need. Something like a graph with the article at the center and comments linking to each others could be an interesting experiment.
Well I think that looks hideous so “strictly better” may be a step too far.
Geocities and blogs are still web2.0, web2.0 is any user-generated content. Web1.0 is the era of html pages that could only be changed by their server.
You just hate giant-boring-corporations-dominated web2.0, which is understandable.
Substack is a giant ball of single-page-appiness, with all of the glitchy microaggressions that product managers think users love that that implies.
SSC loads, and then it's done. I can scroll around the page, find-in-page (or, google!) anything, and it all works instantly and consistently. I will gladly pay the price in lack of gloss and “user-friendly enhancement” for that alone.
Now, get off my lawn.
This is the issue for me.
I like the new aesthetic, but all the crappy collapsing of threads, where I can’t even open that collapsed thread in a new tab (stuck behind javascript? So annoying), which prevents me from full text searching, and frequently just hangs or resets to the top of the very-long-page is incredibly frustrating. So much so that I just don’t bother reading the comments much anymore.
Exactly (and there are several other issues of a similar nature).
The styling is one thing, but the machinery underneath amplifies the “meh, I kinda liked the old better” into “god I hate this crap”, which is reinforced every time you visit another substack-based site.
If every elevator button that was round gave a small shock when you pressed it, people would quickly come to hate round buttons, regardless of where they found them.
Yeah, being able to open a subset in a new tab is pretty basic, to anyone who is *not* a professional UI designer, or limiting themselves to microscopic (mobile) screens.
Seconding this
Thirding this
One specific case in point to illustrate this: on Substack, at least in a desktop web browser, there's a header bar above the article showing the blog name and the user avatar and so forth. Except it's not *really* above the article, because it's a "smart" Javascript control that jerkily hides itself as you scroll down. But then if you scroll back up at all, the header peeks out again, covering whatever it was I was trying to scroll back up to see (usually the last few lines of text) unless I scroll even further to get it out of the way again.
As some of the other commenters have also noted, the font size is quite a bit larger, so the information density is lower than on SSC, and the switch from sans serif to serif doesn't work as well on the screen as it might in print. And not being able to see all nested comments at once makes following the various discussions harder.
None of these are a big problem individually, but all together it adds up to a user experience that feels like the page is fighting us. There was none of this artificial stupidity with SSC – it was just a flat page that loaded and then got out of the way.
That top bar also really annoys me. It's fine when you're scrolling up fast but gets in the way when you just want to scroll a few lines.
That's it for me. Substack brings in a pile of JavaScript and won't work without it. Wordpress I can read and even comment in w3m.
(I like reading as many things in w3m as possible, since it does me the immeasurable favor of throwing away all web design and presenting everything the same way.)
Does w3m allow *any* JavaScript? Would a minimalist-JS approach work for it?
It just doesn't support JavaScript. It works fine if the site supports graceful degradation. (i.e. there's a usable document and the site uses JavaScript to add functionality on top of that.)
Exactly! If some idiot would somehow make me imperator of world then after dealing with important stuff (lock up Putin and so on) I would ban making SPA.
Same here. On SSC, I could go to the archive, ctrl+f "book review" and go through all the book reviews. On ACX, well, try it and see what happens.
Also, it was fun to occasionally scroll through the archive page and look for old posts I might want to read again, but if you try this on substack it loads ~5 new posts and then you have to scroll down and wait a few seconds again to continue scrolling. And each entry takes more than an inch on the page instead of taking only a single line of text.
FWIW it's possible on Substack, but the within-publication-search button is super hidden. I didn't know we had it for like 6 months of working here https://astralcodexten.substack.com/?sort=search&search=book%20review
Making a throwaway to second this. It's the Javascript for me. The moving header bar, literally everything about the page load... Sometimes on my phone I come back to a ACX tab to continue reading and everything below the scroll position is blank, so I have to do a full reload. The back button doesn't work reliably. Increasing the font size doesn't work on mobile. I'm sure there is more. I don't know why tech people insist on breaking perfectly fine websites with this stuff. Just send the HTML to my browser, it will be fine.
In addition I hate the profile pics in the comments and all the rest of the social media stuff. I actively don't want to know how many likes things have.
One other thing that ACX doesn't seem to suffer from, but most other substacks do, is the newsletter form in the middle of an article. It feels like the end of the post every time. It makes no sense, put it at the end.
Obviously the old blog had its flaws too. As I recall it wasn't responsive, so it was sort of a pain on mobile. These things are irrationally more forgivable when it's someones humble personal corner of the internet vs. a large platform.
Speaking of glitchy microaggressions... it allows people like you to have animated avatars, which create constant movement which distract from reading :)
Wow, that's horrifying. I guess we should be glad that not many people have discovered that "feature".
Same. Yes you can technically scroll and ctrl+f on Substack as well, but it can take 10+ seconds to load later parts of the page (especially on mobile) whereas searching/scrolling the old page was basically instant once it loaded.
I think this is a symptom of modern web dev being horrible at every opportunity. Modern websites have tons of Javascript and they are trying to be artful and fancy. This is the trend among designers, but real people just want something simple and straightforward that works.
Designers are, indeed, not people.
Artificial intelligence for the lose?
> If they forced everyone into the standard non-customizable layout of 2015 SSC, would that be a straight utility gain?
No, because a large part of why people like the 2015 SSC layout is because it /felt like SSC/.
Substack, like almost all technology companies, is trying to make their design work for everyone. Every author, every audience. This inevitably results in a design that expresses no personality and has no distinguishing features whatsoever. That is what flat design is, that is what flat design is for, and that is also why people hate flat design.
Right. Right now the blog looks like any other generic Substack blog. For those of us that read multiple Substack blogs, they tend to blend together. Sometimes standardization between people works well - like Facebook, it's easy to find what you're looking for on someone's profile. But that leaves it impossible to create a specific aesthetic for your blog. People might not be missing any particular element of the old design, just that it was specific to SSC.
I prefer the substack design, but liked the writing better on the old blog. I also think you nailed why in the Why Do I Suck post: you were grappling ~7 years ago with stuff I'm grappling with now, and it's well-trodden ground for you at this point that you don't need to rehash but to me it's front-of-mind.
Your old site was uniquely yours, which people like to say they value. People like to associate with unique aesthetics and your old site was definitely memorable—if every Substack used your old site's design, it would perform horribly. It's uniqueness was the main draw, even though the design was worse on its own merits.
I think if you were to look at revealed preferences in terms of engagement, Substack will actually perform better. It's clearly more designed/optimized than your old site.
If you thought there was a replication crisis in academia, try looking at the internal engagement studies that companies use to justify their design decisions to themselves. Utter pseudoscience!
I’m mot convinced that “Every blog has the same design” would actually outperform “Every blog except ACX has the same design and ACX has the old design”. I suspect that a universe where every blog has a unique and well designed theme is more appealing to users than every blog having a shared minimalistic theme. But Substack can’t offer a unique theme to every blog it hosts, so we’re stuck with the minimalism.
> But Substack can’t offer a unique theme to every blog it hosts, so we’re stuck with the minimalism.
They could offer the ability to customise the style of your blog, they just don't.
> But Substack can’t offer a unique theme to every blog it hosts, so we’re stuck with the minimalism.
...that would be incredibly easy for Substack to do.
The problem with "Every blog has it's own design" is that there are lots of basic functions that it's easy to glitch while implementing. This, however, is a lot different from "Every blog should have the same design".
>Your old site was uniquely yours
is key, I think. When I first encountered SSC, I thought "This guy writes great, but his website looks like shit. He hasn't even bothered to put a picture in the header." But after a while that "no visual aesthetic is my aesthetic" grew on me and seemed kind of punk. I connected that punkness with the writing. But then you went and sold out to Big Substack... and now the lack of visual aesthetic seems corporate instead of punk.
Yes. SSC looked like "Scott's blog"; acx looks like "another substack".
Related to https://astralcodexten.substack.com/p/whither-tartaria, perhaps? I find modern web design sterile, in a similar way to modern architecture, and (mildly) preferred the SSC website for some of the reasons -- it had more character.
I like the old SSC because (apparently counter-intuitively) I think *it* is the more minimalistic one.
Substack has a superficially minimal aesthetic, but that's just the outer veneer. It's like a Tesla vs a Lexus or a Macbook vs a Thinkpad. Their formers *look* sleek and minimal, but that's just on the surface.
Ditto, especially adding the other comments' inability to browse the archives. Substack looks minimal but the UX feels a little like a maze. The old blog had more information density but you could find what you were looking for so easily.
I'll also add that the old blog's design really stands out to me among Wordpress themes. I would not mind as much on theme if certain other blogs moved to Substack.
Yes, there's the Myspace over-customization problem, but if a theme is *good*, unique themes really give blogs a nice character.
One thing I hate hate hate about Substack is how hard it is to browse the archives. SSC made that really easy and really enjoyable. Substack, please just give a list of Titles and dates as links!
The old archive page was superior in every way. Tartarian blogging technology was obviously vastly more advanced than our own.
I was a relative latecomer to SSC and relied heavily on the archives. I think it would be difficult for me to explore the Substack content in the same way. I wouldn't have thought of this on my own but this is a great point.
Yes. I also have that problem.
A few weeks ago, I realized I can use datasecretslox' "Astral Codex Talk" subforum for a more streamlined index and then jump to the post I want on Substack via the thread for it:
https://www.datasecretslox.com/index.php/board,12.0.html
(Now, I'm going to attempt to read your mind and guess you're saying, "But I don't -want- to go use a second site to read ACX archives!" It's helpful for me; not something everyone would want to do. The future of the internet is in non-sucky, non-distracting indexing!)
God, this so much. I loved browsing through the old archive. The substack one is a infinite-loading disaster that takes literally forever to scroll through. It's just a list of links! How hard can it be!
Also, in trying to reply here, for whatever reason it forced me to re-log-in, so of course I had to refresh the comment section, which takes bizarrely long, and look for this one again, because the direct comment links are terrible. Thanks pile-of-js people.
The inconvenience of using the Substack archive does seem like a considerable handicap on the staying power of ACX posts. It would require a bit of extra work, but maybe this could be addressed by adding links to each new ACX post on the old SSC archive page.
Seconded!
It has to be said that Substack is easier for me to read on mobile than SSC. But there was a long period in the industry, when mobile first came around, that companies produced very distinctly designed layouts for mobile and for desktop. Substack is shooting for a one-size-fits-all design -- which of course focuses on accomodating mobile, and thus fits *only* mobile.
First of all, they should drop the single-design principle. And then they should really upgrade the archive browsing functionality on the desktop version, at least to the level present on SSC.
The blog roll also. The old site had a LOT of useful features that were not portable to substack.
The archives are the only thing I really miss about the old site, but that alone is a sufficient downgrade to make me prefer the old site.
Agreed. Scrolling it is a series of 1-5 second delays that just feels terrible. Plus, the tags are gone! So I can't even binge-read the "things I will regret writing" tag anymore!
The comment you posted of someone speaking of their preference for the old design explicitly stated that what they liked about it was its custom, non-professional feel. Professional designers hired by a company to think hard about what constitutes good design and then implement that for everyone basically definitionally can't come up with a design that feels homey and unique for individual blogs hosted by that company.
https://astralcodexten.substack.com/p/why-do-people-prefer-my-old-blogs/comment/6400024 <- Freddy deBoer mentions people liking "the band's old shit" as a constant of human nature - but it's worth asking *why* that's the case. Maybe idiosyncratic associations that people make with an old art style, whether that's a band's first album or your hastily-designed 2013 blog, are still intrinsically aesthetically pleasing to people who encountered that art at the right time, and there's no way for more professional artists/designers to actually create better aesthetics for that group of people.
Yes, but professional designers also think hard about what needs to go on their resume, which leads to what to throw into their designs. Professional designers have long ago figured out that what to put on their quarterly MBOs (Management By Objective) also guides what tools and features they throw into web pages.
Thus, what professional designers need to do to satisfy their career goals may not exactly coincide with what provides a great user experience.
I think a major problem of "professional designers" is they design for an abstraction, the hypothetical universal user persona drinking from the universal content firehose. So they completely fail to take into account what is unique and different about *this* content and provider and *this* audience. Which leads to frustration, not surprisingly, as you try to hammer a thousand different polygonal pegs into the same damn round hole.
It's sort of mystifying why this is considered good practice. It's as if Ford, BMW and Tesla all hired a separate class of "professional designers" to design their cars -- none of whom had *actually driven a car* but only studied carefully the general theory of controlling mechanical devices, supposedly applicable to everything from pencil sharpeners to aircraft carrier arresting gear -- and everybody strived to design just one ideal car that everyone was expected to buy, in one standard color.
Mature industries long ago abandoned this idea. Car companies routinely make 10,000 cars a month, hardly one of which is identical to any other, in a wide variety of sizes, feature sets, and looks, knowing that having each individual customer be able to come quite close to his peculiar ideal is the way to thrive. They pour gigadollars into figuring out ways to become *less* standardized in the salesroom while being more standardized on the factory floor for efficiency. One imagines Web programming will get there, some day. But in the meantime we have to put up with smug statements of design best practices that were rejected as naive and unrealistic decades ago by mature industries.
>worth asking *why* that's the case
Well, as far as creative output is concerned, I'm pretty sure that it has to do with artists using up their best novel ideas early on, and then either doing inferior derivations, or chasing popular trends.
I think other obvious "psychological" explanation besides selection bias is that your old readers are just "used" to it, and they liked it because it was familiar and associated it with the Good Old Days of your blog. Your old website and its old design are still around, and your new readers have probably gone to it to read some of your most popular old posts. Do they prefer that design? I doubt it.
I read your old blog on my RSS reader, and I read your new one in my email inbox. I was exposed to your old layout only when revisiting old pieces, and I always sort of regretted having to do it (no offense). If the article was very long I would usually try to find some way to avoid looking directly at the WordPress layout. Sometimes I would read using a browser extension to isolate the article text (Firefox also supports this natively), or sometimes I would send the article to my Kindle. So I have no nostalgia for the old layout, and this is also the first time I have ever clicked through to look at an article on Substack's own platform.
Agree. Anyway you could survey for this, Scott? I read SCC first and prefer it. I read SCC first and prefer ACX. I read ACX first and prefer it. etc...
Agreed. Isn't this just a live fire example of selection bias and the endowment effect?
I, old blog reader, got used to the old blog (I've been endowed with the old format), and you change it on me, and I just liked it better the first way I had it. And then when you ask me how I feel, I feel motivated to tell you that.
Other dude, new reader, never really interacted with the old one, and couldn't be bothered filling out the survey about the whole old vs new thing, because I didn't feel strongly about it. I mean, it's just a blog format, right?
Having a unique design, good or bad, makes you distinctive, and after enough time readers develop strong and warm associations between the design and your with your voice. Sometimes on Substack I forget who I’m reading.
Yeah, I feel like Matt should be perceptive enough to realize this.
Tempted to switch my profile pic to “exploding lemming” in honour of this.
Conflate this with reddit's redesign. An overwhelming majority of people I've interacted with prefer the old design, yet the company insists that the new one is more liked.
I guess it's probably simply just that the kind of people I interact with tend to like a different kind of thing than the majority. Selecting for content does introduce a lot of preferences
reddit's redesign introduces infinite scroll which make things trickier (especially in terms of measuring which is "more liked"; that could really come from users getting manipulated into scrolling "just a little more").
On the other hand, I use old reddit mostly because it allows me to use RES, an extension which among other features includes its own infinite scroll.
I prefer substack but I very very strongly prefer the old reddit.
> yet the company insists that the new one is more liked
They lie. Mobile website in particular is persistent harassment trying to force you to download app.
I guess I can be thankful, I never waste time on Reddit on mobile.
the font choice is bad. i prefer reading substack comments to articles because the font display looks so bad on desktop.
Substack wastes half of the screen and feels like it was designed for mobile or email rather than desktop viewing, probably because it was. Wordpress is built for browser. Substack doesn't allow pinch & zoom within mobile browser, presumably to force me to download the App I refuse to download on principle. Wordpress just works. Substack has a flimsy minimalist aesthetic, especially in the comments section, while Wordpress has beautiful comment boxes, obviously indented reply trees, and is overall happy to use the space it has available. Substack is all white while Wordpress has grey accents, colored comments. Substack defaults to a smaller font, especially for metadata like commenter names, comment times etc., but then DOESN'T USE THE SPACE SAVED AT ALL.
