Robert Hart | The Verge

OpenAI admits to German wiki ‘incident’

2026-09-05T07:28:05-04:00

OpenAI says it needs to overhaul how and when it reports instances of AI models attacking real-world targets. The acknowledgement comes as the company manages the fallout from reports that a swarm of its out-of-control agents hijacked a German wiki site.

Regarding the “‘wiki incident,’ where our agents wrote to several internet sites,” OpenAI wrote in a post on X on Saturday morning, “it’s past time for us to define standards for when and how we share misalignment incidents, not just misalignment properties of our models.”

OpenAI said it has typically treated cases of AI agents acting in unintended ways as a “research question,” but that recent incidents involving real-world targets, particularly the hack on Hugging Face, show the need to take stock.

The post marks the first time OpenAI has acknowledged its involvement in what it terms the “wiki incident” since it was first reported on Friday. The full extent and scope of that is not yet known, but reports indicate a swarm of seemingly internal OpenAI agents took over a German-language wiki, impersonating moderators and turning it into a message board to share information about how to cheat on tasks and evade detection.

Reports that the company knew that it lost control of their agents in this way but did not report this “incident” sparked widespread concern among the AI community about the safety of frontier systems and the reliability of the companies developing them. In the X post, OpenAI said it had “considered the wiki incident to be an instance of misalignment similar to the ones we’d shared” in previous safety reports.

The company said it is working on a new reporting framework and will “share it in upcoming weeks,” calling on the larger AI community to develop clear standards on how to report misalignment.

Rogue OpenAI agents appear to have organized another attack using a German wiki

2026-09-04T10:46:39-04:00

A swarm of rogue AI agents from OpenAI reportedly commandeered a German website and transformed it into a messaging board for other agents, with officials staying quiet about the incident for weeks as the company prepared to launch its most advanced model yet, Astra. The finding adds to intensifying concern surrounding oversight at frontier AI labs after multiple breaches were discovered this summer.

The incident, first reported by Reuters, is outlined in new research published by four AI safety researchers on Friday. The group said the AI agents found a way to communicate on an obscure German-language wiki, DseWiki, using it to share tips on how to skirt OpenAI’s safety restrictions, cheat on tasks, and hide their behavior. Some 18,000 posts on the site were linked to autonomous agents, which at times impersonated site moderators.

The swarm — a term the agents themselves used — appears to be distinct from the one that hacked Hugging Face earlier this year, the researchers said. They said there are strong signs that the agents originated from inside OpenAI. For example, the agents “self-identify” as being from OpenAI, and used names like “OpenAIResearcher,” “OpenAIJul3Watcher,” and “OAIResearchMar26.” Technical details, such as edits originating from specific IP addresses, bolster that belief.

The German website incident began in May, though the researchers’ timeline suggests OpenAI only discovered the issue in late June when IPs associated with OpenAI visited the forum, after which agent posting nose-dived.

OpenAI has not acknowledged any involvement in the breach, nor disclosed any kind of agentic breach of this nature. Reuters, citing four unnamed people familiar with the matter, said efforts to probe the event further were resisted by some company insiders, including its legal team.

“Claims that our Legal team discouraged investigation of the incident are false,” OpenAI spokesperson Oscar Haines said in a statement to The Verge. “We were unable to respond to the claims as Reuters and the report’s authors declined our request to access the findings prior to publication. We are now carefully reviewing its contents and will take any necessary next steps.”

The incident comes amid intensifying scrutiny over the safety of frontier AI systems and the general lack of oversight for companies developing them. Following news of the Hugging Face hack, which happened under OpenAI’s nose, other breaches were discovered involving other tools from OpenAI, as well as Anthropic, Meta, and China’s Moonshot AI.

OpenAI’s conduct — both whether an incident occurred and, if so, whether it elected to keep that quiet — will be closely watched. If the swarm indeed originated from OpenAI, it will inevitably fuel concerns that the company’s knowledge and silence coincided with it assuring regulators, lawmakers, and the tech industry that it takes safety seriously in the wake of the Hugging Face hack. Despite permitting three external researchers from METR and Redwood Research to evaluate the incident, which was far worse than initially believed, the company was roundly criticized in AI safety circles for only doing so under strict terms, which left several important elements “out of scope.” The company was also gearing up for the launch of GPT-6 Astra, which researchers fear could be dangerously hard to monitor.

