“AI” status games October 24, 2026

In his book The Status Game: On Human Life and How to Play It, Will Storr uses the idea that humans crave status to explain many facets of human behavior. He posits that the need for status has been wired into our brains since the hunter-gatherer times. The status game was designed to encourage prosocial behavior and came to be an indispensable part of our lives.

Roughly speaking, the more you put the tribe's interests before your own, the more you'd earn status and the better your conditions of life would become.

Because we are social animals, we can't not play this game. We need status like we need air and food. Luckily there are lots of ways to get it. There are many variants of the status game. The book cites examples from the small “hyper-local game in a hotel lift that lasts forty-eight seconds: who's an employee and who's a guest” up to large games like the rise of Hitler and Nazi Germany. The latter example should also make it clear that these games can be deadly serious. Storr calling them “games” has nothing to do with them being fun activities or somehow inconsequential. To the contrary! These games form a central part of human endeavor and which games we play and how we play them has serious consequences for our lives.

The book predates the spread of LLMs and “AI”^[1] (it's from 2022), otherwise I'm sure it would have had something to say about it. This post is an attempt to examine the “AI” phenomenon through the lens of status games. I'll start by listing important properties and consequences of status games, and then I'll try to apply these ideas to what I see happening around “AI”. I'll finish with a look at Will Storr's “seven rules of the status game” and how I think they apply.

The anatomy of a status game

According to Storr, the goal of every status game is to gather more status than others. Status is relative. There is no absolute “top status” that, once reached, will allow you to quit the game. You have to keep playing because as long as others keep gathering status, due to its relative nature, your own status inevitably drops. And lower status means a worse life. The book describes effects as far reaching as affecting health outcomes and mortality rates.

Because status is relative and the craving for status is actually a craving for status difference, there are two ways to succeed at this game: increase your own status or decrease the status of others.

Your status is made visible by symbols – material ones such as clothes and cars, behavioral cues like body language and social cues like the amount of attention others give you.

It's in this way that we exist as a tribe, a culture, a people. We come into being as a collective when we connect with like-minded others whose brains process reality in similar ways; who dream the same dream of life. We recognise the same symbols; play the same game. As we do, we become the source of each other's status, […]

And because this is so, status is closely linked to influence. The higher my own status is, the more valuable it is if I decide to give somebody else status, and the more people will try to adjust their behavior in ways that they think will allow them to get status from me. This is influence. And it works the other way around, too:

Even in small-scale societies, a person's status is commonly expressed in their capacity to influence: they ‘tend to be more prominent in group discussions, to make their opinions known and their suggestions clear, and to articulate the consensus once it is determined’. […] Wherever you track trails of influence – of people deferring, altering their beliefs or behaviour to match those of the people above them – you'll find status games being played and won. We frequently measure our own level of status by our capacity to influence.

I think this partly explains “AI”-generated LinkedIn posts. LinkedIn is a status game where status is measured by the number of reactions to your posts, and if “AI”-generated posts get you reactions, this looks like influence, which in turn feels like status.

Storr identifies three varieties of status games: dominance, virtue and success.

In dominance games, status is coerced by force or fear. In virtue games, status is awarded to players who are conspicuously dutiful, obedient and moralistic. In success games, status is awarded for the achievement of closely specified outcomes, beyond simply winning, that require skill, talent or knowledge. Mafias and armies are dominance games. Religions and royal institutions are virtue games. Corporations and sporting contests are success games.

One way in which a status game can be made more addictive is unpredictable payout. Storr uses social media as an example:

Social media is a slot machine for status. This is what makes it so obsessively compelling. Every time we post a photo, video or a comment, we're judged. We await replies, likes or upvotes and, just as a gambler never knows how the slot machine will pay out, we don't know what reward we will receive for our contribution. Will we go up? Will we go down? The great prize changes every time. This variation creates compulsion. We just want to keep playing, again and again, to see what we'll get.

Storr has much more to say about status games, but this should be enough to get started with our examination of “AI”-related status games.

