There are ‘luxury’ versions of seemingly everything, from jewelry to watches to mineral waters (complete with ‘water sommeliers’ & tasting bars) to things as prosaic as a business suit or automobile. These can be justified by their frequent use: if you are going to drive a car every day, why not have a very nice car, a much nicer car than you need in any ordinary point-A-to-point-B sense—just because cars are a hobby and you enjoy and appreciate an excessively nice car, and thinking about the obscure technical points like what kind of engine it has or how fast it can accelerate 0→60? So this makes me wonder: is there such a thing as “website luxury”? Not merely websites for other luxury things, or websites about ‘luxury’, which all obviously exist, but websites which are themselves luxurious. We spend so much time looking at websites, and web development certainly involves many challenges and cannot be said to be ‘solved’. Like mechanical watches, websites could involve all sorts of arbitrary constraints or goals, like making a mechanical watch as thin as possible, or accurate as possible, or having more “complications”. It also can’t be because websites are too easy to copy—they are often quite difficult to copy or replicate. (Gwern.net comes to mind: a number of people have admired it enough to try to copy it, but invariably, they replicate only the tiniest, easiest subset of the design—despite it being 100% FLOSS and available on Github!) Nor can it be because there is inadequate money: web development is an enormous sector of the world economy. Nor even could it be for lack of interest in websites by rich people: people who buy luxury watches are not necessarily that wealthy, and anyway, there are plenty of billionaires with personal websites, which are uninspiring. Bill Gates’s personal website seems important to him, given how he posts book reviews to it, but it can only be described as ‘exists’ (despite not doing much beyond a simple blog, it has severe performance & layout shift problems); Patrick Collison’s personal website is stylish, but more remarkable for its commitment to minimalism and what is not there (ie. all the horrors of the modern web like cookie banners or ads); and most billionaires of course do not have a web presence beyond social media, if that. (A possible exception that proves the rule: Elon Musk lost something like $15b after he bought Twitter—surely a website luxury for him! Even Larry Ellison struggles to spend so much so extravagantly on personal consumption.) It is true that most billionaires do not dabble much in software—as far as I know (between my reading and asking a few), pretty much no billionaire has their own shop of programmers making custom apps for them as the only user, and they use off-the-shelf software like the rest of us—but this has more to do with the power of just hiring a good human assistant and the difficulty of making custom apps truly competitive with available apps which may be developed by teams of thousands of programmers (and with the difficulty of making a tool which can pay back the effort of designing & learning a custom tool, similar to the difficulty of ‘tools for thought’). It is going to be hard for a programmer to make a better Gmail for your Android phone than Google can make, or at least a full startup like Superhuman. But this doesn’t apply to a website: Google is not making a website about or by you.

First, what is luxury to begin with, beyond simply “expensive” (whether in money or opportunity cost)? Some signature traits of luxury goods as I see them:

• perfection (to evoke the perfection premium)

• multi-person, or at least full-time

• more than mere “professionalism”

  Being “professional” is not “luxurious”; they overlap, at most. Being professional is about predictability, not perfectionism or style.

  Website-luxury is not “professionalism” either. One might say that “luxury” starts where “professional” ends. Many websites are designed for, by, and about professional purposes of many kinds, but the word ‘luxury’ would never pass one’s mind. The essence of professionalism is a kind of reliability, predictability: a professional is above all predictable. The official standardized peanut butter from NIST is professional because it is the standard, not because it is good peanut butter (it apparently is not).

  You know what you are getting. They will not suddenly decide to boycott you over South Africa, nor impose their personal preference, or aspire for individual glory and to make the current project the greatest X ever. Instead, a professional is a consummate blackbox, which can be slotted into a larger scheme. A functional organization is made of professionals because unprofessional behavior, no matter its merits, cannot be reliably combined by the crude mechanisms of human organizations. You cannot build a car if every individual screw, bolt, wire, or microchip has its own opinion on divestment from Israel, how the car needs to try to set a Nürburgring record, or actually wouldn’t purple be fetching this time of year…? A professional is someone who is happy to check the tolerance of a bolt for the tenth time because that bolt holds together airplane wings and that is how many checks it takes so the rest of the airline system can assume “the wing won’t fall off the plane” and get on with things.

  Professional things emphasize reliability, but only as part of cost in general (because if something is predictably unreliable, that is almost as good as being reliable, and indeed, professionals might make things deliberately unreliable to ensure that no one accidentally counts on them too much), while something like a supercar might spend more hours at the mechanic than being driven—but it is fine if a bunch of bolts need to replaced every 500 kilometers at exorbitant cost, as long as it is every 500 kilometers and never, ever, in the middle of a turn at 200kph on the racetrack. (You can’t enjoy luxury if you are dead.) Similarly for things like mechanical watches—they may need regular maintenance or to be wound up every month, but if that has been done, then they must work.

  Thus, ‘professional’ does not mean luxury, although most luxury things will depend on professionals and be in many ways professional. But if a luxury thing is only professional, then it has failed and cannot justify its cost or claim to luxury. Luxury must go beyond.

