Photo-Illustration: Intelligencer; Photo: Getty Images

A few weeks ago, OpenAI’s Sam Altman circulated a memo assuring his employees that, despite recent advances by competitors like Google, which could create some “temporary economic headwinds,” the company was “catching up fast.” The firm was “doing remarkably well,” he wrote, which he expected to continue. In the meantime, however, he had a warning: “I expect the vibes out there to be rough for a bit.”

AI guys love talking about “vibes.” There’s “vibe coding,” a term coined by OpenAI co-founder Andrej Karpathy to describe writing software by prompting AI without touching any code. In Microsoft’s deranged marketing imagination, this grew into “vibe working.” Meta’s AI-generated social app is called Vibes. Altman has long been fond of the term, fielding recent questions about the company’s planned devices with an aspirational vibe — “sitting in the most beautiful cabin by a lake in the mountains” — rather than specifics about, for example, what it’s shaped like.

It’s an apt word for describing general-purpose technology that produces unusually subjective and personal user experiences. It’s also useful for describing the AI discourse, which is notoriously intense, disputative, polarized, and prone to sudden and major swings in sentiment. This is partly intrinsic to the subject, a story that’s moving fast, and fundamental assumptions that really do keep changing. But Altman’s warning was a nod to a strange fact of the AI boom that’s becoming, even for him, impossible to ignore: It’s unfolding, to an extraordinary degree, on a single social platform owned by Altman’s collaborator turned nemesis, an AI heavyweight in his own right. These “rough vibes” — and the AI industry’s vibes in general — are made up of a bunch of tweets.

Altman is clearly devoted to posting on X, where he teases product announcements, trolls competitors, shares cryptic proclamations, and tries to be funny. This might seem unusual or risky for the CEO of such an influential start-up, but in AI, it’s closer to typical: The entire industry is addicted to X, from its most prominent leaders down to the enthusiasts who want to keep up with what’s going on. It’s where CEOs become internet celebrities and also feud; it’s where scientists and critics debate their theories of scaling and accumulate fandoms (Yann is subtweeting Sam again! Demis is clapping back!); it’s where policy discussions unfold and where start-up deals, in the DMs, get made; it’s where small but influential clusters of AI ideologues build cultish followings and occasionally drive themselves insane; it’s where the latest Arxiv research papers are surfaced, sorted, picked apart, and decontextualized by engagement-baiting influencers; it’s where anonymous rumors and leaks surface and explode and where fast-shifting narratives, and vibes, manifest before trickling out into the wider world, where they are, for now, propping up the American economy; it’s where the vibrant but nascent AI media talks to itself. If you listen to enough of the Dwarkesh podcast, for example, you’ll notice that, whether he’s talking to an AI researcher or a CEO, and whether the subject is reinforcement learning or the end of the world, few guests can avoid bringing up a particularly memorable recent tweet they saw.

The reality of the AI industry is far larger than a subculture on X with hundreds of billions of dollars flowing into physical infrastructure, reshaping jobs, and deploying into products used by billions of people. But that makes X’s capture of the AI conversation all the more notable. Much as Twitter long ago captured, amplified, and distorted elite conversation in media, in sports, in parts of finance, in left politics, and more recently in right politics, the world’s understanding of what’s going on in Silicon Valley right now — the hype and the doom and the bubble and the progress — is first processed through the strange culture and incentives of X, which is now owned by estranged OpenAI co-founder, Google antagonist, and xAI founder Elon Musk. It has been merged into xAI, where the in-house chatbot occasionally confuses itself with Hitler. When the history of this period in tech is eventually written, it will be composed of, among other things, a whole lot of threads and replies from blue checks on X.

X’s AI takeover was slow, voluntary, and perhaps inevitable. The platform was, before Musk’s takeover, the default social platform for tech workers and academics, making it a good fit for a research-centric subject like generative AI. After the release of ChatGPT, much of the resulting explosion of attention around the technology, which users, members of the media, researchers, and industry figures were exploring and processing together in tweets, accrued to what was then still called Twitter. The entire industry’s reorientation around LLMs played out on the platform, which was a natural fit for an exciting story that was moving both quickly and in small increments. AI was both a massive subject and one well matched to the discursive machinery of X. As an abstract news event, it was something akin to an election, except instead of a scheduled vote, the story was to end with at least one of the above: apocalypse, fully automated luxury communism, widespread immiseration, AGI, ASI, or perhaps a bubble pop and a recession.

All the big AI firms are well represented on X, even if they’re attached to other social-media platforms: The executive and research teams at Meta, Anthropic, and Google post constantly. OpenAI, however, is the emblematic X-ified AI concern, even more than Musk’s own xAI. It’s a company run by an X addict with an army of real-name and pseudonymous online personalities spreading its message. Altman’s concern with the vibes on X is deep and in character — industry insiders joke about the X-to-OpenAI hiring pipeline, and the company’s internal hierarchy is informally influenced by employees’ status and reputation on the platform. Altman loves a good tweeter almost as much as he loves his own tweets.