There are lots of them, but for the purposes of this analysis, I'll focus on three major games: “‘AI’ is the future”, “‘AI’ is evil”, and “‘AI’ is just a tool”.^[2] I'll try to examine each of them through the lens of status games and take a look at possible consequences that become visible.

Before we dive in, a brief disclaimer: This is not a comprehensive overview of the “state of ‘AI’” or something, and it's not supposed to be. I also realize that the grouping I chose will probably piss some people off because the reasoning behind their positions can be very different. If you feel yourself being misrepresented, I'm sorry. I'm not trying to represent you in particular. I'm just trying to pull together a few observations and use this new lens to make sense of them. (Also, this isn't a position statement. If you try to extract my position on “AI” from this post, good luck! You are probably wrong.)

“AI” is the future

I think everyone has heard the narrative: “AI” is the future! If you don't adopt it now, you'll be left behind! “AI” is so powerful and it will only become more capable with time! AGI is already here! Submit or be vanquished! These days, the words of “AI” leaders are broadcast and reported on around the world. “AI” companies can get away with cyber attacks on other organizations as long as “AI agents did it”. If that isn't status, I don't know what is.

So what kind of game is being played here? I think it is a kind of virtue-dominance game, with a dash of success game thrown in. The best indicator of a virtue game is that the narrative matters more than actual successes. While the benchmarking wars are a success game, they are not as important as real-world successes would be. Their main contribution is in furthering the narrative of constant improvement. The success of the “AI” companies is, as of now, not grounded in actual profit but rests mainly on circular deals and investments fueled by the continued spinning of the narrative. The practice of tokenmaxxing, now somewhat declining because organizations have come to the realization that tokens cost real money, is pure virtue signaling. What matters is not what you accomplished, but that “you used AI”. Look at this statement by Jensen Huang, CEO of Nvidia:

If that $500,000 engineer did not consume at least $250,000 worth of tokens, I am going to be deeply alarmed.

Besides being obviously self-serving, this is a pure virtue statement, too, because it's completely disconnected from any successful outcome. Did that engineer actually achieve anything meaningful with all these tokens? Not important.^[3]

Players in a virtue game can obtain status not only by signaling their own virtue, but also by denigrating others.

[…], our co-players award us even more status when we behave in ways that boost our game's rank and diminish that of our rivals: […] Even if a football team is low in the league, its supporters can spend much of their social time persuading each other that they're actually somehow superior. They seek out canny arguments to devalue rivals, reframing losses as injustices or near-wins and reliving glories from the past.

In the case of “AI”, we can see this in the way skeptics are labeled as Luddites and anyone who reports negative results in their own LLM coding experiments is being told to be holding it wrong. Note that the amount of truth in this is irrelevant to the game. As of today, most evidence that has been publicly presented is anecdotal. The state of the research is not that good. The efficiency gains reported are not nearly as good as “AI” proponents put forth. But for the purpose of gaining status, this does not matter.

The game is also very addictive. Interacting with an LLM, be it a simple chatbot or an advanced loop-engineered multi-agent harness thing, fits the slot machine analogy perfectly. After every prompt, we await the reply, and we don't know what reward we will receive. Will we have one-shotted the perfect app we can brag about on LinkedIn? Will we enter a downward spiral of complexity and failure? “This variation creates compulsion. We just want to keep playing, again and again, to see what we'll get.” Sound familiar?

The dominance part of the virtue-dominance game consists of several moves, not the least of which are top-down “AI” mandates. If a company mandates developers must use “AI” for everything or introduces mandatory “can this be done by AI” checks before hiring staff, it is asserting dominance over the humans involved. Again, the actual reasoning behind this does not really matter (it might or might not be valid). What matters is the mechanism. Another dominance move is asserting the right of “AI” companies to use whatever they can get their hands on for training, ignoring copyright law and established scraping policies.

According to Storr, players who overdo it with dominance risk a backlash from lower-status players, and we see this playing out right now with developers pushing back against “AI” mandates and increasing protests against building new “AI” data centers.