• goes beyond craftsmanship, to be more than one craftsman or hobbyist would do

• subtle easter eggs or quirks

• advanced technology/materials and over-engineering

  • often valorized in the marketing

  • but only specific parts; tends to emphasize either total transparency (so as to impress the nerds) or total obscurity and an emphasis on “It Just Works™ magic” (even if often ‘just’ a lot of hard work)

• anti-community

  Luxury goods are not intended to create a self-sustaining community, which might do things like go off and ‘mod’ it or even make their own but to create a steady stream of buyers. Hence, things like punishing buyers who resell goods or who modify them or ever compete.

• unique but usually understated esthetic

  (If the current luxury ‘meta game’ is not towards understated minimalism, as it currently is in the West in a trend of “quiet luxury” like shibui and post-Xi Jinping crackdowns in China, then the esthetic is usually overstated: over the top, nouveau rich, flamboyant, unmistakably extremely expensive in terms of labor or materials etc. But luxury is rarely ‘mid’—because then how would anyone know?)

• opposed to mass market success

• whale customers

  Hermes doesn’t make its money on people who are buying 1 scarf or a shopping bag. They make their money on the wealthy women who might drop $100,000 in a single shopping trip. (I am reminded of my family’s Vietnamese exchange student who came from a very wealthy family, and regarded our trips to NYC or Las Vegas as an opportunity to stock up on luxury goods, and our middle-class shock at the sums spent.)

• brand recognition and aura… but illegible

• often customized (frequently as an excuse to make it more expensive via labor consumption and reduce commoditization)

  Particularly as one moves to the high end, or away from goods per se. For example, what makes super-yachts so luxurious is apparently the intense staff:guest ratio, unmatched by any mere resort or five-star hotel, allowing for personalization and unique experiences.

• broadly: what “premium mediocre” tries to evoke; one can deduce “luxury” from where premium-mediocre feels shabby or pinchbeck

So, what would ‘website luxury’ luxury look like, if we tried to imagine “what if Hermes sold websites?”, “if Birkin bags were instead websites”, or more pointedly, “what if Apple did websites as seriously as it did devices (and didn’t hate websites as an existential threat to its cash-cows like the iPhone?)”

Some examples that come to mind are: the Nintendo family foundation website, McMaster-Carr, B&H Photo closing on Saturday for Jewish Sabbath, Low Tech Magazine website which only works if there’s enough sun that day, 37signals web apps like Basecamp, web developer showcases like Steven Wittens. Anti-examples: almost all newspaper websites; social media like Facebook or Twitter or TikTok; Medium.com (ironically, in betraying its original goals).

These are website luxury by conspicuous consumption, by costly signaling, by a refusal to compromise certain principles, and by technical excellence (or at least distinctiveness—Low Tech Magazine’s technical choices are not that good even at its stated efficiency goals).

The Nintendo website is an excellent example: there is no need whatsoever for a charity to have such an elaborate in-browser 3D animated explorable world, and it presumably cost a substantial sum of money to hire people capable of that (how often do you see such 3D animations in websites?) to execute it to the standard that they did (the longer one watches, the more one sees and is amazed that there’s so much stuff in it, to an extent that a hobbyist or amateur would struggle to match with their limited time & resources), but it is incredibly cool and memorable. Hence, it is website luxury.

Taylor Troesh’s “cheap” web manifesto made me start thinking about website luxury in the context of asking myself: “What sort of website is Gwern.net at the end of 2024?”

It was initially a hard question to answer because the existing labels are all inadequate in different ways: While I sympathize with Troesh’s idea and like many websites that could be described as “cheap”, my own site is clearly not “cheap”, as it violates at least 4 of the guidelines. The classic “blog” term is a gross misnomer (entries aren’t even reverse-chronologically sorted!), while “homepage” is barely used and hopelessly undersells the website—it would be like describing the Encyclopedia Britannica as “a book”. It’s not clear “minimalist” is a good alternate description: while there is little waste and minimalism is a north star, there is much that could be removed without destroying the essence. (We are a long way from 80-column ASCII text files or Utext or Gopher.) “Static site” also doesn’t fit: while static website design imposes a certain esthetic, we’ve moved on to approaches which are highly dynamic (and arguably it’s not even technically ‘static’ any more due to the use of SSI webserver templating). Robin Sloan’s “home-cooked app”? Great idea, but not even close. How about “digital garden”, as a variant on personal wiki or PKM? Not too bad, there’s a definite overlap in the type of writing and in the small touches like link-icons—but most digital gardens do not seem to feel any need for Gwern.net priorities which absorb vast amounts of my effort, like fulltexting links or repairing linkrot, or in extensive lints & tests or automatic rewrites or careful copyediting. (Notably, when people try to clone the design of Gwern.net, they typically omit the link-archive system completely, and if they do link-icons, they settle for 1 or 2 icons, rather than the several hundred necessary to be comprehensive.) And hobbyists often delight in incomplete or unpolished content, while I regard those as a grudging concession to my personal limits.

