The beginning of the end at my last permanent full-time job came about six weeks into my time at the job, when I received an email from my CDO telling me to stop reading books at my desk. The line from the email that sticks out even now is that "we not only have to work hard, but we have to be seen as working hard": a truly fucked up line of thinking that prioritises being seen as hard-working technical types over actually achieving anything (which for unrelated reasons was basically impossible at that job). Given that I was reading books on statistics that had an immediate and important bearing on my work, this was particularly galling. And of course, this was all the way back in 2019, back when GPT was mostly used for simulating subreddits on the internet.
Since GPT became genuinely capable (of bullshitting fluently, at least), this kind of cowardly, sordid keeping-up-appearances type behaviour has only gotten worse. Whole swathes of the corporate world have become reduced to people sending ChatGPT-generated emails to each other while pretending to be performatively busy. Job applications have been snowed under by LLM-generated CVs, and companies are increasingly taking to running LLM-based interview processes where the candidate doesn't even get to speak to an actual human. And while the results of this are all obviously shit and have serious and material negative consequences, it takes a ludicrous amount of effort to get people to actually stop doing this.
The core fact that gets lost in the discourse is that the technical capabilities of LLMs are not the main reason why people choose to use or not use the technology. While it does have some impact (as mentioned in my last article, if the tool is clearly unfit for purpose people just won't use it), the primary drivers of LLM use are cultural and ideological. This is unsurprising, as it's a relatively commonplace statement these days that the development and adoption of technology is culturally embedded: culture conditions technology as much as the other way around. Looking at LLM usage and promotion as a cultural phenomenon, it has all of the markings of a status game. The material gains from the LLM (which are usually quite marginal) really aren't why people are doing it: they're doing it because in many spaces, using ChatGPT and being very optimistic about AI being the "future" raises their social status. It's important not only to be using it, but to be seen using it and be seen supporting it and telling people who don't use it that they're stupid luddites who'll inevitably be left behind by technology.
This, in turn, raises the question of what kind of culture and society would create LLMs in the first instance. After all, it's unlikely that engineers from Hamilton would create a tool like that left to their own devices: we have an economy largely built around supporting agriculture and a culture that reflects that: in fact, Fieldays is coming up soon, and you can just imagine how things would go down if you tried to show your average New Zealand dairy farmer ChatGPT when it first became publicly available (matters have changed somewhat, mostly because Altman et al. have run a very effective PR and media campaign). Even somewhere like Poland is unlikely to generate something like this: we still have an economy with a significant heavy industry presence, and even our software engineers are largely focused in on fairly prosaic applications: there's not so much of a culture of building weird shit that nobody wants in Eastern Europe, and in general we won't do something unless we know that there's real demand for it.
I can however think of one place that would absolutely have generated ChatGPT a priori: LinkedIn. The place was already full of largely meaningless dross, posts that were flat out wrong and overt posturing for status at the expense of everything else. Developing and deploying a technology that lets everyone do that much faster is exactly the kind of thing that your average LinkedIn user would really like. Given that LinkedIn is largely a reflection of corporate, entrepreneurial and Venture Capitalist spaces and the lies they tell about themselves, this is unsurprising. And this raises the question: what's so broken about our society that anyone thought any of this was a good ide?
LLMs as a cultural mirror
The rot, unfortunately, goes much deeper than the immediate applications of LLMs. Rather, the only reason that LLMs took root in the first place was because our societies in the anglosphere have already developed cultures solely devoted to gaining status and keeping up the appearance of doing things rather than actually doing them. All other values, increasingly including even the accumulation of wealth (while this is still very much a thing that people pursue, wealth is increasingly becoming a proxy for status more than something desired in itself) are becoming subordinated to symbolic status games completely detached from anything real.
Take Ludicity's article about software engineering jobs as an example. While this is intended as a mostly unironic guide to finding work in the current market, it inadvertently says an awful lot about the attitudes of your average hiring manager. In short, the vast bulk of the software engineering industry doesn't care what you're actually capable of: a software engineering candidate that gets hired is primarily one who looks the part, seems appropriately technical and can raise the hiring manager's prestige. While they do require a certain minimum level of capability, the general perception is that they're far away enough from the (mucky and altogether beneath them) job of actually making things, maintaining things or generally making the world work that they just forget that any of that has to happen. They simply drift along in a symbolic world where nothing matters but relative status, and until something comes crashing down hard enough to make it impossible, they will go to almost any lengths to hold onto that world. This is, more than anything, a GPT-supercharged version of the attitude behind that stupid interaction at my last workplace, and it's similarly ludicrous.