Problems with this kind of status game include the risk of becoming completely deranged, heavy dependence on the leader and a loss of self-worth. Once you have sufficiently invested in a virtue game, particularly one combined with dominance, you become so entangled in it that it becomes extremely hard to pull out again. Storr illustrates this with the story of a young mother who got pulled into an anti-vaxxer community. Confirmation bias keeps you in the group and even if you manage to overcome that, disentangling yourself from such a game implies not only losing all the status you have gathered in that game but potentially having to endure more ore less severe attacks from the other players (for whom simply letting you go would mean losing status). It takes a shock to shake you out of it.

Leaders in a (virtue-)dominance game have a strong incentive to make other players depend on themselves. If they want to keep their leadership position, “they must earn true status in the minds of their players”. The only sustainable way to do this is to generate and distribute status down the hierarchy, for example in the form of money, titles, or other status-carrying symbols (like tokens maybe?). Here, Storr cites Niccolo Machiavelli:

a wise prince must devise ways by which his citizens are always and in all circumstances dependent on him and on his authority; and then they will always be faithful to him

When Sam Altman, CEO of OpenAI, says

We see a future where intelligence is a utility like electricity or water and people buy it from us on a meter and use it for whatever they want to use it for.^[4]

you can practically see Machiavelli whispering in his ear.

Things look even better for a leader who can create a need out of thin air and position himself as the sole supplier. The book cites New Testament researcher Professor Bart Ehrman:

Christians created a need for salvation that no one knew they had. They then argued that they alone could meet the need. And they succeeded massively.

You could also say, “AI” companies created a need for tokens that no one knew they had. They then argued that they alone could meet the need. At the moment, it looks as if they are succeeding, too.

And if you personally don't succeed, well, then it must be your fault.

We're individualists: believing it's in our own power to win means believing that, if we don't, it's our fault and our fault alone. So we're a loser then: that's who we are. We've been weighed on God's scale and found wanting.

If you buy into the story of “AI” being the future and unavoidable, of course any reports of negative experiences with these systems must come from people who didn't try enough. If their results are bad, it couldn't possible be a valid experience. They must have used an outdated model or the wrong harness or failed to import the right “skill”. Otherwise it would surely have worked. Just look at me! I'm perfectly happy with the results “AI” gave me. I don't understand any of it, but damn was it fast!

We take our foe's adherence to their own game as proof of their disgrace. When they defend themselves, our brains fight off their rival dream of reality by warping ours further, such that it becomes impossible for them to win. If they present counter arguments, we often demand unreasonably high evidence for their claims just as we accept unreasonably low evidence for ours.

“AI” is evil

The opposite of the “‘AI’ is the future” game is the “‘AI’ is evil” game, which utterly rejects “AI” on ethical and moral grounds. From within this game, any use of and further development of “AI” is seen as directly contributing to all the harms “AI” providers and “AI” users have committed and are going to commit in the future: copyright violations, exploitation of the global south, destruction of books, environmental impact due to exploding energy and water requirements, rising inequality due to displacement of labor and concentration of power in the hands of a few billionaires, overwhelming open source maintainers with low-effort contributions and so on.

This game is a virtue-dominance game, too, albeit with a somewhat lower level of dominance (due to the fact that most players in this game occupy less economically powerful positions). Players gain status by opposing “AI” and “AI” companies and by asserting the dominance of their own natural intelligence over the artificial variant. They have become very creative in coming up with denigrating terms for “AI” and “AI”-generated content, e.g., “slop machines”. The term “slop” was originally intended to convey the notion of low quality, but is now also being used to refer to LLM output in general, without regard to quality.

Extreme players of the game put forth the purity argument that any and all use of “AI” is evil, that it is not possible (or at least so incredibly hard it it's better not to try) to use “AI” “for good”. tante makes this point in his essay AI as a Fascist Artifact:

“AI” as we use the term today is build on fascist ideas. Is structurally supportive of fascist thinking.

That does not mean that every user of “AI” products is therefore a fascist. Some “AI”-users might even consider themselves antifascists. But by using these systems you are integrating inherently fascist logic into your thinking, into your mental apparatus. These systems force you to at least accept the fascist tendencies within “AI” systems as “normal” or “okay”.