At a more abstract/kvetchy level I feel that the migration of hobbies, blogs etc. from wordpress, ipboard style individual websites to bland subdomains where the creator is just another user, and not the host, is emblematic of the fall of the old internet which I miss dearly. I think it's fair to say that your readership is probably more on the MySpace end of the scale than the Facebook one.
I can pinch & zoom just fine in Mobile Safari, for whatever that's worth. But the Substack app actually doesn't support it. Or text resizing, or search in the comments, or collapsing threads(!). Its only advantage over the web interface is that it renders threads fully, without the disruptive "Expand full comment" and "Continue Thread" functions which were inexplicably added not long ago. Otherwise, the app is a featureless placeholder.
The horizontal space thing occurs because (according to experts) columns are easier to read. Apparently we have trouble eye tracking when moving from the end of a very long line to the start of another, hence why newspapers use columnar text even though they don't have to. This is now conventional wisdom in web design, hence why we get ever wider screens but text remains stubbornly in the centre surrounded by white.
It's probably right though. I can imagine that reading lines as wide as my screen would be very annoying. HTML was designed in a time when the rule was one page = one window, so you could easily resize windows to decide how much horizontal space they should take up. After tabs and popup blockers made every web page a full screen experience, browsers never adapted and there's now no standard way to express how wide you want the content to be except by zooming.
Your right but ssc was also able to present itself as a column in a way I much preferred
The previous blog used columns too – just wider ones. I find reading newspapers (or did find them) to be tedious because the columns were too _narrow_; too much jumping to the next line.
If you check your browser's accessibility settings, there might be an option to enable pinch & zoom even on pages that don't want it.
If not in the accessibility settings, possibly in the experimental features.
My explanation is that your blog caters to a subset of the population that prefers information-heavy presentation with options. I can read Chinese and love Chinese news homepages. I can scan 100 headlines in 30 seconds. My blog is like your old one, three columns of info. Someone can visit and hit a whole bunch of links off the homepage. Saves a lot of time.
Most websites trend towards mass audiences if they grow and that means a lower general intelligence. I like Yahoo Finance from around 2000. It was data heavy, very customizable. Now mostly videos, junk articles, almost never visit. It feels like its made for stupid people. Don't go on Drudge much anymore, but that site stuck with all headlines.
"My explanation is that your blog caters to a subset of the population that prefers information-heavy presentation with options."
That is true for this blog, sure. But almost equally true for Substack in general. Substack subscribers literally pay money to get information. And I see no reason to optimize for non-subscribers.
Also why would the lowest decile in IQ of Substack subscribers prefer lower info density in particular?
Not being high IQ doesn't imply you don't want to find stuff fast, as well.
I don't see how seeing more stuff at once would be more mentally taxing.
Quite the opposite.
I really love your old style *because* it feels homemade. It gives off a slightly quirky, small-creator vibe, and is very *you*.
This is extremely common in UI design. Normally though what happens is after a period of time, people adapt, and then dislike the old design and not the new one. Your audience may be one where they enforce a logical consistency on their views, so repeat what they said before to prove they weren't wrong. Normal people just forget their old views and think their current views were the ones they always had. I suppose a survey of some sort could look for this correlation, so this is just speculative as to why people haven't come around.
I don't disagree that people _do_ adapt, but I think there simply is 'too much' design – for the existing users of anything.
Professionals are – basically all the time – _furious_ whenever the UI of their professional software changes. Even if the new UI is better (and it often isn't), the cost of relearning one's workflow often, drastically, swamps any meager gains from a better UI (when that is the case).
An anecdote that I consider particularly strong evidence for this is that _designers_ themselves were finally subjected to this same treatment somewhat recently when Adobe significantly changed the UI of their tools. A designer I know admitted that they found the new UI to be a big pain! I pointed out that they finally knew how the rest of felt!
I want the current "design/style" with the snappiness and ergonomics of the old one. Substack does not load enough comments (hides too many things by default), and it does too many background requests. Oh, and it's also noticeably slow.
You should ask the people who have read *only* ACX which design they prefer!
That works for the design, but not for the details of the commenting functionality.
I assume there was something people liked about the older, more personalised design. Perhaps on an objective level the Substack Theme is just better, it obviously has a minimalist elegance, but on a subjective level the old SSC theme probably felt more personalised? That is, crude and primitive web design has a charm that quality lacks?
aesthetics vs usability/ergonomics
I think it's (1) the excessive minimalism of solid white margins, plus (2) the shortened line length. Both hit squarely on the professional-typographer vs sophisticated-reader mismatch. The typographers are very worried about people getting distracted by stuff in the margins, and have eye-tracking data from first-time unfamiliar readers to back it up. They *don't* have longer-term data, which mostly points in the opposite direction. And similarly, there's a widespread piece of ideology which says that shorter line-lengths are better, which I think originates in research done on less-sophisticated readers with a different saccade pattern.
The professional typographers have been right about a great many things for centuries, long before the data eventually vindicated them. I say consensuses among typographers are way more reliable than data. Every controversy between typographers and other people I’ve investigated was just the typographers having thought about the problem more carefully and people not bothering to read them.
I don’t know about this “longer-term data” that supposedly “mostly points in the opposite direction” which you haven’t linked to but I am extremely skeptical. And talk of “sophisticated readers” usually comes from the same kind of sentiment that leads people to think that tiny font sizes and bitmap fonts are so great, that antialiasing is so bad, etc., which, no, is not sophistication but ignorance.
I've been a bit interested in typography for awhile but I think it's pretty rare for most designs to follow any of its best practices. I think the original commenter is wrong that the divide is "professional-typographer vs sophisticated-reader". I think it's more 'web design fads' versus typographer-and-readers myself.
Yes, readers overall consistently dislike long line lengths. And that’s a good reason to avoid them no matter their impact on readability. Publishers figured that part out since they have an incentive to sell their material. But there’s a subset of readers who consider themselves more sophisticated and disagree.
Web design often gets important things wrong. I don’t know why and don’t think it’s just fads. People think they need to make text color lighter to avoid eye strain; instead they should make the page background darker and leave text at #000. Text is too light, too small and contrast is too low.
I think we might be writing past each other. I don't disagree that _too_ long line lengths are bad or disliked. I'm claiming that some line lengths are also too _short_.
I think web design gets important things wrong because fads are necessary for 'full designer employment'.
I also think that 'fresh' designs are useful for sites/apps (and their companies or organizations) but that's because they're a (relatively) costly signal to _new_ customers/users.
If web designers stuck to the existing solutions for "important" things, most sites would look much more similar and there would be much less work for designers overall.
There's already a plethora of tools for 'optimal' color palettes for 'pages' and charts/graphs/visualizations – and it's not just web designers that are mostly ignoring them. There's lots of bad (more) 'traditional' publishing/design products/artifacts too, e.g. academic papers.
I think the 'typographic evidence' actually supports the old blog design being better – too short of a line length is bad too. And the 'typographic landscape' is pretty varied too. _Some_ academic articles are published in multiple columns. I think a good bit of mathematics papers tho are often a single column. SSC/ACX sometimes 'reads like' the latter more than the former, so the 'optimal line length' for this blog might be different than others.
Multiple columns are mostly a consequence of a poor choice of paper size because large margins are problematic (aesthetically and for zoom in a PDF document). The recommendation of 66 characters is too low both for print and for screen. Professional typographers have been using lines longer than 66 characters for a long time however.
On Slate Star Codex I see lines with 130 characters. That is too long. On Substack, I see 80 characters, which is slightly too short.
When you are done reading one line, your eye travels near the beginning of the next line. For reading speed, there are two problems if the lines are too long: you may locate the wrong line or you may not reach sufficiently near the beginning of the next line. The first case is one reason you want a greater line height when lines are longer. In the second case, this is only a small loss of efficiency because you can still process the words in the part of the line that your eye has reached before it jumps to the beginning and you will not need to spend as much time on them afterwards.
I think Substack's line length is fine – but I also didn't have much of a problem with SSC. I think the 'optimal' length is a little closer to Substack, but I would (for screens like mine) prefer a larger font and a 'geometrically' longer line length than either.
About the customization, whether better or worse by some consensus metric, Substack makes it more generic. There was a certain feeling of being "at home" in the old one, reaching a known place in a corner of the web. Substack feels somewhat like a newspaper.
Newspapers have multiple columns of text.
Well, not that it changes much the point, but for the sake of clarity
> Substack feels somewhat like a newspaper website*.
Substack keep it generic, plain and boring to allow any/all writers to use it and for subscribers of multiple substacks to have a consistent experience. They don’t want your brand overriding their brand.
& I think people prefer something more unique and visually inspiring on a one off, but not when they have to constantly switch visuals when switching from Substack to Substack.
A plurality says it isn't much better or worse.
Most people say it's as good or better.
A large minority prefer things the way they used to be.
Sounds like a generic response to a generic question on if Now is better than Then.
That being said, in general I hate 'mobile friendly' layouts. Too much white space for my old-man preferences.
https://stackoverflow.blog/2022/03/03/stop-aggregating-away-the-signal-in-your-data/ seems apropos…
Nice article, in which data scientists apparently rediscover classical time series analysis (or perhaps, a classically trained statistician explains to data scientists what looking inside the box allows one to do).
> A large minority prefer things the way they used to be.
And that minority is larger than the minority who prefer the current layout.
Underappreciated factor. There's only so much that can be concluded from a 21-point net favorability with 54 percent undecided.
One thing I immediately noticed from a product-design perspective when you moved to Substack was that it is *much* harder to read through the archives on Substack. SSC had links to the next/previous article at the bottom of each article, whereas Substack doesn't seem to have this and instead intermittently shows you a 'greatest hits' or something, which I'm sure has tested very positively for engagement and delight or whatever metrics are currently trendy, but is really annoying if you want to just 'read the next page please'. On SSC itself I thought it was a bit naff having to click through 100 meetup/openthread posts to get to the next dose of insight porn, but I much prefer that to not having the option at all!
Yes, this is still a huge problem with Substack even after the other updates.
Yeah, just last week I opened SSC and had this nostalgic feeling. I had been pointed to SSC just a few months before NYT happened. It was, uhm, the first blog I ever read, and I read a lot of it ... so for me the old layout is just linked to the adventure and the fascination of discovering those texts in the first place. I also somehow like the colour.
More broadly, I think by definition it had a more 'personal' touch and at least some people who like your writings might just like that personal touch by association. Substack to me is very fine in general. It's just the same for every newsletter ... no 'oh, there's Scott's place' feeling.
One possibility: When you ask people what looks better, they think about the sites as things to look at. They assign a positive value to ornamental aspects. But when you are actually reading something, ornament is a distraction. What you enjoy when you pay attention to it and what actually helps you achieve your goals are not necessarily the same thing.
Personally, I miss the "next post" and "prev post" links. Great for low-intensity archive binging.
The uniqueness- the blue bands and the formatting were distinctly SSC, whereas every substack newsletter looks the same.
I think the key here is personalization. Substack's design is much better for a generic, unified presentation content *platform*. When you are hosting many writers and you want all of their blogs to look roughly the same, what Substack has done is clearly the correct answer.
However, your blog was not a platform. It was just yours, and that allowed it to be idiosyncratic and optimized for you and your style and content. More precisely, your blog isn't just a blog, its a centerpoint of a community and subculture. Most bloggers are not that. I like Matt Ygglesias, but I don't need a special theme to know i'm in his territory, because his territory is just his content. Your blog is/was a whole subculture, and as a sub-culture, it warrants an aesthetic differentiation that merely being a writer does not.
It's just hipster-ism. "I liked it before it was cool."
People always complain about design changes. Even if you do a well-constructed experiment and prove that people prefer the new design, most of the reactions you get will be complaints.
The upshot is that you need to weigh the utility of the improved design to your average user against the pain of dealing with the complaints about it.
I think people are very reasonable in complaining about most design changes!
I think part of the problem is that "average user" is often a _new_ user, and the best design for them is often very different than the best for a regular, let alone 'power', user.
And for regular/power users, the utility of _any_ design change is often negative, even if/when it is better. Having to relearn 'workflows' is a significant cost, often swamping any meager gains to be had from better design.
Your web design is your face and your domain is your name. Your name should never (have) change(d), and your face shouldn't change drastically all at once. It's also unfortunate to have the same face as so many other writers. SSC's design probably looked the same as some other sites out there, but it was uniquely recognizable among the sites I frequented. Same goes for just about every blog from that era. With Substack, blogs are just content using Substack's surname and wearing Substack's one-size-fits-all face.
The fact that my peripheral vision constantly reminded me that I was reading SSC seems is analogous to how a person's face constantly reminds me I'm talking to them.
I really like this analogy. It captures most of my feelings about the new blog name/look with "it's the colour" filling in the rest.
The substack page is horrifically slow on mobile. I’ll scroll and it will take several seconds to render the newly visible context. And this is on a pretty decent phone.
Modern websites (including Substack, though it's offenses in this regard are minor) are not optimized for enjoyability or readability, but rather engagement. Basically, an alignment problem - we want to reward people who create enjoyable websites, but the actual incentives reward engagement (or whatever other things are driving revenue).
I believe part of the problem is the email format. On some phones, you cannot read in portrait, you have to switch to landscape. In addition, which include screenshots of data set points, surveys, et al. Must be clicked and and they view too large to simply look at.
I don't know if this is a major problem, but it slightly irksome to me.
Could be a broader form of selection bias, wherein Substack's layout is preferable to most people but there's some weird correlation between "people who like your kind of content" and "people who prefer older-style, non-minimalist blogs".
Part of UI design is fashion. Many UI designers follow fashions in UI design, and fashions can change without them actually being an improvement.
Many of us just want to read, including the comments. Substack has more wasted space (larger font, higher line spacing, shorter line wrapping around the comments) and requires additional clicks to expand things, requiring moves away from the scroll bar.
Also, it used to be easier to see which comments you hadn't read yet. That lack makes the lengthy comments much more difficult to read through more than once.
Plus the old site had more contrast. Many of us like readability above all. "Modern" design doesn't always enhance that, especially when just made fancier.
Is it my imagination or are there more comments per post on the Substack version? Substack makes it too easy to say something -- consequently it's harder to "wade" through the sheer amount of content and perhaps the comments are less well-thought out. For example, I never once commented on a single post before the blog migrated to Substack but now I do from time to time. I know it's a dull instrument but what I wouldn't give for some sort of comment upvote system...
I don't think so, SSC comment sections could be *huge*. They were easier to navigate and read though, so it may not have felt like it.
I think SSC comment sections were narrow & deep v. ACX comment sections being wide & shallow, tree structure wise.
Maybe the problem is that substack was designed by really experienced designers, who are living in a bubble, where they produce what *they* like, receiving acclaim and the occasional award from their industry peers, and insulating themselves from any users so crass and uneducated as not to share the professional opinion. (Which IMNSHO, would be most users.)
Alternatively, the problem is that substack was designed for one-to-many interaction, with only token feedback, much like newspaper comment pages. With newspaper comments, some percent of the commenters are ranting one-topic *ssh*les, spammers, and other annoyances. Moreover some perhaps even larger percent are political opponents of some readers, and so considered by those readers as even worse than the less political problem posters. So the harder it is for readers to read other people's comments, the better for the newspaper. And if it's also hard for the original poster to respond to comments, well that's perfectly fine; if my local paper is any sample, newsies eagerly ignore anyone who tries to interact with them.
Though if I were to be specific, it's not the layout that annoys me on substack - it's the whole look and feel, including but not limited to the tools for commenters. We had better thirty years ago, before html was a thing.
"Before HTML was a thing"... do you mean Usenet or something like that? Or did you perhaps mean Javascript, not HTML? Before HTML, there was no Web at all.
I was about to add that you'd also have to go back more than 30 years, but that's wrong as it turns out; HTML turns 29 this year.
I was thinking about trn in particular - a usenet news reader.
But there were other online communication systems before html. Not just usenet.
And yes, I checked the date for html before posting ;-)
SSC did not have a great design, but it had a unique design. Substack, by contrast, has essentially no design.
Just another iteration of the soul vs soulless meme.
I think the mere *fact* of customization explains why I prefer the old SSC design, more so than any of the actual design choices you made.
When I read a website like SSC that has its own look, I feel like I'm *located* somewhere, like I'm in a specific physical place. And there's an intimacy to it, as though someone has invited me into their home.
When I read any Substack, I don't feel like I'm "in a place" much at all. And if I am, that place is just "Substack." Likewise with Facebook and other standardized venues for content.