A similar attitude can be seen in our political leaders. While we write a lot about fascism and the degradation of democracy (and for good reason), the fact is that democracy being under direct attack or deliberately subverted is still relatively rare (though I will grant that when it does happen it's really, really bad these days). What's rather more common is the institutions that support a healthy, democratic state being left to gradually lot due to lack of funding or maintenance over years or decades. In New Zealand, for example, our roads are wearing out, our public transport is hopeless, our educational system is creating multiple generations that may as well be functionally illiterate and our healthcare system is slowly but inexorably falling apart. This has been going on for a solid forty or so years, with our politicians talking at each other consistently about the importance of "living within our means" as if that makes sense for a government's finances, constantly jockeying for status among the engaged voter base at the expense of actually making a country work. And while we're unfortunately starting to see more explicit attacks on democracy as our current government tries to dig itself out of its popularity hole, the vast bulk of the damage is still this kind of slow death of neglect, which the opposition hasn't done anything about because, again, they value their status as "serious political people" over the actual function of the state.
In short, this kind of rot is everywhere, and it's deeply ingrained in all of our institutions. From our politicians, to our executives, to middle managers and stupid people online, many, many people believe that status in our society is the only thing that matters, no matter how bad everything else might get. They care about keeping up the appearance of things working much more than they do about actual function. They will run scams, lie, grift, do anything, no matter how morally odious and dishonest, so long as it gains them status. Moreover, in a society that actively rewards such behaviour, this kind of shit quickly outcompetes any kind of actual, practical labour, creating the situation that we're currently in. And in such situations, all kinds of monstrous things begin to emerge.
LLMs are a technology that are fundamentally parasitic on this attitude. When people playing status games are hit with the inevitability of having to actually do a thing, they will very often do the easiest, most half-assed possible thing that might count so that they can get back to fighting for status. This has been the case for a very long time, but previously the real work was usually delegated to junior staff that cost money and were inconveniently human and potential rivals. The promise of ChatGPT is that it allows you to do this minimal amount of stuff without having to rely on any annoying humans or be judged by them. Similarly, LLMs provide a great deal of opportunities to increase your status and decrease other peoples': a core status marker in the corporate world these days is being "innovative" or "tech forward", which means that loudly and ostentatiously using LLM technology can be an excellent way of building your status in corporate spaces. The converse of this is that you can label LLM sceptics as being "irrational" or "frothing" and tell them that "if they don't use the technology they'll be left behind", which is an excellent way of lowering their status in the eyes of the corporate world that they function in. We've seen exactly the same pattern play out with blockchain and half a hundred other overhyped technologies.
It's no surprise, therefore, that in a society where people are trying desperately to hold onto status divorced from anything material while their country and their society falls apart around them, people would latch onto a technology that promises to semi-adequately patch things up without anything having to fundamentally change. The fact that it lets people boost their status also helps significantly, and taken together, for someone who only cares about their status, means that they're very likely to get attached to the thing. This makes sense of why people are so attached to the idea of LLMs being revolutionary despite the existence of significant objective evidence that they're incapable of being good enough for many applications and strong arguments that the existing technology is unlikely to ever be good enough (the fact that LLMs really struggle to work in a constrained environment is just one of these things). While people will eventually change their behaviour, their attachment to this status-boosting technology is so strong that they will suffer considerable amounts of real, material harm before they even start to reassess.
Finally, the idea of LLMs as a technology that's largely parasitic on status games in a crumbling society does a great deal to explain the recent spate of AI use mandates in businesses. While they're obviously, transparently bad for materially getting work done, they're amazing status boosters, and if people are willing to take on material harm in order to improve their relative status, this behaviour actually makes considerable sense.
Winning the status game as an LLM dissident
So, what's one to do when you're a professional who doesn't like LLMs and thinks they're harmful? Material or technical arguments aren't going to work here, at least not in the way we'd like: as we've established, the people who hold the purse strings are willing to absorb considerable amounts of material harm in order to keep playing their status games. We're therefore going to have to win the cultural fight, and that means being able to maintain some status one way or another. But naturally, we don't want to do what everyone else is doing, and we usually have a level of professional ethics that will prevent us from using LLMs, either in their entirety or at the least in the kind of indiscriminate way that your average LinkedIn glazer does. To state things more rigorously, I believe the four points below apply:
- We need to be able to win some kinds of status game in order to survive in our society and earn money;
- Winning the kinds of status games that everyone else is playing is morally destructive, will make us hate ourselves and probably won't work in any instance;
- We cannot by ourselves rewrite an entire country's culture, therefore;
- We need to find status games that exist in our current society that we can compete at effectively.
Fortunately, I think a form of status game that works for LLM dissidents does exist: we need to compete on prestige.