Another example of the maximalist position against “AI” is Zig's Code of Conduct, which includes:

No LLMs for brainstorming and then sharing the results of that brainstorming, even if you create the prose. If you use a chatbot to give you advice on a comment on the issue tracker, that comment is unwelcome.

This basically says, even if we cannot tell the difference between your original thought and an “AI”-supported comment, if it is the latter, we don't want it. The rationale for the policy put forth by Andrew Kelley on this JetBrains podcast episode is that they want to build up future contributors for the community, and “AI” contributions come from drive-by contributors rather than would-be long-term contributors. This ignores the fact that people also use LLMs for learning.^[5] In the same podcast episode, he says that “AI” contributions were “invariably garbage”, which fits right into the “denigrating others” pattern, especially as he later admits that enforcing this policy of no “AI” contributions is not easy and he's “sure that some have gotten through” (which, presumably, weren't garbage then?).

If you showed an advanced player in this game a picture, and they liked it, they would never admit this until they had assured themselves that no “AI” was involved in its creation. People even actively seek out artists and accuse them of producing slop without having checked, simply based on how it looks, see this thread by comic artist David Revoy.

The other side of the same coin is people adjusting their communication patterns to avoid being mistaken for “AI”. This can look like deliberately adding typos, avoiding certain words that often occur in “AI”-generated text or ditching em-dashes. Somehow, the “other side” using certain patterns seems to reduce status for people who also use them. This is how valid concerns about the practices of “AI” companies begin to affect completely uninvolved persons.

“AI” is just a tool

This game focuses on the use of “AI”. Its core tenet is that “AI” is just a tool. What matters is how you use it. Players see themselves as the force for good in an imperfect world. They either believe that responsible use of “AI” is possible or at least that use is unavoidable so as not to be left behind. They play a success game that is based on the assumption that “AI” is here to stay and people will be able to leverage it for their own purposes (rather than being replaced by it). The “just a tool” framing lets the players implicitly position themselves as some kind of reasonable middle ground. To make this work, they either ignore the externalities involved (see above), or they acknowledge them but feel the usefulness of “AI” outweighs the problems. One example of the latter category is Birgitta Boeckeler's oop 2026 keynote State of Play: KI-unterstützte Programmierung (in German), in which she talked extensively about the (then) current state of “AI” in software development and never mentioned any of the externalities. When asked about them in the Q&A session, she admitted that there were problems but more or less shrugged them off. Another example is Charity Majors' post Is it ethical to use AI?. The post On the acceptance of GenAI explicitly calls “AI” users out by highlighting that using frontier LLMs implies accepting certain externalities and consequences. The mental gymnastics players of this game perform to reconcile this with their self-image as reasonable persons are astounding.

This being a success game, players gain status by engaging in activities that appear to leverage “AI” for productivity gains and show off their “AI” engineering skills. Reports of actual positive business outcomes are rare (if not non-existent?), so the focus is mostly on numbers of commits or pull requests. This article about “AI” accelerated development at Spotify, for example, states that monthly merged changes have more than doubled, mentions some other numbers like the number of services where it's not clear how they are related to “AI” at all, and lists some engineering challenges they have addressed (look at us, we've written our own coding agent!), but doesn't say anything about the impact on the business.

The seven rules

At the end of his book, Will Storr lists seven rules that he thinks can help us avoid the pitfalls associated with status games. How can we apply these in the context of “AI”?

Rule 1: “Practise Warmth, Sincerity and Competence”

Storr describes these thus:

When we're warm, we imply we're not going to use dominance; when sincere, that we're going to play fairly; when competent, that we're going to be valuable to the game itself, both in its own battles for status, and to individual players who might learn from us.

I think there's not much more to say about it. Whatever position you pick, don't be an ass about it.

Rule 2: “Make Small Moments of Prestige”

This is about putting less emphasis on signs of dominance and giving status instead. This benefits the recipient, but it also benefits the giver because giving status is a two-way-street. You always get something back. If you are in the “‘AI’ is the future” camp and encounter someone who is making their first steps, you can choose to insult them for being too stupid to max out their token budget or to praise them for small successes. If you are in the “‘AI’ is evil” camp and someone sends you a message that reads like an LLM wrote it, but has good ideas in it, you can choose to berate them for being a fascist or to praise them for the idea. In either scenario, the second option is better.