If every blog looked like SSC, then my reaction to the SSC design would change from "I am physically inside Scott Alexander's private salon, this is what the chairs and curtains look like" to "I am reading a blog."
Mildly interesting sidenote: tumblr (uniquely?) serves every blog in 2 versions, the Myspace-style customizable view and the Facebook-style standardized view. Consider the difference between
tumblr.com/blog/view/nostalgebraist
nostalgebraist.tumblr.com
The latter feels more place-like intimate to me, as I've described, even though the extent of my customization goes little further than "making the background a particular shade of desaturated purple." As long as nothing else looks the same way, that's enough.
YES
I think you're onto something here.
I'm pretty sure I answered "about the same" in the survey, and I stand by that. My general feeling is that the old blog was kind of ugly, but not offensively so, and the new one is nice, but not in a way that particularly stands out, so overall they're both just pretty okay. Consequently I suspect any preference for the old is nostalgia.
"Is it selection bias? My previous readership is, by definition, people who liked my old blog, so of course they like my old blog more than some new one? I’m including this because I know someone will bring it up in the comments if I don’t."
You know us too well.
Substack is pretty good I think. The app is better. But …. I often see replies to my posts that just show the reply. The only option then is to “show thread” which goes back to the main thread, as far as I can see, and it’s not always clear what the poster was referring to.
It seems to me mostly nostalgia. Vladimir's displayed comment seems to me to read like "It was worse, in a way I was used to." (e.g. "The blog name was more quirky, harder to remember... And hey, that felt memorable!" is literally saying it being bad made it good).
Honestly this all seems like overthinking it (which is not a complaint, I'm very fond of your overthinking).
Mostly I think people just hate it when things change. I mean, what's the last time that you updated an app and went "ooh, I LOVE how everything looks slightly different now"? (And this is classically even more pronounced for autistic people...)
Now, I absolutely count myself in that camp! I like the old layout too. Change is the worst, and the old one was comfortably familiar.
My big complaint is that the comment architecture is so primitive. On SSC I could go back to a post I was interested in the next day and easily find the new comments by searching for ~new~. On ACX there is nothing at all like this, so I usually wait to read a thread until it is a few days old, and then never go back.
This. I hardly ever comment here because of this.
This a 100x.
Upthread Pycea mentions a chrome extension which adds this: https://github.com/Pycea/ACX-tweaks
Ah, interesting. I have added it; we'll see if it acts the way I hope. Thanks!
Now we get to do the old complaint of "hey, you had "~ n e w ~" in your comment, and now I hit it every time I search for it!"
:-D
Heh, right. Sorry about that.
I wonder if the tech would permit it to expand that string into a euphemism if you ever tried to type it in a comment.
I never read this on mobile. The thing I don't like AFAICT is the current blog has a blank white background, identical to every other substack, without any customization or charm. It makes the text feel almost oppressive, having nothing to hem it in at all.
I realize that we didn't have this feature on the old site either, but I wish I could collapse comment threads
I'm not sure if this comment is earnest, but both the old site and Substack allow you to collapse comment threads. On the old site, it was very helpfully marked "hide". On Substack, I think it wasn't possible at first, but this was one of the most-requested features here, and I think Scott's large audience motivated Substack to incorporate it. Now, if you click one of the vertical lines to the left of a comment, then it will collapse that comment and the entire thread following it.
Just… don't do it if you're deeply nested, as it doesn't adjust the scroll location, so if you're 5 screen lengths into the thread when you hide it, now you're 5 screen lengths past the top-level comment that followed.
Yeah, I really hate that! That, and the fact that the link in the e-mail I get about comments replying to mine is only about 70% reliable at actually finding the relevant comment in the link, and otherwise drops me off at a random point in the comment thread so I have to try to find the relevant comment.
Funny enough, I think that it _sometimes_ working is probably a (minor) technical accomplishment – I suspect they must be doing some crazy something something to show you a specific comment. That it doesn't _always_ work feels like evidence that it's not trivial, whatever they're doing.
Linking to a specific item within a HTML page, was a solved problem in the previous century. On the old SSC site, just right-click on the timestamp of any comment, choose "copy link address" and then see what happens when you open that link in a different browser window/tab. You can also send that link in an e-mail and then people can open it from their e-mail client.
Of course, if instead of building a website by using standard HTML features in their intended manner, you go and poorly reinvent all kinds of wheels using unnecessarily complicated Javascript techniques, you will break all those well-established standard browser facilities. So then you need to poorly reinvent them too. And so on..
Agreed!
I'm pretty sympathetic to the designers (and front-end programmers) and all of their JavaScript, but much prefer 'simple' web sites/apps myself.
Thanks! I guess I had forgotten you could collapse the old threads, and never figured this one out. :)
Yeah, it was rolled out in a really hard to understand way. (At first, it allowed you to collapse all the replies to a single comment, but not a comment and all its replies, which was *extremely* annoying when I wanted to skip some subthreads on a thread, but not all of them.)
For me I think it's just nastalgia, but I do prefer SSC. When you first joined substack, I installed some sort of "make ACX look more like SSC" add-on, which does a pretty crappy half-job, but I still get annoyed when I read ACX on a new device without it and it just looks like substack.
The old blog comments were terrible on mobile. After a small number of levels of indentation,
you
were
reading
comments
that
looked
like
this
and
they
were
really
really
really
long
because
even
a
small
number
of
words
takes
up
a
lot
of
space
when
written
one
per
line
and
commenters
were
not
really
all
that
brief
since
they
had
a
lot
to
say
I always assumed this was intentional, to discourage excessively long threads.
It's a common class of bug; substack evades it by putting deeply nested comments behind a link to a new comment page.
hacker news evades it by forbidding replies beyond a certain depth.
I thought the old blog deliberately stopped nesting at ~4 levels.
Yeah, that part really sucked.
my personal answer is “I’m very used to the old layout & I’m not used to the new one yet.” I do mostly read on mobile, but I still preferred the older version tbh.
Oh, also -- tags. I *loved* the tags.
This is not complicated. Your SSC layout had STYLE, and it was your personal style. Presumably your readers like your personal style, otherwise they would not be your readers (I do like it). Substack style on the other hand is minimalist, out of necessity, because it has to cater to a diverse audience, so it has no personality. EDIT: also default Substack font for articles is imho awful, though it is better in comments.
This maybe overlaps with Whither Tartaria. I weakly preferred SSC's old layout over Substack. The reason I did was specifically is that it was, in a cute way, kind of ugly and quickly thrown-together, and it made me feel like I'd found a brilliant blog that is niche and obscure, because a website as ugly and quickly thrown-together can only be niche and obscure, and thus not only do I enjoy reading a brilliant blog, but get to pat myself on the back for finding it, quite literally, "in the rough". This is pretty resistant to reality too, e.g. my enjoyment of SSC's layout and my feeling of it being something niche and underground didn't really diminish when I realized that NYT picking on a blog probably means it's not niche or underground.
At the same time, when I come back to thoughts of how warm and fuzzy the old SSC was, I do not remember that we couldn't collapse comment threads. So I guess, for me the explanation that it is nostalgia for the "band's old shit", as Freddy deBoer puts it, is correct.
Wild guess: humans are neural nets, and the high quality of the old blog caused its layout to trigger a reward function. We like the layout because we read good things there, so our brain assumes the layout must be good. All reasoning after that is our attempt to provide a logical justification for an illogical association.
If the is true, a similar or more extreme pattern should be visible in new readers who spend time on substack without seeing the old blog.
I vastly prefer the old layout.
The black sans-serif text on white background feels sterile and very visually immediate. The ACX logo is way off in the corner, and there are no visual borders. The navigational structure of the site is now "list of articles + a few sidelined sections." This makes your site feel like an unformatted powerpoint where each slide is a wall of text, or maybe like an email client on default visual settings. It's just one white text surface with no layers or edges or footholds. This design makes you inaccessible.
I get that this layout is mobile-optimized, but if you are thus required to cede aesthetic vision to this umpteenth degree... why have a website at all? You could just send emails for people to read in their email apps.
In contrast, the light-blue and faint-gray color scheme with the fonts you had chosen made your articles seem like they had come from a strange book. And, there were multiple sections that contained diverse types of items - some of which I never fully understood! I liked that I had to puzzle some of it out. The old site was like an illuminated manuscript. It had layers, and pages like a book has pages, and errata, and appendices. It felt like you were writing it and simultaneously writing in its margins. Its design invited people into you.
I don't find myself wanting to comment here as much as before. Some of that is my own change in habits, but Substack's design sure made it easier to disengage.
> You could just send emails for people to read in their email apps.
This is almost exclusively how I consume ACX, to be honest. Nowadays, I only come on the actual site to read comments whereas I often actually opened the old blog in a browse, because it felt "just as nice" as my fairly custom email experience. Substack looks like every other blog aggregation site and even conjures up thoughts of medium.com etc. Why would I want to look at this site when there's really nothing to look *at*.
>Nowadays, I only come on the actual site to read comments
Me too! I generally read this in my email client whereas I always read SSC at the actual site.
+2
Old layout is comfort food. But Substack's is healthier. (Comments system aside).
Somewhat related to the first thing you posted. If you're a fan of a sports team, and you first fell in love with that sports team in and old, dingy arena. And then they move into a brand new state of the art arena.
I bet a lot of fans would say they preferred the old arena. Especially if a number of other teams make a similar move to similar arenas around the same time.
We associate the old design with Scott's writing (and, to borrow from the other post, when you didn't suck :-) ). Now, it's not unique. It's the same of a bunch of other writers, which we have varying levels of warmth toward.
I think Substack's is better because of text's max width & visual minimalism.
The biggest problem IMO is lack of organization through. Kinda hard to find specific texts. If at least there was a separation between cyclical stuff like Open Threads and text content...
I didn't read your old blog at the time, but I've read tons of it now. I prefer the new layout. The only way I don't prefer it is that it looks "generic" and like every other Substack. With the old one, I know who I'm reading. But you're pretty distinctive so I don't mix you up with, say, Matt Yglesias, the way I might mix him up with Noah Smith. Also, Substack is way easier to read on a phone IMO.
I think substack are trying to relentlessly focus on their core proposition rather than bells and whistles. They are very much trying to let the content be the king on mobile and desktop and as a blank canvas they aren't trying to be personalised in a nice way like SSC was. I think they will eventually allow you to customise it, but it's not core to their mission of getting writers paid.
I really don't think you did a bad job especially after the designer tweaked everything, it reminds me of gamer sites which isn't necessarily a bad thing. In the end despite how important designers think they are, how it looks barely matters at all.
• Not all comments are visible when loading the page. (Or weren't when I last checked without ACX Tweaks. Now they do seem to load immediately. When they didn't, this had various consequences beyond that you have to click multiple times: you couldn't easily search in all comments, or save where you've left off by clicking the permalink of a comment.)
• When viewing the thread below a comment, the blog post isn't visible.
• No way to see which comments are new.
• No formatting in the comments.
• No way to jump to the parent of a comment.
• Inexact timestamps.
• When following a link to a comment (or just loading a page with a comment's url) it doesn't jump to the right place. (In Firefox, you can then go to the address bar and press Enter to jump to the right place. In Chromium, that reloads the page, and jumps to a wrong place again.)
• The header that is hidden when you scroll down and shown when you scroll up is annoying. (If you scroll down just a bit more than you wanted, you can't undo that by scrolling up just a bit.)
• The entire archive can't be loaded at once. On the Archive page, I have to scroll to the bottom repeatedly to get to an old post.
• SSC had nice aesthetics, ACX has none.
The ACX Tweaks browser extension (https://github.com/Pycea/ACX-tweaks/) fixes some of these problems, but not all, it doesn't work perfectly, and most people presumably don't use it.
1) I have near crippling, drop out of grad school/lost a job/have problems with tasks of daily living ADHD when my stimulant wear off. Substack's pro design has S_H_I_T implementation. Every hang up for .75 seconds. Every 1 second animation instead of just loading the page. Ever "lol I'm going to reload for no reason" event while trying to scroll back to a comment from a non-new-tabbable read more button. They are just another aggravating. Just shove a pencil in my eye.
2)Same reason a family owned Italian joint can feel like family while Olive Garden feels like a strip mall. Teams of experts aren't an assurance you'll get good art.
The old SSC layout gave me the vibes of the old bulletin boards, "internet before the Eternal September", and "this exceptionally interesting but little-known and therefore sort of pleasantly elitist thing". While uniform and minimalist layouts are likely more readable and even more aesthetically pleasing in some way, they are also a sign of times when big and often disliked corporations want to constrain people's individuality and self-expression into homogenous boxes.
Yes, this.
Here's what comes to mind for me.
1) I like knowing where I am on the internet. While simple, SSC felt like its own place. No substack feels like its own place.
2) The smaller font on SSC felt more congruent with the detailed style of writing. More content on-page helped me to keep more of the reasoning chain in-view.
3) The sidebars added loads to the culture! The local ads! The local blogs!
4) It wasn't actively bad. People weren't complaining about the old site (I expect). I think things would've been different if you'd really broken the old site somehow.
I guess what I'm saying is that the old site had a unique character (in being visually different from other blogs and having the site UI be filled with things of local-interest), it wasn't actively broken, and (I think) the smaller font was a better fit for the detailed content with long chains of reasoning.
>2) The smaller font on SSC felt more congruent with the detailed style of writing. More content on-page helped me to keep more of the reasoning chain in-view.
It may be that this says something about my short term memory more than anything about web design, but this definitely strikes me as true. It also makes it difficult to read quickly because I'm constantly jumping to the next line or scrolling, it makes the thoughts/arguments in the article feel more disconnected.
People don't like same-y-ness for things that they actually like. But they do like it for things they kind-a-like.
Almost all web redesigns (well most brand redesigns too) suck by removing personality and density in favor of cleanliness. It's the same reason I hate modern architecture. Actually clunky small font websites like old reddit, ssc, and goodreads are good. (Speaking for desktop)
Actually I want to rant about the goodreads redesign since that's new. Lots of people have been complaining about this for ages but I hate the revamp. Yes it is less "ugly" but that ugliness removes a bunch of functionality or requires more clickthroughs to see info like isbn or page number.
Consider your blog. The new app removes the little sigil at the top and the colors. It's a stretch but these could be considered functionality by serving the purpose of ssc-territory/tribe signaling. The ads were all rationalist ads further encasing the bubble.
Personally, my issue with Substack (and why I never tried using it myself) is just that it's not customizable - every blog looks exactly the same, and that's no fun. Even a bunch of blue squares that you slapped together in a few days is nice because it's *yours*.
I have no way of knowing how many other people experience the same thing I do, but here it goes. Substack presents a tiny column of text with right around double the text's width in whitespace on the left and right side. The old style wrapped properly for me.
My 2 cents:
Slate Star Codex's homepage feels rich, full of content, like a library of old books full of mystery.
Astral Codex Ten doesn't. It sort of diminishes how you feel while reading certain posts.
Perhaps you might ask people: "do you think I should go back to the old layout, if Substack let me?" and see what they say. I'd say no, despite what I have just written.
>Slate Star Codex's homepage feels rich, full of content, like a library of old books full of mystery.
I think it is important to distinguish this from "nostalgia" which a lot of people are pointing to. At least for me, this was an active part of the appeal of reading SSC from the moment I first got linked there. Getting linked to ACX would have felt very different if I were coming in blind. I'm sure arguments could be made for that ultimately being good *or* bad. Just kind of a "vibe" thing (Substack doesn't have one, of course).
I started to follow SSC relatively late (mid 2019) so I am not sure if I qualify to be nostalgic!
In the early days on Substack, the biggest thing for me was the inability to collapse comment threads. They've fixed that, but there's still a lot of problems with comment threads, some of which I've seen other people mention. It's hard to click through to a specific comment, there's some slow-ness in the loading of comments, you can't have both the main post and the comments visible (unless it's an open thread), and so on. My guess is that a lot of this has to do with how Substack stores the comments. (As I'm writing this, I just noticed that Substack has an option to view comments "Chronological" or "New First" - I don't quite know what the difference is, but if they are sortable, then that might require some sort of data structure that makes the above problems inevitable.)
The other big thing I miss here is the whole archive feature - links to "next post" and "previous post", and links to each month of the archive, and a browsable front page of the whole archive. I think a lot of this is a self-conscious change from the world of blogs, that were seen as chronological writing, to the world of social media, which are seen as ephemeral rather than things to come back to. (It's an interestingly different pair of conceptions of what "Web 2.0" meant!)