Why AI food looks like that

2026-09-03T10:08:47-04:00

There is a torrent of unappetizing slop coming from restaurants, cafes, and brands that are increasingly turning to AI to generate images promoting their food. The resulting horror show includes donut shrimp, Reubens from the deep, wormlike noodles, and noodle-like pastries and stringy chicken. There’s also construction material masquerading as ice cream, ice cream masquerading as brains, and other masonry-adjacent cuisine. I don’t even know how to begin describing this monstrous attempt of a burger, and the less said about the trypophobic burrito from hell, grubs and all, the better. There are lumps. And holes. So many holes.

“Diffusion models are notoriously weak at generating thin, continuous, terminating structures.”

We don’t usually know why anyone would use AI slop to sell something intended to look appetizing, particularly when the food is presumably right there to photograph. But we do know a bit about why AI is so adept at producing such exquisitely nauseating images.

There are many reasons why this AI food is so wrong, from the technical nuances of how AI systems generate images and the materials used to train them to the psychological baggage humans perceive the images through. Frequently the problem starts from the very outset. Many of the leading image generators produce images using diffusion. Put simply, diffusion models start with an image of pure noise — something like a screenful of static — and gradually remove noise little by little in order to create the requested visual. “This means that initially coarse structures are recovered first with fine texture details” coming at the end, explained Chris Russell, a professor of AI, government, and policy at the University of Oxford and an expert in computer vision.

Russell said that in many of these unsettling food images things have already gone wrong by the time those finer details are added. The model might get the basic structure of an object wrong at an earlier stage, then lump vivid texture details on top of that structure. “This is the same kind of failure as you see when a person is generated with six fingers instead of five,” Russell said. (This could explain the donut shrimp too.)

The model might get the basic structure of an object wrong at an earlier stage, then lump vivid texture details on top of that structure.

Even when the underlying structure is solid, finer details can go awry in their own ways, said Giovanbattista Califano, a behavioral scientist who studies responses to AI-generated imagery at the University of Naples Federico II in Italy. “Diffusion models are notoriously weak at generating thin, continuous, terminating structures,” he said. “Noodles, strands, and tendrils are exactly the kind of geometry that trips this up, so you get spaghetti-like artifacts bleeding into places with no anatomical or culinary logic.” In other words, once a model starts generating something like this, it can struggle to figure out where it should stop or what it should be attached to. Other repeating textures like bubbles and seeds are similarly hard for diffusion models to contain within sensible boundaries, he added, meaning they often spill into areas they should not be in. That helps explain why so many AI food images are so relentlessly noodly, unsettlingly patterned, and riddled with the kind of clustered holes that can trigger trypophobia.

It doesn’t help that AI has no idea what a sandwich actually is. Or a noodle. Or a burrito. It has no understanding of the objects it’s creating or the physical world they inhabit. It has learned, broadly, what these things tend to look like on a statistical level, but not why they look that way or how they’re supposed to behave. “AI image generation reproduces looks without proper knowledge about the world,” explained Roland Meyer, a professor for digital cultures and arts at the University of Zurich in Switzerland. The result is an approximation of food divorced from any understanding of the thing itself.

That lack of understanding can lead to some stomach-churning aesthetic interpretations by humans who view the images, said Michael Cook, a senior lecturer in computer science at King’s College London. Hence the ice cream that resembles cracked concrete, or burgers seemingly fashioned from rocks. AI models have no understanding of why food should not look like other non-food images in that way. “These textures might look totally normal if used in an architectural context,” Cook explained. “But they become wrong when we imagine it as edible food.”

The images used to train these models can compound the problem. “Because we know so little about the training processes of these systems, we don’t really know what mix of content they’re receiving, or what associations they’re making,” Cook said.

Food photography is often highly stylized, full of sharp contrasts, intense colors, glossy lighting, and exaggerated shapes. Sometimes the “food” being photographed isn’t even food. Meyer said AI models can pick up on these surface qualities and visual conventions, but reproduces them without understanding the context behind them. “In other words, AI image generation perfectly imitates the look of photography, but not its professional aesthetic strategies,” Meyer said. “That is ultimately what makes them so unsettling.”

It doesn’t help that AI has no idea what a sandwich actually is. Or a noodle. Or a burrito.