For example: pic.twitter.com/qcEEjfG8g0

— Sam Altman (@sama) December 16, 2025

For the public, the upside of such an X-brained industry is real: The AI world’s insatiable urge to post means that many of its principal characters, interesting thinkers, and most compelling skeptics are engaged in constant public conversation and conflict about a consequential story. It’s chaotic but relatively transparent, too, and a vastly richer resource for keeping up with what’s going on in AI development than reading the mainstream press. It’s evidence, in an otherwise diminished or disrupted platform, that Elon Musk’s X has still got “It.”

But in 2025, we know a little bit more than we used to about what X’s “It” actually is and the perils of a particular community, or movement, or industry, becoming too entwined with the platform. For one, overexposure to X/Twitter, a platform that favors breaking news and conflict, and where novel narratives drown out the status quo, has a broadly radicalizing and polarizing tendency, particularly for people with lots of followers.

Another recurring story, as it relates to capture by X/Twitter, is of cultural elites talking each other into progressively less relatable positions and conventional wisdom while also performing their insularity in public. This is something Democrats, for example, struggled with in the 2020s (Twitter is not real life, etc.) and something that freshly X-ified Republicans seem to be grappling with now, as their movement’s most powerful leaders mingle and tweet with the MAGA fringe, directing their attention to increasingly esoteric concerns and away from regular voters’ concerns, producing strange figures such as the latest iteration of J.D. Vance, an X account given human form, the nominal sitting vice-president who no longer seems capable or interested in communicating outside the style of a withering tweet. (If indeed a small, late rift is developing between Trump and parts of his reliable base, I think the MAGA elite’s addiction to X has at least a little bit to do with it.) Also, despite pockets of vibrancy on X, the place is clearly a mess, even to some of the industry neo-celebs who continue to build identities there:

the amount of completely wrong utterly worthless falsehoods on this website is really astonishing, and it’s even more astonishing watching sitting senators and CEOs fall for it. Postman’s Information Glut every day. despite grok and community notes it’s never been worse

— roon (@tszzl) December 9, 2025

In AI, too, the allure of X seems to be getting the better of people who seem like they probably have other things they should be doing. There even seems to be some emerging concern among prominent voices in AI — including the White House’s own AI advisers — that the way they talk, largely on X, is off-putting, alienating, and perhaps even alarming to the rest of the world.

I’ve come to have mixed feelings on “AGI” and “ASI” as terms to convey where this technology is headed.

On one hand, AGI has played a key role in motivating talented people who obsessed over this problem (@demishassabis, @ShaneLegg , @ilyasut , many others) and enabled the flow…

— Sriram Krishnan (@sriramk) December 3, 2025

Then, of course, there’s the Musk factor. The fact of his ownership creates obvious risks for all of his competitors, who rely on his platform to supply the vibes they need to help keep their companies, and the AI boom in general, going. It’s worth remembering that, back before he decided to purchase it and let it drive him nuts, Twitter was, to Musk, a platform with clear functional utility, a social network that helped make him wildly famous and use that fame to convince thousands of retail traders to buy into Tesla, the company that made him the richest man in the world.

Even if we assume, against all reason, that Musk won’t leverage this unusual sort of power over all the competitors gathered on his platform, X’s influence over them and how they’re seen by the rest of the world is undeniable. Along with “AI Twitter,” some of the other communities thriving on Musk era X include MAGA guys and crypto guys, meaning that the background vibe of the primary gathering place for the industry that’s trying to remake the world is quite a bit more conservative and mercenary than it otherwise might have been, and while the primary factor in the public’s deep skepticism of where AI is going is clearly fear about labor, association with polarizing Muskian politics probably doesn’t help assuage ambient fears about job loss and surveillance. (One important consequence of having the AI elite gathered on X is that their media diets are now substantially made up of one another’s posts, effusive and dismissive responses to their own posts, and standard “For You” page Musk-era algo-slop. It’s hard to overstate how thoroughly their information diets are influenced by their timelines, and you can get a pretty good idea of what those look like by jacking into AI Twitter yourself.)

Depending on how things go for xAI, this alone might have made Musk’s Twitter purchase worth it. Perhaps he gets an edge in an AI race that’s looking increasingly close. At the very least, he gets a measure of soft power to extend his animating crusade against the “woke mind virus” to AI, too.

The AI industry’s X obsession is obvious enough for insiders to joke about, but all evidence suggests they’ll be riding with their feeds until the end. (It’s certainly worked out so far!) As analysts try to identify sources of financial risk and evidence of an AI bubble — will circular financing backfire? Is Oracle’s stock price a worrying bellwether? — the industry’s capture by X, a chaotic platform full of capricious audiences owned by an impulsive and invested tech magnate who holds a grudge, represents an underappreciated latent risk. (For all their resources, the finance world doesn’t have a better way to keep up with the AI story either, and you’d better believe that massive trades are taking place based on tweets from AI Twitter influencers.) It’s a place regular people can get an unusually clear glimpse of a fascinating and consequential new technology and witness real candor from some of the people involved in building it. At the same time, it’s a platform that may have helped the lucrative, contested, and now load-bearing story of AI spiral out of control and where any shift away from that story — be that a subtle shift in the vibes about AI development or a more disruptive narrative contagion — will take hold first as a meme.

Sign Up for John Herrman column alerts

Get an email alert as soon as a new article publishes.

Vox Media, LLC Terms and Privacy Notice