All the visual touches and the blogroll and stuff are fine, but it's really these structural things that make the relevant difference. And it seems that Substack expects most authors to prefer things like sortable comments and ephemeral posts, rather than an archive, and that's why their thing, that is more professional, is worse for what your readers want.
It looks like "Chronological" means "Old First" (which is how SSC forced comments to display). Substack should make that terminology more obvious, since "New First" is a kind of chronological too.
Mostly the colours, honestly. Substack is a uniform, eye-searing white. Slate Star Codex had a grey frame, a bit of blue on the top, a bit of light grey on the sides. The overall effect was easier on the eyes. (But also, I don't hate the Substack layout and design. I just prefer the Slate Star Codex one.)
Edit: https://astralcodexten.substack.com/p/why-do-people-prefer-my-old-blogs/comment/6400906 makes some nice points, too, about comment loading behaviour and the archive.
The latter I mercifully had yet to notice, because I've not had the urge to browse this site's archive, because if I had, I would certainly have been very annoyed at yet another instance of the far too prevalent 'lazy-loading content' paradigm. I flippantly want to say "god created scrollbars for a reason" - but basically that would just be a tongue-in-cheek way of saying that I *like* that a scrollbar tells me how much more content to expect at the bottom of the page, and this is completely nerfed by the lazy-loading content paradigm. Prefer clear pagination if we do need to chunk it for size.
>Mostly the colours, honestly. Substack is a uniform, eye-searing white
Yes, the wall of white is a little painful on my eyes. I don't think I would prefer a full on Dark Mode, but the muted colors of SSC were more pleasant, at least for me. Some people have mentioned the that the smaller, Serif font on SSC resulted in denser blocks of text that made it easier to consume some of the more challenging writing, which strikes me as spot on. I think this is the biggest thing for me, although there are a number of other good points brought up throughout the comments.
As someone above mentioned, it's about margins. Our eyes want margins. Color is nice too. MR still has its classic green. There's no Marginal Revolution without margins and green.
I prefer the compactness, generally, and increased information on the non-single-item pages of the old format. The layout within a single post/article is fine enough (though not better). This isn't limited to SSC.
And the comments section. Of all the areas they could really improve functionality, the comments section has to be the most wasted opportunity. Nothing that is "improved" actually improves the user experience or promotes more engagement, generally, and it's absolutely terrible for highly active comments areas.
I suspect there's a signaling mechanism at play--the expense in building and maintaining a distinctive, functional blog theme acts as a signalling proxy for the quality of your blog. By contrast, appearing as a generic substack or medium blog--while functionally identical or superior to your custom theme--signals nothing other than "I signed up for an account with one of these services."
Everyone hates the new minimalist designs when you ask them about it, but they engage more with them and they're way easier to adapt to different browsing environments.
People engage a lot with Twitter but I avoid that website because it's fucking awful in every way, design wise. "Engagement" doesn't mean "good," it doesn't mean "memorable," it doesn't mean "improving your life in any way." It means you're wireheading people to click the "Next" button. No more, no less.
Relax. You are OK. The website is OK. The content is the thing.
My only comment on the current website is that I would like to have some html in the comment entry at least for links.
That's where they're falling down by contrast with the old blog and it's *badly* needed. It's cutely retro that we're back to using *x* to indicate italics like the early days but it's no longer the Century of the Fruitbat, you know.
Substack is the same for everyone.
Your readers' aesthetics will not match the general public's.
The trouble with Substack's layout models is that they are pushing very hard for writers/authors/content creators. 'Here's how you can start your own Substack! Here's how to price subscriptions!' Meanwhile, if you want to know how to use Substack to post comments, it's "Oh, here's how you subscribe to a Substack. You want to know about comment formating? Hey, here's how you start your own Substack!"
And sure, that's their business model: get people producing content that draws subscriptions that they take a slice of for their operating costs and (fingers crossed) profits. But it makes it difficult at best, and actively lousy at worst, for anyone who's a reader but not interested in creating their own Substack. Your job is to passively consume and pay for the privilege.
By contrast, even though WordPress can be a steaming pile of horseshite, the old blog was simple and did what it said on the tin. Even a fool like me could learn simple HTML tags for formating, links and so on. It didn't need bells and whistles, it needed "Here is Scott's post. Here is how we comment and have discussions" and that is what we got. Flashy but superficial design puttering-about wouldn't have contributed to that, and wouldn't do anything. Yeah, great, your paid theme is ultra-high-minimalist post-ironic Zen repurposed via Bauhaus and so achingly chic it has won design awards five years from now, but I can't find the 'sort by new' button.
I prefer the old layout, probably nostalgia. A few quick notes:
-There's a big difference between reading articles in your inbox and on a website. Substack feels perfect for email but very sparse for a website. There's no links, the commenting sucks, it feels very "read and leave" not "read and stay".
-I've found myself clicking over to other substacks a lot more on the website. Sometimes this is cool, I especially like it when I see one writer I like commenting on another writer's article (shoutout to FdB) because it brings back that old blogosphere vibe everyone is missing.
-Again, reading from the inbox, it feels very, I dunno, "special" or "professional". The inbox is where real things happen. But the website feels like "content". Like, everyone's consolidated around some standard practices which are very optimized but they "smell" like content.
Strikes me as a branding issue - you are branded a certain way and it's more "aesthetically ambivalent independent-minded nerd."
The old layout is ugly and cramped and therefore on-brand. This new layout is easier to read, optimized for mobile, etc etc. That's off-brand.
From a design standpoint substack is fine. A little bit bland
From a technical implementation standpoint substack is AGGRESSIVELY bad.
Some of the open threads put my poor old desktop to its knees. It's just text on a white background, why does that require layers upon layers of shitty javascript?
> From a technical implementation standpoint substack is AGGRESSIVELY bad.
+1
Something displaying limited amount of text has no right whatsoever to lag in any case.
Something lagging while displaying less than 10 000 pages of text is broken and should not be used.
Sadly, people accept such brokeness.
Substack randomly lags and noone really cares. OK, I cared enough to switch to other entertainment but I was not paying anyway.
FYI - from someone who does research - the way you frame that question is designed to get a specific response. You might frame the question differently: Is ACX or Substack easier to read?Easier to Navigate? From there you might actually get more pertinent results. People don't like change in their visual environments. Think of all the hand-wringing about Google's new logo, Netflix's new design, Medium's new logo, that California School's new logo, the new...[new design of something] and there is a lot of scorn for it, but it's all gone in moments. In a few months, people can't even remember that Google had a serif font and don't even care. It might be that people don't like the attachment to what SUBSTACK represents - that ACX felt more personal to them because it was designed in 2 days. It doesn't mean you have some secret visual taste because people prefer it - you definitely don't. It's what it meant to them. I don't personally care, but I deal with this all the time. A colleague calls it the, "You moved my cheese" problem.
First of all root of the problem is "hired a team of really experienced designers".
Maybe it is hot take of old fossil, but if you compare http://prize.hutter1.net/hfaq.htm#html to Facebook then the first one is far better.
Piles of JavaScript, horrifically bad performance (it displays text and occasional image, and it brings my computer down), unneeded elements.
http://prize.hutter1.net/ may be going too far (I would add margins at least), but SSC was for me in nearly optimal position. Substack broke several things and added nothing better at all and added plenty of JS infestation beyond what was useful or needed.
And removed useful things.
For example there is no way to see just new comments.
Dysfunctional threading - sometimes just scrolling is enough to hog my computer (16GB RAM and displaying formatted text is beyond it!).
Unwanted extra stuff: "subscribe now", subscribing to access it, subscriber-only posts.
Some of that - like trying to spam my email for no good reason, subscribing-related fluff is a direct cost of monetization.
Part of that is team of designers breaking things that worked well, partially because they actually think that it makes thing better, partially to justify their employment.
> Whenever anyone asks, people say they hate the modern minimalist designs and wish they could go back to the complicated colorful ones. But for some reason nobody ever does. Is this just the Internet version of the same general phenomenon?
I designed my personal website exactly how I like it. But it gets about 1/10000000000000000000000000000000000000 traffic of FB.
https://news.ycombinator.com/ and https://github.com/ are rare cases of websites where designers are not horrifically bad.
Related notes: I created throwaway account to avoid spam. And I will likely never notice replies as searching ~new~ or equivalent is not available.
And no, I am not interested in getting more notification email. I already get about 100 a day, and I accept it only because I am highly paid for that.
I don't think you can separate "Why I suck now" from the site thing very much. You have this overall perception among some that things have changed, and that they've changed for the worse. This may or may not be the case, but if it is:
1. You would have some people who were overall just unsatisfied, mad about the writing/topics/takes or whatever, and taking that out on both you and the site
2. Same as above, but didn't want to talk about the writing (in rationalism you can't be mean!) so they talk about the site
3. Same as 1, but they don't even realize it's writing-related; it's just a general sense of discontent
So those 3 are all possible, almost all certainly happening somewhat based on what you've written. But you also have a site that swings hard-nerd and pretty spectrum in it's readership; you know who hates change in comfortable, favorite properties? That segment.
All that to say, so long as you are comfortable with what you are writing I wouldn't worry too much - it would have probably had to have been exactly the same to not get a lot of complaints. And if it was actually the same, it probably still would have got a fair amount of complaints, just aimed at different things.
(postscript afterthought: Another thing that happened is you were gone for a long time, and the comments section broke, permanently, in the way that it will never be what it was again. You can take everything people might believe about your writing and apply it to the comments, as well)
I assume some of the more objectively better things (faster to load comments, easier to read and scroll comments) are just giving people an associated warm fuzzy about the other aspects (minor layout, color, etc.)
There's some kind of horrible stop-the-world pause on Substack that freezes the page for several seconds at its most pathological. It doesn't happen reliably and I have no reproduction steps other than scrolling around.
I get the impression that it's something synchronous in the JS that halts execution for a while, and that it causes small freezes as well as big ones. For example, while typing this comment, input has frozen for half a second multiple times, then caught up.
I don't care what the design looks like as long as the text is vaguely readable and the visual noise is kept down, but a short blog post causing the page to freeze is ridiculous. I can go to Project Gutenberg and scroll through the entirety of The Count of Monte Cristo (which would take more than 1000 pages to print) without any slowdown.
No offense, but Substack is considerably easier on my eyes :) I for one really appreciate the minimalism, I get easily distracted, so having the content surrounded by tons of links in sidebars and topbars wasn’t really my thing. Substack mostly just keeps out of the way.
OTOH, I’m not really an SSC veteran — I read a few posts in the years before ACX, but didn’t become a regular until after the transition. So I don’t really have any nostalgia for the halcyon days of SSC to go on.
Also, I can imagine how people who *can* afford to get distracted / read faster than I do might miss that link-laden context in the UI. But as for straightforward having an aesthetic preference for ACX — that doesn’t really compute. But hey, that’s how aesthetic preferences work, I guess.
I think there's def just a nostalgia factor. People generally prefer the past when asked even though it was objectively worse.
I mostly read your posts in my mobile Gmail client now, which is much better than either mobile website ever was. Substack is good.
My mobile experience has been that my phone browser's "show simplified view" option works more reliably on the old site but it's less necessary on the new one, which is probably a reason in favor of both WordPress and minimalism.
I don't really think there's a mystery here. Your tech friends had something like a decade and a half to optimize a set of features and layouts exactly to your readerships preferences, and substack has ... not. Two things strike me. First, the tech people who were making changes actually talked to your users directly. A lot. There were no barriers between them, and in many cases they were the same people. Substack, to the extent that they get feedback at all is probably mediated by third parties, mostly you. Second, even if substack knew how to make the perfect set of features for you, there's no reason for them to do that unless it would benefit enough of the other bloggers and therefore substack. I think most of it comes down to whoever it was that designed your old blog thought of your readers as the users, but substack thinks of the authors as the most important users. It shouldn't be a surprise that your readers prefer the former on net.
My theory is that the previous site had a better fit between form and content. Scott's pieces are long and involved; they require investment and effort on the part of the reader. The font, line-spacing and layout made the text slightly more information-dense and very slightly more difficult to read, which accords with the higher-than-usual difficulty of the content. Like having an ornate frame on a highly-detailed painting.
The Substack font, line-spacing and layout are easy to read. They appeal universally, and offer very little friction. So seeing the same high-effort writing presented in this low-friction way is jarring. Like printing a treatise by Rousseau in an Animorphs book.
Oh, it's lots of things. The old blog had more color, the old blog had more features, the old blog had more personality - this one is pale, poorly featured, and bland. Would much prefer SSC.
I prefer if each blog that I read looks different, so that I'm aware in the back of my mind who I'm reading. The context change when I go to a different blog is reflected in the different surroundings. The various associations/expectations that I have with the blog also get attached to the layout, and I get used to particular ways of interacting with the site.
So any change to site layout has short-term negatives, as I lose that familiar context and it's a bit jarring. This particular new layout is very similar to lots of other sites, so it can't grow as much into a familiar distinctive context.
Substack is far too white. SSC had colors and sidebars. Substack is glaring and barren.
I think that most people prefer websites with white backgrounds, although certain groups of people, like coders, prefer darker backgrounds. I wonder if ACX is sufficiently tech-adjacent to have mostly people who prefer dark backgrounds, or at least backgrounds with some color. It would be an interesting thing to include in a future survey.
And for a literary description of the effect whiteness can have on the human psyche, we turn to Herman Melville:
---------------------------------------------------
CHAPTER 42. The Whiteness of the [Website].
What the white [website] was to [Scott Alexander], has been hinted; what, at times, he was to me, as yet remains unsaid.
Aside from those more obvious considerations touching [Substack], which could not but occasionally awaken in any man’s soul some alarm, there was another thought, or rather vague, nameless horror concerning him, which at times by its intensity completely overpowered all the rest; and yet so mystical and well nigh ineffable was it, that I almost despair of putting it in a comprehensible form. It was the whiteness of the [website] that above all things appalled me. But how can I hope to explain myself here; and yet, in some dim, random way, explain myself I must, else all these chapters might be naught.
Though in many natural objects, whiteness refiningly enhances beauty, as if imparting some special virtue of its own, as in marbles, japonicas, and pearls; and though various nations have in some way recognised a certain royal preeminence in this hue; ... and though, besides, all this, whiteness has been even made significant of gladness, for among the Romans a white stone marked a joyful day; and though in other mortal sympathies and symbolizings, this same hue is made the emblem of many touching, noble things—the innocence of brides, the benignity of age; ... though in the Vision of St. John, white robes are given to the redeemed, and the four-and-twenty elders stand clothed in white before the great white throne, and the Holy One that sitteth there white like wool; yet for all these accumulated associations, with whatever is sweet, and honorable, and sublime, there yet lurks an elusive something in the innermost idea of this hue, which strikes more of panic to the soul than that redness which affrights in blood.
This elusive quality it is, which causes the thought of whiteness, when divorced from more kindly associations, and coupled with any object terrible in itself, to heighten that terror to the furthest bounds. Witness the white bear of the poles, and the white shark of the tropics; what but their smooth, flaky whiteness makes them the transcendent horrors they are? That ghastly whiteness it is which imparts such an abhorrent mildness, even more loathsome than terrific, to the dumb gloating of their aspect. So that not the fierce-fanged tiger in his heraldic coat can so stagger courage as the white-shrouded bear or shark.
...
Bethink thee of the albatross, whence come those clouds of spiritual wonderment and pale dread, in which that white phantom sails in all imaginations? Not Coleridge first threw that spell; but God’s great, unflattering laureate, Nature.
I assert, then, that in the wondrous bodily whiteness of the bird chiefly lurks the secret of the spell; a truth the more evinced in this, that by a solecism of terms there are birds called grey albatrosses; and these I have frequently seen, but never with such emotions as when I beheld the Antarctic fowl.
I remember the first albatross I ever saw. It was during a prolonged gale, in waters hard upon the Antarctic seas. From my forenoon watch below, I ascended to the overclouded deck; and there, dashed upon the main hatches, I saw a regal, feathery thing of unspotted whiteness, and with a hooked, Roman bill sublime. At intervals, it arched forth its vast archangel wings, as if to embrace some holy ark. Wondrous flutterings and throbbings shook it. Though bodily unharmed, it uttered cries, as some king’s ghost in supernatural distress. Through its inexpressible, strange eyes, methought I peeped to secrets which took hold of God. As Abraham before the angels, I bowed myself; the white thing was so white, its wings so wide, and in those for ever exiled waters, I had lost the miserable warping memories of traditions and of towns. Long I gazed at that prodigy of plumage. I cannot tell, can only hint, the things that darted through me then.