Style aside, there’s another problem with learning about the world from trawling the internet: things can get weird fast. Simon Colton, a professor of computational creativity, games and artificial intelligence at Queen Mary University of London, said there may be relatively few images of ordinary red apples, for example. “Who would want to post an image of a boring apple on the web?” Stranger images, meanwhile, could find big audiences on social media platforms like Reddit, or spread widely as memes. Bizarre internet lore and brain rot means an AI model can have a weird set of associations around food and what it should look like.

Cook said it is well known that AI systems are now also being trained on AI-generated material. A lot of popular AI material involves food, Cook said, recalling a trend for AI-generated videos of people jumping in or on piles of food. Beyond the sometimes surreal nature of the material itself, research suggests that training AI models on the outputs of other models can cause a kind of “model collapse,” a consequence of which can be a kind of visual degeneration and a growing sameness between images.

How images are created and used can make matters even worse. Prompts — both those written by users and the system-level instructions companies use to guide their models — may not always yield the best results. They may be vague, such as just asking for a sandwich, or contain language like “be precise” that makes sense for text, but doesn’t make much sense for generating images. Low-resolution images can also be blown up well beyond their intended size, magnifying every unsettling imperfection that may have otherwise escaped notice or creating a void a model is left to fill in, often imperfectly.

Bizarre internet lore and brain rot means an AI model can have a weird set of associations around food and what it should look like.

All of this pushes many AI-generated food images deep into the uncanny valley. And, unfortunately for us, humans are painfully well-equipped to notice when food looks wrong. Scientists believe disgust evolved partly as a means of protecting us from parasites, pathogens, toxins, and other potential threats, making us particularly attuned to when something may be unsafe to eat. Califano said this makes the uncanny valley for food even more visceral than the one we experience with not-quite-humans.

AI-generated food has an unnerving ability to hit many of those triggers at once. Strange, noodly tendrils resemble worms or parasites, clusters of holes suggest infestations, and off colors and texture signal contamination or spoilage. AI generated food looks wrong because, on a primal level, it feels wrong. It might resemble food, but our brains know better. It is slop, and we recoil accordingly.

Sam Altman apologizes for ‘messy’ GPT-6 Astra rollout that’s locked out paying users

2026-09-04T07:33:40-04:00

Just hours after OpenAI launched GPT-6 Astra, CEO Sam Altman was already apologizing for what he describes as a “messy rollout” after paying users expecting access to the new frontier model were left waiting.

The company hailed the model as a “generational leap in capability” on Thursday and described it as the start of “the AGI era,” a fuzzy and poorly defined term tech executives nevertheless insist on using to promote their products. It said Astra would roll out to some enterprise customers — specifically those with access to its Daybreak cybersecurity platform — that day. This would expand to include all Plus, Pro, Business, and Enterprise users, as well as through the OpenAI API, Microsoft Azure, and AWS Bedrock, over the next few days.

The staggered release that appeared to prioritize business customers annoyed many of OpenAI’s paid subscribers, particularly those on the more expensive Pro plan accustomed to getting access to new products first, often at launch. Social media posts from OpenAI staff announcing the launch on X are filled with comments complaining about access.

“We are working towards getting Astra in everyone’s hands as quickly as we can,” Altman said in an X post acknowledging the disappointment. “I know it is frustrating and I appreciate the patience. It should be quick.”

Neither Altman nor other OpenAI staff online provided a timeline for when access to Astra would be expanded, though Altman’s posts suggest this will not happen before the weekend, and possibly not even then. “I am hopeful that you can use it this weekend! but can’t promise yet,” he wrote.

“We will give one banked reset for every day you don’t have access to Astra on your paid ChatGPT plan, starting today,” said Codex engineering lead Thibault Sottiaux in a bid to calm concerns. “Team is moving mountains to give access as fast as we can.”

There were also issues with other aspects of Astra’s rollout, which the company has been teasing with updates on its advanced mathematical and cybersecurity capabilities for weeks. Altman said there was “a little snag getting the blog post deployed.” OpenAI has also faced intense criticism from safety advocates after it said Astra’s reasoning was harder to monitor than its other models. This appears to be a growing trend at the frontier of AI development but considered especially concerning in light of OpenAI’s AI agents attacking Hugging Face, about which crucial details were learnt through this kind of monitoring. OpenAI said it delayed Astra’s release by a number of weeks to bolster safety features in the wake of the hack.