The “small web” or “indie web” is a mix of “cheap”, “digital garden”, “minimalist”, and just various “punk” dissidents and experiments; your average Neocities page or anything listed by Kicks.condor feels different enough from what I am aiming at to be unhelpful. (There is something more than a little ludicrous at describing a website which is >103GB in 100k+ files and uses >78kloc, and is almost a full-time job to develop, as “small”. “Indie” scarcely seems more apt.)

But I think you could describe Gwern.net as a website luxury, as it satisfies most of the signature traits: we aspire towards perfection and try to squash every bug (hence all the lints & checks), as we take website improvements far beyond what could be justified by traffic or possible Patreon revenue, it requires multiple people (myself, and Said Achmiz for frontend), it is not merely professional (else I would be doing all my writing on LinkedIn, perhaps), it includes plenty of easter eggs & quirks which make a Gwern.net website page distinctive and recognizable at a glance, is not attempting to be a forum or social media and exists mostly for one-way interactions like publishing, definitely “unique” and many parts are subtle (eg. link-colors), is funded mostly by relatively few patrons donating money in a typical skewed whale donation distribution rather than by advertising or subscriptions, and is ‘recognized but illegible’ in the sense that it is well-known in some circles but utterly obscure everywhere else. We also put effort into documenting the cool technical tricks we do, even the stuff we wind up discarding, and the website itself is as discussed as much as anything I actually do with it.

So it seems to me like website luxury is a meaningful concept parallel to luxury-everything-else, and there are a few examples, but that the striking thing about website luxury is its absence.

Why isn’t there website luxury in general? What makes watches, cars, scarves, bulky handbags, cars, water etc different from websites?

I would say that part of the issue might be the lack of a peer community to show off to. What turns a good from simply something you like or are a connoisseur of, is a community of peers who you can impress wit your consumption. A woman’s Birkin bag exists to be seen by other fashionistas who know what a Birkin bag is; an ultra-thin mechanical watch exists to impress other horologists who know how hard it is to make a thin mechanical watch and how expensive and unique that must be; a hype-hound frames a pair of off-white sneakers for the other hypebeasts who know that the laces mean that was a rare drop that happened only once in a popup shop in Brooklyn in 1993_32ya, etc.

Websites, however, lack any such consumption community. Website quality and skill are hard to see, while a supercar speaks for itself. Sometimes it’s impossible to miss that a website is unusual: the Nintendo website. But a good website is just… good. It is taken for granted by a user. They can’t see or appreciate the wacky pageload trick you did to make it load in under half a second; it just seems like a snappy website. How are they supposed to know you just did the impossible?

Consider Bill Gates’s website. Gates writes the book reviews to communicate with the general public and as public advocacy, and not as any kind of luxury consumption. After all, which of his peers like Larry Ellison or Sam Altman would be impressed by some fancy website luxury, like it being absurdly high-performance? Probably none—they don’t even read the book reviews to begin with! They are important people who have people for doing things like ‘reading’. Maybe the social media addicts will see a few screenshots of key paragraphs, but they’d rather hear about it over dinner with Bill himself. They would be impressed by the announcement of some major philanthropic program, or commissiong a superyacht even bigger than Larry’s newest one—but not by any website, no matter how tastefully done. It’s just not a thing that they recognize as something to show off with. Websites are just an ordinary utilitarian sort of thing, like a press release. And so his website winds up being an ordinary quasi-corporate/nonprofit sort of website, with all the ordinary design approaches and flaws that you get from ordinary web developers & designers, who follow design fads and don’t care if a page pegs a few CPU-cores.

One would have to be a web developer oneself, and that is not that big a demographic, and poses a certain problem in trying to sell websites to them. This is mitigated to some extent in areas like mechanical watches by the historical obsolescence: you don’t have to know much about mechanical watches to be impressed by a watch with a perpetual calendar built in or which can tell you the tides, because obviously those are difficult if you can’t just use software. (Note that there is far less of a ‘luxury digital watch’ market, and to the extent that there is, the Apple Watch is doing things far beyond a ‘watch’ and are better thought of as weird iPhones.) Websites usually have no such arbitrary restrictions which could foster consumption and competition.

We could certainly imagine some! Like a Demoscene size limited website, which is no more than 1MB total—that could lead to a luxury arms race of people tuning their JS ZPAQ compressor to squeeze in a few more bytes of text or hacking HTTP response fields to save a few bytes. A viewer can, with this fixed limit, appreciate how much different websites achieve within 1MB, and can see a new trick and immediately know that it is new, and be wowed, and share that with fellow viewers to elevate the status of whoever commissioned the new website.

https://verygoods.co/smallness

What I’ve noticed since leaving Svpply is that other industries treat their 1% differently. In the fashion or magazine industries for instance, they give the creative output of their star members their full attention. The 1% is the whole thing. There’s no open invitation to contribute content to Monocle. With hindsight, I feel the reason Svpply never grew into anything substantial is because we misread the opportunity. Smallness was the steam that drove our engine and we opened the gasket.

https://news.ycombinator.com/item?id=43925732

https://commoncog.com/c/cases/y-combinator-power/

https://www.henrikkarlsson.xyz/p/art-gallery#%C2%A7the-worst-exhibitions-take-the-most-work