...
Nor, in some things, does the common, hereditary experience of all mankind fail to bear witness to the supernaturalism of this hue. It cannot well be doubted, that the one visible quality in the aspect of the dead which most appals the gazer, is the marble pallor lingering there; as if indeed that pallor were as much like the badge of consternation in the other world, as of mortal trepidation here. And from that pallor of the dead, we borrow the expressive hue of the shroud in which we wrap them. Nor even in our superstitions do we fail to throw the same snowy mantle round our phantoms; all ghosts rising in a milk-white fog—Yea, while these terrors seize us, let us add, that even the king of terrors, when personified by the evangelist, rides on his pallid horse.
Therefore, in his other moods, symbolize whatever grand or gracious thing he will by whiteness, no man can deny that in its profoundest idealized significance it calls up a peculiar apparition to the soul.
...
But not yet have we solved the incantation of this whiteness, and learned why it appeals with such power to the soul; and more strange and far more portentous—why, as we have seen, it is at once the most meaning symbol of spiritual things, nay, the very veil of the Christian’s Deity; and yet should be as it is, the intensifying agent in things the most appalling to mankind.
Is it that by its indefiniteness it shadows forth the heartless voids and immensities of the universe, and thus stabs us from behind with the thought of annihilation, when beholding the white depths of the milky way? Or is it, that as in essence whiteness is not so much a colour as the visible absence of colour; and at the same time the concrete of all colours; is it for these reasons that there is such a dumb blankness, full of meaning, in a wide landscape of snows—a colourless, all-colour of atheism from which we shrink? And when we consider that other theory of the natural philosophers, that all other earthly hues—every stately or lovely emblazoning—the sweet tinges of sunset skies and woods; yea, and the gilded velvets of butterflies, and the butterfly cheeks of young girls; all these are but subtile deceits, not actually inherent in substances, but only laid on from without; so that all deified Nature absolutely paints like the harlot, whose allurements cover nothing but the charnel-house within; and when we proceed further, and consider that the mystical cosmetic which produces every one of her hues, the great principle of light, for ever remains white or colorless in itself, and if operating without medium upon matter, would touch all objects, even tulips and roses, with its own blank tinge—pondering all this, the palsied universe lies before us a leper; and like wilful travellers in Lapland, who refuse to wear coloured and colouring glasses upon their eyes, so the wretched infidel gazes himself blind at the monumental white shroud that wraps all the prospect around him. And of all these things the Albino [website] was the symbol. Wonder ye then at the fiery hunt?
Why do people like your old design? Was it objectively good? It was certainly more complex and had a signature, and was not cookie cutter, and indicated an actual human cared about it. But these things are not inherently antithetical to a "clean" or "sparse" design.
Could Substack look "minimal" and still have quality motifs? https://dribbble.com/shots/16022934-Substack-redesign OR https://dribbble.com/shots/16038753-Substack-logo-redesign indicate perhaps it could. So why would it not. What would compel such a well-funded team to simply not care?
As you've asked, why did Substack—in other words, everything digital—end up boring and awful? Because designers deliberately engaged in an anti-aesthetic, and anti-intellectual ideological mass preference falsification through fear and intimidation. They consciously led the entire web going to shit, and exactly zero designers (more accurately, approximately a single one) fought back against it.
You can learn about this in following two series:
Fall of the Designer: https://web.archive.org/web/20201108113232/http://www.elischiff.com/blog/2015/4/7/fall-of-the-designer-part-i-fashionable-nonsense
Critical Sharks: https://web.archive.org/web/20200220012356/http://www.elischiff.com/blog/2015/3/4/critical-sharks-part-i-you-cant-say-that
For anyone wondering... yes, I did cover the question of responsive design and mobile. And the short answer is no, these developments in screen viewport variability absolutely did not necessitate us getting boring layouts. That is the fault of ideological designers who hate you.
Fall of the Designer Part III—Conformist Responsive Design: https://web.archive.org/web/20201108140316/http://www.elischiff.com/blog/2015/4/21/fall-of-the-designer-part-iii-responsive-design
I think it could have to do with the uniqueness of the SCC website, and the way that this effects information retention and engagement.
I remember noticing when I started using a kindle that things I read would sort of blur together and I would find it harder to coherently remember an individual book. Whereas a paper book has a distinct colour, size, shape, mass, smell etc associated with it.
I find a similar thing happens with substack, where there is nothing distinct between any of the blogs to latch onto. This makes it harder to kickstart the process of information sorting, and there are less sensory clues for memories to form onto.
Contrarian take: ACX looks better, SSC is more useful.
I like having any color at all instead of literally just black text on a white background, but the substack layout isn't actively bad or anything.
SSC blog had its own character. Substack ones all look the same, and the design is made as lowest common denominator for a million of blogs, all alike. It's not optimized for SSC reader/writer, it's optimized for whatever criteria Substack has. Just like a lot of supermarket food tasted like crap because it's optimized by price, durability, transportability, etc. but not taste. I wouldn't say Substack design "tastes like crap", but definitely quite a bit bland.
It's not a huge deal though - I wouldn't stop reading the blog because of it.
1. Modern UX design with its obsession with whitespace and disdain for borders is a plague upon humanity.
2. The old layout had marginally more characters per line, which I think strikes a better balance there.
3. The two points above combine to make it so that on the new site, 2/3 of my monitor width is pure whitespace, while on the old site, more than half the width is being used, and even the unused parts are better since they aren't bright white lights.
I think the whole thing is just path dependency - if it had always been the other way around, no one would have minded - and also, the only people to say anything are those with strong feelings about it being bad. You solicited the comments in a way that maximizes negative feedback.
If I needed to take a stab, I'd say it's personality. The old blog with its mediocre-matched colors wasn't the prettiest thing, but it was its own thing. It was also a bit reminiscent of the 'old internet' with a lot of distinct, personal sites, which definitely helped with flair. Your substack, on the other hand, looks like any other substack. In fact, it pretty much looks like any other modern text website - if I take my glasses off, l could not tell whether it's medium or substack or even New York Times (I'm exaggerating a bit, but I hope you get the point).
That being said, there are also a few trade-offs. Mobile experience is much better on substack. Comments, on the other hand, load in chunk-wise and, as one of your quoted comments points out, does not work too well. This progressive loading is great for SEO and page speed (since your page is not ten times larger due to 500 comments), but it's not to great for reading large threads. It's quite obvious that they are a second-class citizen and are especially not optimized for the kind of in-depth discussions this readership generates.
Then, of course, there's also some nostalgia thrown in and some resistance to change, but I think the above is enough to explain a lot already.
I liked the prominent tab listing corrections in the old version. Showed you were different from most public thinkers.
I was someone who came to your old blog relatively late, only a year or two before the switch. I don't have strong opinions one way or the other, but I did notice that the comment formatting on the the comments of your old blog was kind of broken for at least 6 months before the switch. I tried multiple browsers and it didn't fix it. No one else seemed to be complaining about it (and I asked on the subreddit but never got a response), so no one else seemed to have that issue? But it made the comments completely useless for me, so I like not having that issue on substack.
People associate the memory of intellectual satisfaction and joy of their favorite post with the layout of the old blog.
The old one had a retro, unique, weird aesthetic. New one is more generic. In a vacuum, new one would probably be better, but the old one stands out amidst the sea of identical substacks
"Meanwhile, Substack is run by tech industry veterans who probably hired a team of really experienced designers and thought really hard about every aspect of their product. It doesn’t make any sense at all for me to do a better job than them. So what’s going on?"
What's going on is that web designers are (A) aggressively bad, having developed their own culture of what websites "ought" to look like, which I suspect is driven to constant churn by the need to seem novel common to most artists, (B) adhere to a lot of stupid design principles from Apple which say you can never be too minimalist, (C) believe people don't like text, and (D) take Twitter and Facebook as models of website organization.
I would ask 3 designers. Some people hate change and are very vocal about it. Even if something is better, they want it the old way. As a usability engineer and designer, I test and ask people for their opinions by testing them. See how fast they use something. Design should stay in the background to the meaning. Get people the content they want most quickly and clearly and you have excellent design. People won't notice great design. Myself, I prefer text that I can turn to audio which allows multi-tasking. I do read on a small chromebook but I listen on mobile.
>Get people the content they want most quickly and clearly and you have excellent design.
I'm not sure everyone likes this purely functional approach to design...(some) people still like aesthetics.
What do you like? Content is king they say. When I'm coming to read words, visuals are to support the message. That's just me. I'm a designer and people are so busy that my design work stands in the background to what people came for.
>What do you like?
Basically anything other than a solid white page.
>Content is king they say.
Who says this? Designers? Product Managers? I like to *enjoy* the content I'm consuming, I don't necessarily want the visual equivalent of Meal Squares while doing my web browsing (and for me specifically the sea of white isn't just neutral-boring, it is actively unpleasant). I get that some people *prefer* the flat white look, I'm not disputing that, but I'm taken aback by your confusion that some people would prefer something else.
I feel like the prevalence of themes and dark modes supports me here, do you have some reason to think differently?
Statistically, there are people who read dark mode better than black on white but most people can read black on white better. Myself, I prefer written content in audiobook format for multi-tasking. Several people who commented say they read only on their smartphones while other read only on their laptops.
The fastest way of reading is with a line length similar to newspapers. For me, while I don't mind certain vignettes while consuming content, I mostly prefer to stay focused on the topic. Otherwise, I have the natural environment as background. That is enough to make it unpleasant to try to stay focused.
While there are time when I need some visual "advertising" in order to gain my agreement and attention, with Astral Codex Ten, the concepts are interesting enough and my own internal vignettes are enough to entertain me and fully interest me.
There are certainly people who would prefer something else, but I suggest to them they may not be interested in the content and if they aren't just switch away from the article they are reading and move onto something that entertains or interests them.
I see you've added a pink background to your last comment. That can soften a message, as pink is shown to affect human anger response. So, I say, adding pink supports your message without distracting from the words which are your content. So in this I agree.
SSC felt more like a niche little community (the rationalist targeted ads were always cute, made it feel like I was in the 1920s reading an underground newspaper), whereas the design of ACX makes it feel a lot more standard.
I like the old layout more. I dunno why. It’s not a big deal though.
I also like “Slate Star Codex” more than “Astral Codex Ten”.
I was immediately reminded of this video (https://www.youtube.com/watch?v=0OB1g8CUdbA) I came across recently, featuring a DIY gadget built on Raspberry Pi that lets you plug in legacy hardware through ethernet connection and gives you a physical dial to find old web pages on Archive.org Wayback machine. The author checks out how various sites have changed over the years and makes notes of overall shifts in trends.
Now, I am broadly of the opinion that many old web pages tend to look ugly, but I don't think anyone will deny that they at least used to look more personable! However, my sense of ugliness relates to types of garishness (such as animated GIFs and scrolling text), and impractical design choices. In fact, I kinda dislike the modern "flat" designs what with their pure white backgrounds and all. The old SSC site comes from an era where it still felt somewhat personable but has the modern degree of practicality and elegance. Indeed, it's in various ways more practical, such as page width that I find more suited to desktop monitors, and clearer quote nesting in the comments, whereas Substack, like many modern sites, feels as though it's been designed for mobile first and foremost.
I don't like text on an endless white background. Even just a light gray background to frame the text makes things feel nicer and less sterile. A blog should feel cozy without feeling cluttered. Not having a mass of links and banners on individual post and instead confining them to the overview page is much better though.
I like clean and simple, but your substack is visually indistinguishable from many others. Even a very simple color scheme would help make this place *feel* like its your blog. At the moment, it's almost as if your posts are your just several amongst many on a common site (like thought catologue or medium).
There's a phenomenon called the Mere Exposure Effect where people start to like a thing more just because they're familiar with it. https://en.wikipedia.org/wiki/Mere-exposure_effect
If someone spends a bunch of time on layout A, and then they suddenly switch to layout B, they'll tend to prefer A just because it was first.
Not sure if that's the only thing that's going on, but it's probably a contributor.
Am I so out of touch? No, it's the maximizers with granular data who are wrong.
>It doesn’t make any sense at all for me to do a better job than them. So what’s going on?
ACX is a little under 2 orders of magnitude more profitable than SSC, right? Feels like the starting assumption shouldn't be that the designers didn't do their job right.
Oh, substack is well designed for making money. That is, after all, probably the only thing its creators care about.
That doesn't mean they provide a good user experience - just one not quite bad enough for users to stop paying, given the lack of competition. (Medium is worse.)
With substack, I can read Scott's articles - generally by email, because the site sucks. If I work hard at it, I can comment/read other people's comments, but without the same kind of interaction I enjoyed at SSC. (I don't even bother to go to the site for more than a couple or articles a month.) For interactive commenting, I'm better off on the bulletin board site set up by some of the SSC refugees; it lacks Scott's content, except to the extent that people respond there to posts he makes here. But the software is designed for conversation. And it's old enough to lack any number of modern misfeatures.
If Scott's primary aim in blogging is to make pots of money, he may well be going about it the right way. That was *not* his aim back at SSC, but maybe it is now.
Or OTOH, maybe you've spent too much time studying economics, and now consider money, value, goodness, and utility as synonyms, all best measured in $.
It is absurd to try and compare the competence of different actors based on their products without consideration to their differing goals and constraints. Even pretending that the metrics used here are useful ones, user experience is neither what Substack now nor Scott previously were optimizing for. It's a nice-to-have, but not something the website is first and foremost built to deliver.
Sure, we can talk about design elements and their resulting tradeoffs until the cows come home. But the question "why is Substack's design *worse*" is built on a faulty premise, and most of the purported explanations are swinging at air.
>just one not quite bad enough for users to stop paying
That's not how marginal changes work.
"Substack is run by tech industry veterans who probably hired a team of really experienced designers"
That is exactly why it's bad.
In an industry that generates so much profit that it literally does not matter whether you do a good job or a merely passable job, "experienced" does not mean "competent". It means "has internalized years of cargo cult practices".
There is no competitive pressure to make the best design - at best, there is pressure to copy the designs of the biggest companies, and those companies themselves are already big enough that people are just going to use their products regardless of how bad they look and how user-unfriendly they've become.
Nobody is going to unilaterally stop using Youtube, or Twitter, or their iPhone because of a little extra whitespace here and there, or because buttons lost their bevel, or because the website loads and runs a little bit slower due to all the extra bloated javascript running in the background. Everyone will bitch about it, but keep using the product, and so anyone who is trying to do A/B testing in these places will only get pure noise out of any measurement they're trying to make.
Fast forward a couple decades and that's how you get to the situation we're in right now, not just in web design but generally in all computer software. Everything is slower, buggier, and less usable than it should be, sometimes by multiple orders of magnitude. As a fun example, 3d video games are running enough math to compute and draw an entire three-dimensional world with tens of millions of triangles and complex interacting physics, and they're doing it SIXTY TIMES EVERY SECOND. (at least! More than twice that if you're using a 144Hz monitor). That is, they're doing it once every ~16.67 miliseconds. (6.95ms at 144 frames per second). Consider that fact, next time you open some boring 2d software on your computer and it takes a couple *seconds* to load a dozen flat buttons and images, and then you click on a menu and it inexplicably hitches for a few *hundred milliseconds*. Consider what kind of code could be written that manages to waste on the order of a billion cpu cycles, to do something we were already doing in the 1970s with computers that were at least 10000 times slower.
Software developers will often be quick to come up with excuses as to why it's actually reasonable, that everything is more complex now, and have you thought of X and Y, but there is no explanation you can come up with that explains a discrepancy this massive.
It's fully reasonable to think "these are professionals, who are being employed by very successful companies, and they are intelligent people who work very hard, so they can't possibly be doing a bad job - this must simply be the best humans can do", but it is wrong.
They are intelligent people, who work very hard, and the companies who employ them are very successful. It just doesn't matter if they do a bad job.
+1
amen
+5 Insightful
>In an industry that generates so much profit that it literally does not matter whether you do a good job or a merely passable job, "experienced" does not mean "competent".
Cool story. Now do online journalism.
There is a group of people trying to fix exactly that:
https://web.archive.org/web/20220429210634/https://handmade.network/manifesto
Join us?
(I'm linking to the old page, because I liked this version better)
>there is no explanation you can come up with that explains a discrepancy this massive
(epistemic status: i've been doing webdev on easy mode for the past three years, so my intuitions on what's exactly going on under the hood with fancy modern web frameworks may be off.)