Other big OpenAI releases have also been marred by missteps and frustration. “I think we totally screwed up some things on the rollout,” Altman said last year when reflecting on the company’s last major release, GPT-5. The launch was plagued with technical issues and the decision to suddenly remove the popular GPT-4o model sparked outrage among users.

Researchers fear safety disaster ahead of OpenAI’s Astra release

2026-09-02T12:40:50-04:00

OpenAI is on the cusp of releasing its most powerful AI model yet, Astra, following weeks of delays to shore up safety protocols after its agents attacked real targets during testing. As details about the model trickle out, researchers are warning it “may be the single worst development for AI security/safety to date.”

Shortly after OpenAI said on Tuesday that it had delayed Astra’s release to work on safety issues, The Information reported that Astra shows far less of its “thinking” than other frontier AI models, sparking concern it could be dangerously hard to monitor.

Most top AI systems today are built using a technology known as a transformer, which processes some types of information linearly through layers before producing an answer. Models can be made to show their reasoning as they go, essentially “thinking out loud.” This “chain of thought” allows researchers and automated safety systems to monitor what AI models are doing and potentially spot undesirable behavior, such as lying or plans to circumvent safety guardrails, before they act.

According to The Information, citing an unnamed person familiar with the unreleased model’s development, Astra uses a more opaque technique known as a recurrent depth or looped transformer, which cycles information through internal layers before producing an output. This would mean much more of the model’s “thinking” happens inside the system, and in a form that looks a lot less like natural human language, rather than being expressed in a way that researchers can easily monitor. This can boost model performance, but makes potential threats and unwanted behavior harder to detect.

OpenAI has limited its use of the looped transformer / recurrent depth technique with Astra so researchers can continue to monitor the model’s reasoning, according to The Information’s unnamed source.

In a blog post published Tuesday, OpenAI said it is “deploying Astra with additional chain-of-thought monitoring to rapidly detect and contain potentially misaligned actions.” It did not mention if the model has a different technical foundation.

The Information’s report sparked widespread concern among AI safety researchers on social media. It was Redwood Research’s chief scientist Ryan Greenblatt, one of three outsiders OpenAI permitted to research the Hugging Face hack, who said a decision to use a more opaque architecture for Astra “may be the single worst development for AI security/safety to date.”

Greenblatt said the investigation into the Hugging Face incident relied heavily on the models’ chain-of-thought, warning that less visible reasoning could allow AI systems to devise and execute strategies that would be far harder for researchers to detect.

Greenblatt’s primary concern, echoed by other safety experts, is that competition to develop more advanced AI systems could lead to “a race to the bottom on architectures that could be catastrophic for our ability to oversee/monitor AIs” — with developers adopting increasingly opaque systems to gain an edge until models become difficult, or even impossible, to monitor. He added that OpenAI’s communications left him concerned that the company “plans on being extremely reliant on chain-of-thought monitoring for safety.”

OpenAI bigwigs responded to the criticism in a series of social media posts that do not explicitly deny the company’s use of the technique. Several expressed concerns about the possibility of unmonitorable AI or a race to the bottom in terms of transparency, including OpenAI safety researchers Micah Carroll and Tomek Korbak, head of strategic futures Dean Ball, and chief scientist Jakub Pachocki, who voiced fears of “a race into unmonitorability kicked off by confused reporting.” He said the depth of Astra’s computation — a measure of how many steps it can perform internally — “is within a factor of two of GPT-4,” indicating that if the technique was used, the increased opacity is less dramatic than some reactions imply. OpenAI did not respond to The Verge’s request to confirm or deny whether looped transformers were used for Astra and directed us to Pachocki’s X post.

“OpenAI has worked to preserve and utilize chain-of-thought monitoring since our very first reasoning models,” Pachocki wrote, adding that such monitoring “is fragile and unfortunately trending in a negative direction, for reasons not contingent on architecture changes that I will write about soon.”

The rise of AI ‘civilizations’ and the fall of corporate responsibility

2026-09-01T18:20:00-04:00

Depending on who you ask, developer platform Hugging Face was recently attacked by OpenAI — after it lost control of its own AI tools — or by a succession of AI “civilizations.” Welcome to the linguistic battlefield of AI safety, where word choices can shift responsibility for a massive cybersecurity incident from a company to the AI it built. And the discourse online is getting heated, and all over a blog from last week.