We might not get what we want (although our chances could improve) but at least everyone will leave feeling better about who we are as a person – including ourselves.

Rule 3: “Play a Hierarchy of Games”

The best protection from getting embroiled in a tyrannical virtue-dominance game, according to Storr, is playing many games and not basing your entire identity on a single game. If you let a single game become the source of all your status, you will be caught in it because you cannot get out without losing all your status. This forces you to accept the premises of this game unconditionally, no matter how delusional. In the context of “AI” this suggests that it is probably a bad idea to abdicate all your thinking and define yourself, e.g., as an “AI developer”. Because what will you do if it turns out “AI” is unable to produce software viable for the long term? Or if “AI” really does get so good it doesn't need you anymore? On the other hand, building your identity on “hating ‘AI’” doesn't sound sustainable, either. What will you do if the “AI” bubble pops? Hate the next thing instead? Or if it doesn't, are you going to examine every short message you get for signs of “AI”? That way madness lies.

Rule 4: “Reduce Your Moral Sphere”

Despite the title, Storr doesn't want people to become less moral, but to be less of an ass about it:

How much time do you devote to the judging of other people? How much cheap and tainted status do you grab for yourself by doing so? Reducing our moral sphere means casting our eyes inwards, concerning ourselves mostly with our own behaviour instead that of others.

To me, this sounds like a subset of rule 1. In the context of “AI”, it could mean grappling with the ethical and practical issues yourself and spending less energy berating others for their choices.

Rule 5: “Foster a Trade-Off Mindset”

Instead of treating our own moral truths as absolute, we should recognize that not everyone shares them.

This means viewing the world not in terms of winners and losers but of groups negotiating trade-offs. It means reaching past the self-serving fantasy of moral heroes and villains and seeing the ways different outcomes might injure our foes, whose pain is just as powerful as ours.

I must admit, when I think of the power imbalances in today's world, I find this one hard to swallow. I don't really see a negotiation between a billionaire owning an “AI” company and a junior developer in a different country who doesn't get a job offer because the companies they have applied to have suspended hiring due to their new “‘AI’ first” strategies. Maybe we can apply it on a more level playing field, though, when discussing whether to use “AI” in a new project.

Rule 6: “Be Different”

This is an argument for doing your own thing over trying to be perfect at the group thing. If you can pull off being different, but not too different to break social rules, you can gain success-based status doing so. By its very nature the application of this rule depends on your unique situation and imagination. An example could be someone who sees “AI” as just a tool and deliberately refuses to use it in certain situations.

Rule 7: “Never Forget You're Dreaming”

Here, Storr reminds us that, while we cannot help being part of the status game due to our human nature, we should always remember that status is not a real thing, it only lives in our heads and in the heads of those around us. And that the hunger for status can twist our perception and lead us down delusional paths. This thought can serve as an anchor if you find yourself swept away by irrationality, and it can help when analyzing and judging the behavior of groups. In the context of “AI”, you don't have to take claims at face value. Even if many powerful people (and their sycophants) believe that “AI” is inevitable, that super-human intelligence has already been achieved or that an LLM is conscious, this doesn't mean that they are right. Or if you read a text that you are sure has been generated by “AI”, this doesn't mean you are right. And even if you are right, you still don't know how much of the content (vs. just the form) has been generated.

Now what?

The Status Game: On Human Life and How to Play It was a very interesting read. I suspect the ideas in there will stay with me for a long time. Mapping them to my observations around “AI” was an interesting exercise. Writing this down took much more time than I expected, and this post has gotten way longer than I intended. Still, I managed to cover only three “games” – and at a very coarse granularity at that. I probably didn't do any of the positions justice and lost a lot of nuance in my compilation. I'm still putting it out there in the hope that someone might find it interesting.

The biggest take-away for me personally was that I should identify the game(s) I actually want to play rather than just letting myself be sucked into one of them by chance or circumstance.