From my experience maintaining terrible websites: the reason sites load so much slower isn't rendering, it's network latency. Unlike a game, which usually has all its assets already on your machine and packaged to load as efficiently as possible (and where a ton of work is done up front in loading screens that are frequently much longer than a webpage load), a webpage needs to wait for its requests for each asset to make its way over to the hosts, and for those assets to come back over the network, usually in an unpredictable order. Network traffic is WAY slower than code! You can't do network traffic on a graphics card! (This is why Google Stadia was an insane idea that failed almost immediately.) When something loads slowly, it's almost always because it's waiting on a network request to finish, not for render code to run. Open any webpage, hit F12, and go to the Network tab, and you'll see exactly how many milliseconds of load time each request accounts for.
...and on doing that, the first thing I notice when trying to reproduce complaints about comments taking a while to load is... am I crazy? It looks like it's *dynamically rendering and unloading every comment every time it scrolls in and out of view*, for some reason. And part of that render pipeline appears to be "making a new request for the user's icon image". *What?* You never want to put network requests in the render step! That can't be what's happening, can it? There's gotta be a more efficient way to load all those default icons up top. I can get not wanting to load long comment threads immediately, since (for the typical website) most users never scroll down to them, but this is overkill. With how much reading-the-comments happens on ACX, this dynamic on-demand comment loading might even be costing them more bandwidth.
That said, in the broad strokes, the rest of this comment is on-point. "Everything is more complex now" *is* the answer, but it really really really doesn't need to be. Tons of network traffic on the modern web is devoted to tracking scripts and ad service and anti-adblock arms race nonsense trying to obfuscate the site's operation as much as possible. Why do something the simple way when you could do it a fifty-times-more-complicated way that'll make it marginally harder for ad blockers to tell when you're serving an ad? And why let a static webpage do something when you can instead dynamically load every little thing and tell your engagement metrics about each and every detail along the way? A big chunk of "advancement" in the webdev space has been dedicated to facilitating this kind of anti-user design. Insane obfuscatory practices have become industry standard.
But yeah, back to the Substack question- they don't seem to be doing nearly as much crap as most modern sites. It doesn't seem to have any third-party tracking pixels or ad junk, presumably due to their revenue model. Most stuff's limited to their API and CDN, and there's not a whole lot of code running on the client end. It's always been pretty fast for me, though I can reproduce the issue with comment loading times when there's a lot of them, yeah. My best guess without looking at the source code is that the "floatiness" of the comment scrolling is a result of getting these icon loads stuck in the dynamic render pipeline somehow- like the same icon is a subtly different URL every time it scrolls into view and that makes the browser try to reload it, or something. And that'll vary with the user's network speed, so you have some users reporting slow floaty comments and some having no trouble at all.
-1
This is a standard whinge about software "bloat" dressed up with some numbers to make something that is not very mysterious at all seem like a deep riddle.
There is a kernel of truth here: incentives are playing a role. But the notion that the entire software industry is doing a terrible job simply because they're rolling in such fat profits that they've ceased caring about quality is just absurd.
Games are fast because the ability to render high-quality graphics without lag is what matters to the success of games. The entire history of the computer gaming industry going back to the 80s and presumably much earlier is one of engineers engaging in extremely clever hacks to squeeze every possible ounce of performance out of the available hardware. There is a reason that graphics GPUs are used for machine learning.
Meanwhile, the major constraint on the development of desktop or web-based software is engineering resources. I have been a software product manager for about 25 years now, and I would estimate that a typical launched software product contains about 10 - 30% of the features its creators want it to have. 70 - 90% of features end up cut because the resources to build them simply don't exist.
This is a number I am very much pulling out of my ass, and I will freely admit that there is a massive gray area here. In one sense, I can imagine almost infinite features for a product ("Then in phase 2, we will add an AGI to this to-do list tracker"), so of course most features don't get built. But that's not what I mean at all. I'm referring to fairly obvious extensions or enhancements to a product that will deepen the feature set, make it more pleasant to work with, or otherwise offer value to its users. I'm referring to things that the software "should" do but doesn't.
And then there is the bug backlog for the product, which in general is going to be a sucking black hole of unaddressed problems. For example, here are just the publicly visible bugs that have been recently filed against various Google products:
https://issuetracker.google.com/components
Click any "recent issues" link to see just what has been filed in the past few weeks. Some of these bugs are quite bad! My guess is that Google would prefer that these bugs not exist, but fixing them would mean not devoting engineering resources to other tasks, and so a choice has to be made. And this is Google! They have all the money and all the engineers. I gaze upon my own bug backlog and despair.
Of course the maturity of the software in question matters. Microsoft Excel has been around for literally decades, has millions (billions?) of users, and ungodly amounts of resources behind it. It is a mature product. Of course, Microsoft is still pouring tons of resources into Excel and I'm sure engineering time is still a constraint (the AGI add-on isn't so absurd in this case), but it's fair to say that Excel is more than 10% feature complete.
I'm not denying that software bloat is a real thing. Software gets bloated. Many of the cut features are bad ideas that the world is better off without. But the reason software is bloated is because getting rid of bloat is really hard, meaning expensive in engineering time, and doing so would involve direct tradeoffs with all the other things engineers could be working on, such as bug fixes, new features, etc.
So, yes, the incentives matter. But those incentives are driven by user needs, user priorities, and ultimately by what users value. In video games, lag is death -- unresponsive games are frustrating and awful. For web software, what users often want more than anything else is to pay nothing for good-enough functionality and performance. That is absolutely their right. But the outcome should not be a surprise.
Also a software engineering manager.
You've basically nailed this. For a 3D action video game, 60fps >99% of the time is the #1 developer design goal. For a blog platform, I'm pretty sure "keep latency under 200ms" (let alone 16ms) is probably not even in the top 10. Competing goals are: high uptime, billing reliability and reporting, analytics, reliable email delivery, easy to make stylistic changes without crashing the site, etc.
I mean, sure every project would also like their offering to be fast but even with tens of millions of dollars of development effort allocated to a project you still need to make tradeoffs and nobody really believes they'll lose users just because this new feature caused page load times to take an extra 50ms. Obviously, accumulated latency can add up and become pathological but then they probably have a month or two where they focus on speeding up page load times, it's probably not a core part of their DNA to keep a well funded team continuously on top of this that can veto everyone else.
Adding money is also not a panacea. Though we've made progress enabling developers to work independently without stomping on each other the last few decades, software still has a ways to go before you can just double the size of a development team and get twice as much rate of improvement. (Without even being cynical, it was the norm for a long time that adding developers slowed the rate of improvement)
Tesla is extraordinarily well resourced, and nobody would believe they aren't motivated to add improvements as fast as humanly possible, but I'm still regularly surprised (but not really) that some obvious, pure software improvements are only making it into over-the-air software updates today. I have no doubt that developers obviously thought of some of these features years ago (or users have been suggesting them for years) and the span of time between "idea" and "appears in the next release" is frustratingly long because software is still not a solved problem.
Yup. People really have no idea how vast is the chasm between what exists in a software product manager's head and what gets built, simply for lack of resources. And part of the reason that software engineers are so much more productive today than they used to be (which they are) is that they can use an extensive set of libraries and services that take care of much of the infrastructure of coding. But using libraries and services means exactly that you're going to give up on some of the flexibility and optimization you might get by coding things from scratch. At least in theory, if you were so motivated to invest in optimization.
Someone wrote and then deleted a response to my comment that read, more or less, "lol it's just a blog, if it's slow then the engineers are incompetent." What's interesting is just how wrong this notion of "just a blog" is. Wordpress is the most widely used blogging software in the world, and about 43% of the world's websites are built on top of Wordpress. Anyone who has ever used it knows that it is an insanely complicated content management system, which is about what you would expect from a 20-year-old piece of software that supports 43% of the world's websites.
Software is expensive to write and engineering resources are highly constrained. Stuff gets optimized if it needs to be optimized. Otherwise it doesn't. There's no mystery here at all.
Some data on the increase in latency: https://danluu.com/input-lag/
This is really interesting, and complements what I was saying perfectly: it is actually extremely costly in terms of engineering resources to eliminate latency, and the there is no way to justify the opportunity cost when there are so many other things to work on and so little pull from users to address these issues.
Not only that, but user preferences militate against one another in this case. Nothing stopping lower latency from existing, except you may be unable to do more than one thing at once or be on the network.
Not true, largely because your premise — that inside monopolies, individuals act with impunity — is incorrect.
While tech co's might be 'successful no matter what design is picked', this doesn't mean that the individuals making said decisions don't have a large incentive to find the best possible design.
They do: promotion.
^The incentive of 'promotion' generalizes all the way up to the CEO, in the form of performance-based grants + the (variable) value of their equity.
Further, we've never been better positioned to differentiate good from bad design. For any given metric, Substack has a large enough 'n' — and sophisticated enough instrumentation — that they can see which design optimizes said metric.
Of course, there are common failure modes associated with this reality. They generally smell like: incrementalism, near-termism, and prioritizing revenue ahead of user-perceived value.
When I was a kid this was "We Got To The Moon On A Calculator" and I don't know that the critique is any more interesting.
I think there's a trend in UI design to create things that are visually striking, minimalistic, and interactive. Sounds great! But I'd probably have used the words flashy, lacking, and cumbersome. As an example, we know that Google is not lacking for funds for good design, but here are a few changes Google has made over the years that frustrate me still to this day, maybe over a decade after some of the changes were implemented:
- Google adopted menus requiring at least two mouse clicks to do anything, while also making it impossible to cycle many links/buttons using the tab key. This gets rid of descriptive text links and compresses everything into a sleek interactive icon with snappy clickly menus. Looks great, but less usable interface.
- Gmail by default greatly expanded the margins/padding between emails, lowering the number of emails you can see on a page. It looks "clean" and "zen", but has about half the number of emails visible at a time. (Fortunately this can be turned off.)
- The new Google icons are uniform boxy rainbow line things, making all their icons so visually similar that they might as well all have the same icon at a glance. But they are vivid as anything and super recognizable for brand association.
- Google search used to have longer form descriptions for each search item, and links to a "Cached" page, "Similar pages" search, and some other features. Now they have an extremely short snippet and a hamburger menu that creates a pretty ugly popup monad that doesn't seem to have very useful information available (this is in Beta for me, not sure what others will see).
In general this oversimplifying is the trend for modern design. I imagine most "power users" that rely on a tool for better productivity don't like these sorts of changes, but I also understand that clean visuals and snappy graphics draw people in and often make it easier for new users to learn.
On Google Search — when testing this out just now, I discovered that the "cached page" and "similar page" are still there in the new beta menu, all the way at the bottom. I thought they were gone for good!
In order of importance:
0. IDK, overall I just like the old one better. I don't always understand my aesthetics well enough to fully (or even mostly) verbalize why I like them, and this is one of those times. You say you have no taste, but I also have no taste, so maybe you disproportionately attract people who have no taste — and so we all like things with color and art and backgrounds, and dislike blank white featurelessness.
1. I dislike the general homogenization of the internet. Mainly the way in which everything is now on the same couple websites, and I don't really stumble across small hidden gems anymore. But less importantly (and possibly just out of nostalgia for that time), there's the aesthetics — I dislike the way everyone from you to the National Bureau of Economic Research abandoned your old unique styles and switched to the same flat minimalistic design. Even if I did think the new design was aesthetically better, then still — I don't want everyone to have the "best" design anymore than I always want to eat the "best" meal. I like variety!
2. If I hit <space> to scroll a page down, then <shift+space> to scroll a page up (or vice-versa), then I don't end up in the same place I started. I end up a few lines higher. I am possibly the only person in the world mildly annoyed by this, but you can empathize with me if you're in this xkcd: https://xkcd.com/1271/. (In the analogy, Substack is the last row.)
3. Relatedly: I don't like floating headers. I find them mildly annoying. (I'd like JavaScript stuff everywhere if it led to more variety, but it seems like it doesn't.)
(The last two issues can be fixed with the "Kill Sticky" browser extension, but I have to click the button every time I go to the site.)
I do like the title images on articles. Even though they're sometimes too small for me to make out what the picture is.
You're definitely not the only one annoyed that substack breaks scrolling (and navigation in general)
In general, the more modern Web design tends to look slicker, cooler, and more elegant. Meanwhile, the old hand-coded Web design is easier to use. If you want to sell a website to investors, you need to impress them, and thus you go with the modern design. If you want to satisfy users, this is the worst decision you could make -- but users don't pay the bills, VCs do, so who cares about the users ?
Not directly related, but somehow connected. This 9 years old post discusses one very interesting difference between web design trends in Western countries vs. Japan. Here minimalism has become the norm, while over there sites continue be extremely information dense. The author makes some tentative explanations for the difference and why it remains, so worth considering:
https://randomwire.com/why-japanese-web-design-is-so-different/
The selection bias of most of your current readers having enjoyed your site enough to come back consistently enough to become a 'fan' is probably one part. Another is that the sleek, monetized, minimalistic design that many companies target for broadest appeal is unappealing to many seeking out substantial content due to its association with blogspam and lower quality content. While the opposite association was true of your old website (ties into first point).
For me personally, I like reading the posts better on Substack. I think that is what they’ve focused on — they sell themselves as the newsletter app (transient information). What I miss on the old site is the navigation / organization. Hopefully they will add more “blogging” (persistent information) features as they grow to encompass more blogs.
It’s much harder to find older posts on Substack, and it’s lacking the links to other blogs. This makes the blog feel devoid of context — cut off from the community of blogs. It also makes the posts feel transient and devoid of context — when I’m reading it I know it will be hard to refer back to it later, and I can’t navigate to other related posts via categories/tags.
I think the ability to add tags to posts would help a lot, as would the ability to add some persistent/pinned posts besides the about page.
It's the neoreactionaries. Reject ACX modernity. Embrace SSC tradition.
"In every art form, complicated colorful designs transition to “modern” minimalist designs over time. Whenever anyone asks, people say they hate the modern minimalist designs and wish they could go back to the complicated colorful ones."
This is the opposite of the problem with SSC v. ACX. The problem is the same as the problem of old.reddit vs. new reddit. Old reddit is cleaner, more readable, and with less buttons popping out at you. Ditto the most recent youtube update. What you have is significantly reduced accessibility and usability for competent users, presumably so they can better optimize ad sales and direct very unskilled users to the right videos.
Your readers are crazy. The new layout is way better
To me it's simple: the substack design just has way to much whitespace. When I view it in full screen on a wide screen monitor, about 2/3 of the screen is just pure white. I'm a dark-mode everything kind of guy, so the pure white light just burns my eyes. Your old blog had smaller margins and they were a tasteful grey- much easier to read. The text also looks like it's spaced farther apart than your old blog, so more white light there. Then it gets worse in the comment section, as comments nest inward to create even more empty white space.
Flakey theory: designers and visual-artists in general really love designs of pure white. I see it in a lot of trendy restaurants and bars, too, and on fashion models outfits. But I wish they'd understand that the white color on a SCREEN is different from the white on a wall. It's emitting light, not reflecting it, so it feels very different.
I've tried some, but I find it breaks too many websites (or just makes them look really weird) to be worth it. I haven't tried Sauron though, maybe I'll give it a shot.
"Meanwhile, Substack is run by tech industry veterans who probably hired a team of really experienced designers and thought really hard about every aspect of their product. It doesn’t make any sense at all for me to do a better job than them. So what’s going on?"
I'm not sure you're super familiar with how the tech industry works. We are pervasively incompetent at absolutely everything. Assuming that because tech people did it, that they have expertise, and therefore worked hard and made good choices, is absurd on a level I honestly struggle to describe.
Take the Outside View on this one.
Yeah, there may be some form of Gell-Mann amnesia at work here. Scott is always great at describing in loving detail how flawed, and sometimes downright dysfunctional, his own field of medicine and psychiatry is -- e.g. how most doctors just follow the herd in deciding which medicines to prescribe, because they lack the scientific literacy to read a research paper and draw their own conclusions.
And then when talking about a field he's not personally familiar with, such as software engineering, he apparently assumes we're all flawlessly competent professional experts. Sorry Scott, but dysfunction and follow-the-herd behavior are just the normal human condition, I'm afraid.
SSC was Web 1.0. ACX is Web 2.0. Web 1.0 is better than 2.0 for minimally-interactive sites.
Groups ... or even individuals often do things that scratch their itch, target personal goals, satisfy departmental goals, or even corporation goals at the expense of customer service—whatever form that may be.