Prestige as a concept is subtly different from status. Where status is something akin to "being the most famous or the most powerful", prestige works more along the lines of being the best at a given thing: being so good that you cannot be ignored. Prestige buys you a certain amount of status, but it's quite possible to have relatively low status but very high prestige in a given space. Importantly, prestige is still something that people want, regardless of what the most popular status games in a given space are: there's a halo effect that comes with owning prestigious goods or knowing prestigious things, and when Andrew Tate, for example, buys a Bugatti, he's implicitly trying to communicate his good taste and his ability to have the best product in a given class, not just whatever's accessible that looks good (you can judge for yourself how well this loathsome excuse for a human being succeeded).
Prestige is also to a large extent about detachment: about not having to compete or not having to participate in status games because you have the taste, refinement and sheer capability to be able to avoid them. This means that we can, to an extent, disengage from the bullshit and have it actively be a productive thing. We can be our best selves and have it actively be a positive things.
Prestige is completely antithetical to the LLM ethos. Where an LLM presents something that's good enough for some purpose or other, prestige emphasises the exactly right tool for the right job. Where an LLM produces masses of largely meaningless text, prestige in writing means using an economy of expression to dig down to the exact point you want to make. Where an LLM proposes the most well-known thing or the mass equivalent of a prestigious object (Gucci or Louis Vuitton as high-status brands), prestige means having your own tailor. In a world where people are almost illiterate and certainly can't write, being able to consistently produce a 3,000 word essay almost every week and being able to demonstrate that you're extremely well-read is a highly prestigious thing to be able to do.
This, in short, is something that a genuinely strong professional (which LLM critics generally are) can effectively compete on. We might not be able to bloviate like the LinkedIn people or talk out our asses by the executives, and we might often do things the slow, old-fashioned way, but we can certainly be the best people in the field. We can be the people that those in the know work with: the people whom you go to when you want only the best and won't compromise on anything. The only question this really leaves, therefore, is whether this angle of attack has an available target audience, and I think it probably does.
While this probably doesn't apply to the likes of Musk or Zuckerberg (I have no fucking clue what's going on with them), the average executive where I live suffers from a crippling level of class insecurity. They register, on some level, that their status is ill-gotten and that they're nowhere as intelligent or capable as people claim. Many of them, therefore, do things like going to the opera and art galleries in order to appear cultured, or reading the Atlantic in order to seem like part of the cultural elite. An entire ecosystem of thinkers like Steven Pinker and Matt Yglesias has sprung up to pander to these people, playing on their wish to seem simultaneously as though they're a part of the cultured elite and as edgy, countercultural people. And these institutions, post Trump 2.0, are going through a massive crisis. This is something that we can absolutely exploit.
The approach we can take, therefore, is to make our criticism of LLM use a prestigious counterculture. We can make not using LLMs or using them strangely an interesting, exciting, almost seductive thing that the best and brightest among us do, and we can create a cachet about the software a company uses being untouched by LLMs. The process, I believe, has already started, and we can do much to accelerate it.
The first thing we can do is write. Being properly literate is already a prestige marker, and as LLM use continues to gut literacy in the general population even more than it already has, being able to write well and being able to hire people who write well is only going to get more and more impressive. People who are well-read, who can write well and who can demonstrate it are people whom we can create great demand for, and I'm attempting to begin that here already.
In our work practices, consistency is also going to be key. LLMs are notorious for their inconsistency, so being able to demonstrate that we can do good work every time is key to our success. If we write, we need to demonstrate that we can output good writing on a consistent basis, and while we obviously can't compete with LLMs on sheer volume, we need to demonstrate that we can compete with them on volume of actual quality work. If you're in a workplace where quality is impossible (and unfortunately I've been in quite a few of them), you'll need to do work and demonstrate it somewhere else.
Finally, while we don't need to work at that level all of the time, we need to produce the occasional thing that really sparkles: work where you almost can't believe that it's that good and where it's clear that an LLM could never have done it. This usually means work that means something important to people. An article that reaches many people and touches many minds, a truly elegant website or a software tool, art that expresses a truth in a way that AI-generated art just can't: we need to be able to touch people in a qualitatively different way.
When enough of us do this consistently enough, capitalism will catch on, and as we know, capitalism has a wonderful way of co-opting counterculture and absorbing it into itself. This is a massive political problem for me, but here, fortunately, it's actually a good thing. People will see this, they'll absorb it into their self-image, and before too long we'll be seeing people post about working without AI being the trendy, fashionable new thing. People will see our work as novel, desirable and the thing to copy, and a whole lot of people will doubtless try and copy it badly, bringing our skills into even sharper relief.
And at that point, we'll have won our struggle.