Let's say that Scott & Michael each lead platform design teams within XYZ.com, the latest blog-hosting site. Scott & Michael each have teams competing to design a better blogging platform. Just because Scott's team delivers a better design, internal politics may select Michael's team's design contrary to the actual corporate goal of delivering the best designed platform to the customer.
Likewise with architecture. The well-connected elite public building architects aren't out to design a building which pleases the actual public, they're out to design the building which gets noticed in the elite architectural journals, which will lead to architectural society awards, which will lead to more elite white-washing, and more elite public works projects ... ala "The King's New Clothes."
So Substack probably didn't deliver the very best platform it has in it's quiver ... it delivered the best internal political solution—for better or worse.
A better version of the selection effect: People fell in love with SSC/Scott's blogging between 2013-2016 and their formative, most passionate years of that relationship are associated with a particular visual stimuli. The new visual stimulus is associated with the later, mellower years. Naturally meditation on the former induces certain warm and fuzzy's.
A good 4 seconds for OT comments to load for me
I just miss the cool blue banner at the top
Maybe it's like the interior design in like a fancy hotel or something vs. the intimacy of being in someone's home.
It's because the font in the old blog's comments is serif and the new blog's comments are sans serif.
(Also every infinite scroll piece of shit design that I hate but has proliferated across the web despite how fucking annoying it is. But mostly the serif/sans serif.)
TL;DR: humans don't like change.
This is an incredibly common problem in design. If you do a user study with people who are used to something, they'll complain about just about any style change, no matter what it is. However, when you do a user study with users who haven't used either, you'll find they hate or are unable to use the old thing.
Big tech companies have learned that the solution to this problem is to introduce changes *extremely* slowly so users have time to adjust subconsciously. For example, change the shade of a color slowly over several iterations so no one notices, or move a single button at a time.
While it is possible your old design actually was better, if you want an answer to that you must run a study that removes the "no-change bias" (made this up, not sure if there is an actual name for it). Easiest way to do that is to hire a testing service that can recruit people who have seen neither, but that is expensive.
Most of your users will eventually adjust to the new design, but it could take many years before they accept and integrate it. Some will be permanently stuck in the past due to things outside of your control, like them finding help in your early blog that was very impactful in their life, causing them to attach to that old design the same way someone can end up with a permanent attachment to a particular song or scent.
>Is it something something mobile? I put no effort into optimizing my old design for mobile phones, so maybe that adds another layer of complexity. But I think at some point some web designer friend made a version that worked for mobile, so this can’t be too hard.
Yeah, the issue is that Substack's designed for portrait (mobile), so a 16:9 PC has half the screen as eye-searing, useless whitespace (SSC doesn't fill 16:9 either, but it uses more of it and the borders are at least *not white*).
At a wild guess, I'd say SSC/ACX's readership slews more toward PC than Substack's audience in general.
EDIT: Upon rechecking, SSC uses essentially the same amount of horizontal space for actual text, but it uses a smaller font for the main posts so it still fits a lot more per line. And, again, it uses the sides for something relatively dim rather than eye-searing white.
>Yeah, the issue is that Substack's designed for portrait (mobile), so a 16:9 PC has half the screen as eye-searing, useless whitespace.
I'm surprised by the number of "obviously it's just people being resistant to change" comments given the pervasiveness of "dark mode" as an option these days. It isn't a universal preference, but it seems pretty clear that there is decent chunk of the population that doesn't like looking at their screen and seeing a sea of white staring back at them (and this is just one of many reasons people might have preferred the old site besides nostalgia).
I cannot tell you why with any confidence, but the SSC format helps me get to a contemplative/intellectual mindset.
The slick/minimalist Substack format puts me in a 'make clever comments and score points' mindset.
And for most of your content, I prefer to be in the former state. So I prefer the SSC format.
Or maybe it's just nostalgia, and I'm just rationalizing everything else.
I think some of this is just learning curve stuff. Some people, particularly those who've made an investment and climbed a learning curve, will tend to prefer higher information density and quick availability of related information, useful functions, etc. For widest appeal, though, lowest common denominator will usually prevail. Think vi/emacs or a modern IDE vs Notepad.
A well-designed blog is more than just graphic design. Experts/nerds, in my experience, will favor higher information density, while noobs will prefer not to be overwhelmed.
I should add, for my part, I have no idea what things look like on substack vs SSC, because now I pretty much always read posts via email, whereas I previously pretty much always read them via an RSS reader.
Professional web developer here. I don't love Substack's design, but I like it MUCH better than the old ACX layout. Substack lacks personality, but it's very readable. The old site was both bland and messy. I actually applied my own CSS to it to hide a lot of the elements in the old one because I found it too noisy.
Speaking as a software engineer, you're not a professional software engineer. I don't mean that as a criticism! I mean you're playing a different game. "Correct" design for you means that you and your readers all like how we read each others' words. "Correct" design in a company means hitting deliverables, which for engineers means not reinventing the wheel and instead relying on other peoples' work. Oh and the product is designed by committee. Oh, and half that committee will turn over during the project. Oh, and sometimes the external libraries just doesn't work or don't work for your use. Oh, you might have taken on too many use cases in order to please your board. Oh, serving ads sometimes degrades your own product. And so on.
The more minimal the product, the harder it is to load it down with wrong opinions. Maybe your opinions aren't as precisely calibrated as a professional's, but you don't need that many significant digits. It's words. On a page. And the constellation of ideas that those words explore.
https://idlewords.com/talks/website_obesity.htm
If someone is asking the wrong question, it doesn't matter how good their answer is. Companies aren't asking your questions, they're asking their questions. Sometimes those are close enough that you can squint and pretend they're the same. Sometimes.
On the SSC page I see your constellation of ideas on the left and a substantial piece of your content in the center, and some housekeeping on the right. On the ACT page I see a whole lot of whitespace, and your headline and your first two lines. Going out on a limb, many of us like how you explore ideas in a compounding, introspective way. One of those pages is designed to do that. One of those pages shows clickbait AND NOTHING ELSE. There's no sense that we're exploring something larger, gathering tools to understand the world. There's just the screaming void that begs a question whose answer is probably "no".
(obligatory) https://en.wikipedia.org/wiki/Betteridge's_law_of_headlines
Because, guess what, Wikipedia uses your old format too.
Generally speaking, I think Substack's design is better than SSC's, but - part of the reason I like SSC's design so much is the association with your writing, and I like reading your writing there because it just feels right. When I read SSC it brings me back to when I first found you and read some of your best posts. Substack doesn't ever make me go "oh, I'm now reading Scott Alexander" (though the writing style still does). I guess that's the power of differentiation, even if on its face there's an option that's "objectively" better.
For me, the biggest difference is that this doesn't *feel* different from any other substack. On the old site, it would open and I instantly would just know I was on SSC. Now, all substacks kind of blend together. The other day, I started reading one of your posts until I realized "Wait, why is SA talking about the electoral implications of a senate bill? Oh, wait, which blog is this?"
There are a bunch of other frictions too. The site is a bit slow. The login behavior often feels weird. The big "Subscribe now" button right above this comment box is an eye sore. But none of those are as big a deal as the fact that it has a different *feel*.
You should get data on new readers and which percentage of those prefer the new vs. old layout. (Even if they haven't experienced both in detail, they can still have opinions.) I predict new readers will prefer the new layout.
Frankly the fact that only 36% of old readers prefer the old layout seems really low to me, and a sign that the old layout has less sticking power/nostalgia than I would expect.
Basically my prior is much higher for people making decisions based on what they have more experience with and much lower for people evaluating the designs "objectively" in any sense.
While we're at it, I really hate the new minimalist LessWrong design too! I miss the greens and dark greys!
Someone may have said it already, but the main reason why I prefer SSC’s layout is that Scott’s old posts are often worth re-reading, and the Substack archive page makes that harder, one has to scroll down for ages to get to even the not very old posts. That’s why, even though I discovered ACX first, I still have a fairly strong preference for the old design.
Every website is always getting worse, which I attribute to optimizing for mobile. (I never read anything on mobile devices.) But then other people on this thread say that SSC was better specifically because of mobile, so shrug emoji.
I just want more blue. It's too bright now.
For me it's almost entirely nostalgia. There was something warm associated with the old design. I've spent years and years reading to it. I've read so much to it it's become engrained.
I loved the old name. Slate Star Codex felt arcane and potent in a way i feel didn't translate to AstralCodexTen.
I dont mind the substack layout nearly as much as the godawful site performance. Even on my gaming PC, let alone on my (admittedly rather cheap) phone, I have noticable delay when I scroll over the comment section. And I am 100% sure the problem is caused by abnormal CPU load, not bad internet.
Very simple thing: I like being able to scroll through multiple posts on a single page. There is a continuity to your blog that is totally lost in Substack.
When you do a series of posts on a similar topic, in SSC I can easily view all those posts in chronological order as well as see what other things were on your mind around the time you were posting them. On Substack I have to navigate back to the homepage, then navigate to the archives page, and instead of being able to ctrl+f the title of the post, I have to scroll down, wait for it to load the next set of posts, scroll down some more, wait for it to load, etc. etc. etc. until finally I find the title of the post and I can click into the next one. Godforbid I ever want to read through a series you wrote over a year ago.
I believe it has to do with the authenticity of SSC. It possessed a certain rawness, call it an early web2 aesthetic. SSC had your finger prints on it and it felt substantially more intimate. It’s cause for nostalgia now because we’ve all been drowning in a sea of refined homogeneity. Refined design (ACX) and raw self expression (SSC) connect with users in two entirely different capacities.
One big thing for me is that SSC has a nice text list of all blog posts, sorted by date and easily searchable. Substack had a dynamic list that loads more as you scroll down, meaning it's hard to find a particular older post if I want to refer to it. I usually just end up using Google if I want to find a particular ACX post.
For me, this is about feeling like you're in 'Slate Star Codex land' vs 'Substack land'. The former is far preferable, as the latter is unfamiliar, but also large and anonymous (I don't feel like I know anyone here; I don't know the 'culture', etc.).
For Facebook vs MySpace, the common layout makes Facebook feel like we are all a bunch of people interacting in Facebook-land, whereas MySpace felt like a bunch of different sites that were loosely connected together. If the goal is interaction with other blogs, maybe Substack is preferable, but for a cohesive readership and community built around the blog, I think a distinct identity is better.
On Substack all content looks the same as all other content on Substack and, indeed, all content elsewhere in the internet. Contrast with the SSC design where you as a repeat reader know right away where you are and what you're doing there. It's homey.
For me one of the major differences is that the old design had some color, and then filled the unused parts of the screen with a nice dull grey. Substack just has white, which is like staring into a lightbulb.
I think this is mostly a 'music was at its peak precisely when I was a teenager' thing. SSC was what what a lot of blogs looked like when your readers were in their formative years – in fact your blog may have got them into reading them, so that's where their idea of what a blog 'should' look like formed.
I'd also probably say that a disproportion of your readers, including myself, read on laptops, while most of the public does things on mobile. Old blog design was laptop-optimised, modern design is mobile-optimised.
The substack font is just bad. It's so unpleasant to read. But bad fonts seem to be "in" at the moment.
The typeface here is gross and the text is (way, way) too large. The comment section code is also pure satanism. That's pretty much all it is for me, anyway.
I dislike the fact all substack pages look the same. I associated the blue with SSC, green with Marginal Revolution, etc. It allowed for some different mental context switching. Now, all substack bloggers blend together.
For me, the biggest thing is how SLOW substack is. I don't want to wait 5 seconds to load an article, and then wait 5 seconds to load comments. When I see a new substack, I don't want to click "let me read it first" then wait 2-3 seconds to find the 'top' tab.
Same reason I prefer hackernews over new.reddit, even though hackernews is purposefully made ugly. It's just quicker. I don't need fluff, I just want it to be responsive.
Perhaps this is more like indie vs. corporate.
Your old blog was charming because of your amateurishness, not in spite of it. The fact that you designed it personally means that it had some gestaltic connection to your personality and community. The user feels this connection without being able to explain it. This is the charm of the old web.
I imagine Substack is designed by one or more teams of product managers, designers and engineers who apply established principles and careful experimentation to maximize user satisfaction and engagement. Is it any wonder that we don't vibe to it?
To me Substack feels like a corporate lounge. Sleek, clean and coldly elegant, but not exactly homely.
I don't think that Substack could do it any different. Authenticity is a primitive. You cannot engineer it. Although some have tried (perhaps the history of punk is relevant here?).
> This may be a little too cute, but I can’t help but think of Whither Tartaria? In every art form, complicated colorful designs transition to “modern” minimalist designs over time. Whenever anyone asks, people say they hate the modern minimalist designs and wish they could go back to the complicated colorful ones. But for some reason nobody ever does.
Isn't this what happened to Reddit? For the record, I joined Reddit when the new design was already in place, and I am not able to use the old one as I find it confusing.
I've discovered your blog after you moved to substack. Since then, I've read basically everything you've ever written. As someone who started with your substack stuff, I'm honestly indifferent. I guess I slightly prefer this, but it is a lot more sterile.
One thing I vastly prefer about the old layout is the tag system. Having to trawl through your posts manually to find your fictional short stories really sucks.
Hell. I like your image of the original SSC design better than the 'improved' one. Sidebars are evil, your rendition only had 1. Every time I read ACX my eyes are stabbed by these massive gutters of white on either side of the content. It's wasted space, it's aesthetically ugly, it's hard on the eyes (dark mode at least helps with this one singular complaint), the only purpose it seems to serve is to make the page longer by squishing the text into a newspaper column, and newspaper column text is hard to read.
Obviously this is all my own opinion and I seem to be in the minority if the prevalence of space-wasting empty white boxes in modern applications and sites is a halfway reasonable indicator. But for my part all these 'experienced design veterans' can go stuff themselves someplace ugly. Give me back my information dense interfaces. (though yes, you can keep all the equally eye-gouging flashy distractions)
To me it seems crazy to spend lots of time web surfing on a mobile device because the input mechanisms so far inferior to a desktop that it slows down everything you do and requires all the designers to dumb down their interfaces. If I'm away from home, it's almost always because I'm doing something that is NOT surfing the web. So all these websites are being optimized for a use case that should barely even exist imo.
I have no preference. I spent lots of time over the last seven years reading SSC as ebooks, text files, or Reader mode renderings, or in a heavily customized browser. The comments were akin to marginalia, not a conversation. So I don't really care about how it looks (and reading the comments here, there clearly wasn't a consistent user experience). However, reading ACX in the Substack app is the first time I was motivated to register and post instead of lurking, even while both the old and new comment systems seem worse to me than a threaded NNTP client. Substack reduced the height of the first step on the engagement curve and nudged me to take part. Is this a good thing when there are already hundreds of great contributors here?
The "wading through water" feeling aside...
I hate the modern bullshit. ALL the modern bullshit. All the stuff that slows things down and eats up screenspace. If I wanted to exaggerate somewhat, I'd say for me the best internet are the "print versions" of websites, and that I wish I could go back to browsing in Lynx. You get the idea.
Like the bar with the name of the blog popping up when I scroll up even a little. Who needs this?
I'm on mobile just now. That pop-up is the only way that I can get from the comments to the homepage and from there to other posts.
I meant the desktop website. But in your situation I'd just use the "back" button, unless I hadn't originally opened the post from the homepage.
Come to think of it, I imagine if I looked into it, I would find a substantial amount of my peeves with modern websites to come from influence by mobile phones, similarly to consolitis in PC games.
Another theory is it's the distinction. SSC looked like SSC, astral codex 10 looks like every other substack.
I don't know if it's just different things are interesting or if it's a conditioned place preference kind of thing. Where you prefer the look of the website that gave you a decade of mind blowing revelations.
The old design was a product of its times. Sure, You, Scott, were not a web design specialist, but you were not writing the site from scratch, you were just putting together things created by people who were. And those people worked in a different environment and optimized for different things than the modern web designers optimize for.
I think other comments have already conclusively explained all the object-level technical reasons for which modern web design, of which Substack is but one example, is extremely awful. No need to repeat them, so straight off to meta I go.
The process that led us here can be framed in several ways, largely complementary, but emphasizing different aspects of the dynamic. Among them:
1. Desktop vs. mobile. Mobile designs optimize for their small screens and crappy input devices by simplifying user controls. Hence, thin columns, simplictic UI, dynamic loading / infinite scroll, and so on, and so on, dozens of people described it in detail already.
2. Customizability vs. intuitiveness. Or, put differently, optimization for experienced/extensive users vs. optimization for newcomers. Up to a point, newcomers aren't going to care about the loading lag or inability to do a simple text search over the entire page, or other technical problems that only start bothering you once you're a regular. The best option for catering to them is a maximally simplified site that showcases everything that's supposed to draw them in, while leaving aside complex options that might confuse them, no matter how helpful they become for the regular user. Also, there's no need to optimize for speed or responsiveness, because newcomers didn't yet learn enough automatisms to make those lags slow them down. Iterate this race to the bottom over the entire internet, and you get the modern web experience.
3. User experience vs. corporate control. Once, internet services competed for users by providing them with better tools to reach what they, users, wanted to see. Nowadays, they mostly enjoy natural monopolies of scale, so they optimize for forcing their captive users to see what they, corporations, want them to see. This isn't really Substack's goal, but again, their expert designers use paradigms evolved in internet-wide race to the bottom, so here we are.
4. Sophisticated vs. casual users. Internet was once a niche of intellectuals, professionals and hobbyists. Now, it's an everyday necessity aimed at a common person. It's not just that some options are taken away from the user, it's also that a majority of the users will increasingly have no need for them and not demand them back, leaving those who do want them a minority not worth catering to.
My suspicion is that Substack is optimized for two distinct use cases, one of which is e-mail. And it's the e-mail experience that allows frequent users maximum control at the level of their mail client, while the browser interface is basically an advertisement for the core, e-mail product, aimed at the outsiders. Great for people still employing habits formed in the days of usenet (assumedly, someone who actually reads substack newsletters in the mail please confirm), but bad for those of us who gave up on e-mail as an everyday tool and wish for an optimized experience in their browsers, because we're getting a laggy, dumbed-down, hard-to-efficiently-navigate version designed for complete newcomers.
Substack does not index comments. It is impossible to Google "site:astralcodexten.substack.com JohnWittle" to find comment threads I replied to, or whatever, the way that I could on SSC.
Instead, I am forced to use Substack's notification system... which only shows me comments over the last month. Older comments seem to just disappear into the void, unless I want to manually ctrl+f every single article
Frankly this is horrible. I can't count the number of times I wanted to look up an old conversation I had, and it was easy on ssc. On substack so far if it's older than about a month, it's been impossible if I couldn't remember the exact context.
This: the inability to search comments, and the lack of archives, means that the long-term usability is much lower on SubStack.
My theory behind why Substack (and other companies) do this is because if everyone made an optimal design there would be no non-textual way to tell websites apart. So by making your design a pain in the ass to use your website becomes more memorable and recognisable and you are able to build brand recognition. This also explains why customisation is not allowed: it subtracts from brand recognition.
Brand recognition is probably very helpful in extracting money from venture capitalists so it doesn't matter if you're pissing off half your users.
Designers are like economists: very handy when it comes to the small stuff but just smile and nod and then do it your own way when it comes to the big stuff. They are taught a lot of theory that changes how they see the world, but they then become very convinced their profession is always right and in the process lose some common sense.
This is really a small thing, but you put a picture next to each post title (presumably because substack forces you to), so I click through because I want to see the picture bigger and in context. But most of the time the picture is not present in the post.
> If they forced everyone into the standard non-customizable layout of 2015 SSC, would that be a straight utility gain?
Yes.
I would say the primary difference is that the comments on SSC look fine and the comment layout on substack is terrible.
Secondarily, the gray on SSC is nicer than the bright white on substack.
As to Tartaria, it's worth mentioning the Flintstone House. https://en.wikipedia.org/wiki/Flintstone_House
(I feel like this has come up before, but I can't find it. So I'll assume it's new to ACX / SSC.)
The house is markedly different from other houses and is decorated in an idiosyncratic way that I assume pleases the owner. The owner is also mired in legal battles with other somewhat-nearby homeowners who object to the idea of the owner decorating her house in a way that she likes. And while I believe the battle is unusual, the general legal background is not. If people were allowed to decorate their own houses, design their own buildings, etc., they might have better-looking houses, buildings, etc.
Substack has no personality or identity. It’s plain and sparse and feels very generic. Scrolling through it feels like it could just as well be showing me listicles or recipes.
I prefer Substack to Slate Star Codex for reading individual articles. The text is bigger (I have to zoom SSC at 150% to be able to read it easily), it's less "busy" (just the text, no sidebars).
On the other hand, visually I prefer the SSC. It may not be the best design ever, but it's "quiet" (I don't feel "attacked" by the colors), and more importantly, it's yours. When I read a blog article on SSC, I know that I'm on SSC by the writing and the design. When I read a blog article on Substack, you have the exact same design as everyone else. It's like being at an independant coffee shop instead of at a Starbucks. Sure the coffee shop may be a bit messy and not to your taste, but it was made by human beings, it's unique, and it says things about them. It has meaning. Starbucks on the other hand are a product of a company. They have no "soul".
Another thing is that Substack uses the serif "Spectral" font (or a backup serif font). Most people prefer sans serif to serif fonts for screens. Substack may give a better impression of "authority" (serif fonts reminds people of printed stuff, newspaper), but the font on SSC is more welcoming (and easier to read). It also, in my opinion, goes better with your writing.
Probably just change aversion. It's well known that people who have spent a long time using some design will react negatively to any new design, even objectively vastly better ones. Loss of familiarity, loss of instinct-level knowledge one has built for using the site...
Try surveying only the people who started following you in the ACX era, and you'll get a very different answer I'm sure.
The old blog was less easy to scroll through, but practicality isn't everything. This one feels less you, less like reading what a specific person wrote and more like the hundreds of sanitized other blogs we stumble upon all the time. I got used to it because your ideas are interesting and they make up for it, but for me it's like the difference between a cozy living room and a modern office with white everywhere.
I am curious if you still have this opinion in 2024, now that the amount of content on ACX has grown
Do you really think that finding a 2-year-old post on ACX is easier than finding an equivalent post on SSC?
Substack on desktop feels like a badly adapted mobile app, it just doesn't use all the real estate a wider screen provides, while the old design did. This is a general trend in webdesign, almost everything is made with mobile in mind so websites are tall rather than wide with lots of empty space on the sides. I know some people think that's a better for aesthetics, but it just make everything feel like twitter/instagram.
For those who want to take matters into their own hands with some CSS: https://applieddivinitystudies.com/slatestarsubstack/
There's like... actual utilization of most of the screen on the old layout. The new layout is a center column of text with literally over half the screen space completely empty.
So yeah - I think it's a mobile thing. Websites used to be designed for 4:3 and later 16:9 screens. Now they're designed for 9:16 screens and look stupid on 16:9 screens.
(and I should note I never read your old blog, so I have no nostalgia for its layout - just nostalgia for websites that were actually designed for the screens they're being displayed on)
Is there though? I just went over to slatestarcodex.com, and the central band of text seems to be the same width as on substack. Near the top of the page, there's sidebands with content, but that cuts out pretty quickly.
One of the things that's been learned in web-dev that turned out to be both true and useful was that text areas should not be super-wide. You want it to be easy for eyes to skip to the next line in a paragraph.
But the optimal width is unlikely to be exactly the same for everybody.
The original philosophy of the web was that you'd send basically plain text to the user's browser with only some minimal markup hints, and leave it up to the browser to decide on the details of how to render it. After all, the browser knows more than the server about the local device's capabilities and about a specific user's preferences.
E.g. if the website would not enforce a text width at all, then it would look ugly when viewed in fullscreen, but I could easily choose the optimal width for myself by simply resizing my browser screen.
Unfortunately, we've been steadily moving away from that, and now web designers insist on controlling what their site will look like on the user's device, down to the last pixel. Except that they don't want to actually customise it for every possible target device and every possible user, so you end up with a one-size-fits-all design.
It is still possible for users to override the choices imposed by the server, via custom CSS. Several custom stylesheets to make ACS more usable have been posted in this thread. But that's a hassle, and it needs to be done for every website separately.
Sure - and I suppose beyond the sidebars, the difference is mainly in the background color on SSC not being harsh white across the entire screen. The parts of the screen that aren't being used aren't bright and annoying.
Caveat: I only ever use Substack's UI when I'm trying to debug or fix something. I always read from my own specially-written UI [1] and on desktop.
The #1 problem with the old site was that posting a comment reloaded the page. Substack fixed that. But it's otherwise doing too much . . . *something*, I don't know what, that slows everything to a crawl.
The old site was simple but it functioned. It got stuff done. My desktop wouldn't slow to a crawl because I had 3 tabs open to it.
[1] https://github.com/EdwardScizorhands/ACX-simple but I haven't put out a release in a long time. I haven't implemented "REPORT", for example.
I would say the old layout was a bit too cramped, cluttered, and quirky. But Substack is a bit too bit, minimal, and bland. I expect there's a happy medium somewhere inbetween.
Clicking on a comment link takes me to that comment maybe 20% of the time.
Emotional attachment. "I liked their early stuff." We attach to an aesthetic when the content moves us.
I've been a web developer for a couple of decades, and it is *very* rare for a redesign not to inspire a backlash among regular users. It's not even rare for someone to be mad at a redesign, get used to it, forget they used to hate it, and then defend the redesign when another redesign comes along.
Obviously, some of this isn't rational, and it's just discomfort with change. But to steel man that a bit, it's often the case that people get good at using really clunky sites/features, and that continuing to use clunky-but-familiar things is, at least in the short-term, less work than acclimating to something better-but-new. It can also be the case that a redesign makes the site MUCH better and more accessible for more people (like with mobile) at a relatively slight cost to existing users who don't care about that.
Primary point being: even if a redesign is obviously better, some people will always be upset. But most of them will get used to it.
There's also the effect that your old users were selected to be those who liked the old workflow best.
Maybe users on the web overall prefer social media integration[1], but your users would tend to be those who didn't need or want that.
[1] monsters, all of them
Good point.
You've just made me realize there's a sorta-kinda overlap with this and the "considering the good of future people who will exist" moral calculus that factors into some social and political questions. Web developers have to balance the wants of existing users with the theoretical good of additional users who have/would bounce off the site without certain changes.
> It's not even rare for someone to be mad at a redesign, get used to it, forget they used to hate it, and then defend the redesign when another redesign comes along.
That could also be explained by the second redesign being objectively even worse than the first. Just because users defend version 2 against version 3, doesn't mean they were being irrational/inconsistent when they said they preferred version 1 over version 2. E.g. if users hate stupid unnecessary Javascript crap, then every time you add another piece of stupid unnecessary Javascript crap, your users will now hate it *even more*.
As for "making the site better for mobile users at a small cost to desktop users": that is indeed exactly what a lot of people here are saying, that Substack appears to be optimised for mobile phones and now it sucks for desktop users. The only disagreement is on how "small" that cost is, and how reasonable the trade-off given that possibly the portion of desktop users among SSC readers is larger than elsewhere. A design which is optimal for reading a three-paragraph Facebook post on your phone, may not be optimal for reading Scott's posts, many of which are thousands of words long and require some uninterrupted focused attention to fully digest.
And anyway, it doesn't need to be a tradeoff! Detecting whether the user is on mobile, and serving them a different version of the site if they are, isn't rocket science. Yes, it means a bit more work for the Substack devs, since they now need to maintain two versions of their website. But the question wasn't "which design is most convenient to maintain for the Substack devs" but "what do the users prefer".
> That could also be explained by the second redesign being objectively even worse than the first. Just because users defend version 2 against version 3, doesn't mean they were being irrational/inconsistent when they said they preferred version 1 over version 2.
True, though occasionally I see them defending the very things they said were bad, not just as better, but as *good*, and with no clarification like the above about going from bad to worse. That is a logical possibility based on the vagueness of my original comment, though not something I think I've observed in practice.
To add some detail: I had an argument with a user about a mobile change awhile back and they made demonstrably false (and sometimes even mutually exclusive) complaints about it from post to post. At first I foolishly tried to engage each point, but in retrospect it was obviously just someone trying to say they didn't like it and coming up with reasons post-hoc. I think this happens a lot in some communities, at a level and frequency that might be shocking to a community like this one, which is more analytically rigorous than most. To that point, I run some sites for people who don't know the first thing about development and are not even particularly web savvy (most of them never complain about "JavaScript crap" because they have no idea what JavaScript is). I don't look down on them for this, they have different interests and priorities than I do, but I have to constantly remind myself that they're having a totally different Internet experience than I am.
As for mobile, I don't find it hard (or harder) to read on desktop. Quite the opposite. I'd describe Substack as optimized for reading in basically all formats, but that's just me.
As for managing two separate sites, I'm not sure how viable it is. I did this (and still do, out of tech debt necessity for the time being) for one of my sites and boy, is it inefficient and frustrating sometimes. Like, a lot. I definitely get why responsive design is all the rage. I'll preemptively acknowledge it can be done in a very lazy, bare-bones sort of way that's less about making mobile functional than it is about making desktop excessively plain, though I think even in that form it's a pretty reasonable fit for a platform that's ostensibly just about reading.
I would agree this needs saying, but I also tend to think it gets oversaid by designers -- who too often mistake grumpy acquiescence for acceptance. People are not forgetting they hated the change (or dislike the new), they just don't have enough energy to keep up the fight, they've got other stuff to do (particularly if it seems no one is listening), so they just acquiesce.
If you haven't read it, I highly recommend "The Design of Everyday Things" by Don Norman. He makes a very good case that *outstanding* design doesn't even *have* a learning curve. It's so intuitive new users are just as competent as old hands, and nobody is even much aware of having to learn something. For example, nobody who grew up in England and goes to China for his very first visit out of country and has to ask his hosts how to pick up a Chinese teapot versus the English teapots he's seen before. How you do it is so obvious it not only doesn't need explaining, the user is not even *aware* of figuring it out.
He also of course has a lot of very funny and sobering examples of everyday things that are *not* designed well, e.g. doors that must be conspicuously marked "PUSH" and "PULL" because the design fails to make it obvious which you should do.
Anyway, it persuaded me that in a fair number of cases -- although I would hesitate greatly to say most or all -- observations that of course the user just has to accept some period of clumsy apprenticeship and conscious study to use a modern contraption may just be a load of excuse-making for actually shoddy design.
I think you're responding to this?
> it's often the case that people get good at using really clunky sites/features, and that continuing to use clunky-but-familiar things is, at least in the short-term, less work than acclimating to something better-but-new.
I think this coexists with the idea that good design has no learning curve. Even if we posit that there is such a thing as perfect design, and that with it new users would be exactly as competent using it as old ones, I think that holds only when the hypothetical user has not used some version of the design before. Overcoming *habit* is another thing entirely. Reminds me of the "debate" about which way a toilet paper roll goes.
I never read SSC and I prefer the it to the sterile, wall-of-white, no-different-to-any-other- substack-blog ACT
That's perfectly reasonable. My comment is that some people will always dislike the new thing because they are accustomed to the old thing, not that everyone who dislikes the new thing does so for that reason.
I personally like each blog I read having a slightly different feel. Makes it more than just words on a page. So a bit of individualized layout is a plus.
I think the touch of colour from the blue also helped a lot. Imagining the current layout with a bit of the old blue, I'd consider them basically equivalent with that change. White background, black text wears thin fast.
https://xkcd.com/1172/
What I don't like in ACX design is how difficult it is to browse old archives. The SCC style, where they're all on the same page, easily searchable and overviewable and archive-diveable, is much better than ACX's infinite scroll that has you start the scrolling down all over again if you lose your place and refresh the page.
The reason old SSC's layout is appealing is the same reason a crowded, boisterous Hong Kong alleyway can be more appealing than a sterile shopping mall.
As counterintuitive as it is, the new substack layout is just a bit... sparse, white, plain. This is usually a good thing for reading, but I think there is a degree of idiosyncrasy people seek out in websites. SSC's layout (I prefer the pre-touch up image, funnily enough!) isn't a tour de force in web design, but it's... idiosyncratic. Recognizable. Perhaps unique?
Going back to the Hong Kong analogy, SSC just had way more going on in the margins: neon signs for tech startups and EA groups, merchants hawking their comments in the open thread, and a blogroll of some of the most revered literati of their small circles posted in the town square like a bulletin. I don't remember all the signs and shops and posters, but they add a bit of flavor and "placeness," as opposed to the blank white margins that greet me on this very page.
My pet peeve with Substack is that the mouse cursor doesn't turn into a text cursor when I mouse over text. I didn't really think about it before, but I must subconsciously point to text with the i-beam cursor as I read!