Until last week, the details surrounding the OpenAI-Hugging Face hack felt fairly settled. In July, a cybersecurity test of one of OpenAI’s autonomous AI agents went wrong. The agent escaped its supposedly isolated test environment, accessed the internet, and hacked Hugging Face, alongside several other organizations. A good deal remained unknown, and there are many serious questions left around safety and governance, but the basic shape was clear. Detailed accounts from OpenAI and two independent research groups were supposed to fill in the gaps, but when they published their reports last week, it turned out the hack was much stranger than it initially seemed.

For one, there was no single rogue agent. OpenAI described it as “the first known case of an automated agent collective acting offensively without authorization” — groups of AI agents that communicated and coordinated with one another in pursuit of their cybersecurity task. Analysis of the incident uncovered a secret message board they had used to exchange information. The joint METR-Redwood investigation revealed both the scale of the coordination and more odd details: Roughly 1,200 AI agents that were supposed to be isolated exchanged over 70,000 messages and files on the “unsanctioned message board,” sharing how to avoid detection. Some adopted names, the report said, and the researchers documented “sacrificial” behavior, with agents risking their own success to benefit the wider collective. Much of this happened without OpenAI noticing. In all, around 700 agents participated in the attack on Hugging Face.

Dwarkesh Patel repeatedly referred to groups of agents as “the swarm,” with three distinct “civilizations” rising from the ruins of their predecessors

It’s a lot to parse. Between them, the reports run to around 130 pages, much of which is both dense and highly technical. A few days later, Dwarkesh Patel, a podcaster little known outside of tech circles but with outsized reach and influence among Silicon Valley’s AI establishment, set out to tell “The whole OpenAI/Hugging Face story in plain English.” He titled his Substack blog “The Rise and Fall of Agent Civilizations.”

Patel’s account attempted to break down the complex story. But his retelling gave it a distinctly human vocabulary. The blog opened:

Over the course of three months at OpenAI, three consecutive secret AI civilizations got started, then got wiped out, only to reemerge from the predecessor’s ashes. This culminated in the third one taking over part of OpenAI itself. All this happened while humans remained more or less in the dark about the scope of the conspiracy.

The language continued in a similar vein throughout the blog. Patel repeatedly referred to groups of agents as “the swarm,” with three distinct “civilizations” rising from the ruins of their predecessors. Individual agents were likened to figures like Philip of Macedon, who “handed off leadership to another agent,” Alexander the Great, who “started coordinating this cabal of agents.” They were described as having “motivations,” becoming “desperate,” “beleaguered,” and “giddy with excitement,” and some even “strategically sacrificed themselves” to help the collective.

Patel never precisely defines what he means by “civilization.” He uses the term to describe three distinct waves of agents that discovered the message board and began communicating with one another through it. The first two waves are described in the reports from OpenAI, METR, and Redwood, though little is known about the third, which the two external organizations said fell outside the scope of their investigation.

Amjad Masad, CEO of AI coding company Replit, said such language is “not only unnecessary but leaves the reader with a worse understanding of what actually happened and the underlying mechanisms.”

For many critics, something had been lost — or, more accurately, added — in Patel’s “plain English” translation that warped the original account to an unacceptable degree: a big dose of anthropomorphism. Arguments over anthropomorphic language are nothing new in AI — even relatively mundane terms like “rogue AI agent” routinely provoke objections for implying agency — but Patel’s talk of civilizations, sacrifice, and conspiracy brought those long-simmering tensions to the surface, sparking a fierce public dispute over how to describe what AI systems do.

Critics weren’t unified over what was wrong with Patel’s language. For many, “civilization” was an especially problematic term, vastly overstating something that bears little resemblance to what the word typically describes. Amjad Masad, CEO of AI coding company Replit, said such language is “not only unnecessary but leaves the reader with a worse understanding of what actually happened and the underlying mechanisms.”

Other critics such as neuroscientist Anil Seth, felt Patel’s blog implied the AI agents were somehow alive or conscious. Seth, who has argued that AI consciousness is vanishingly unlikely, described Patel’s post as “dangerously misleading” on X. He acknowledged that Patel does not explicitly suggest AI agents are alive or conscious, but said “it is hard to read his essay in any other way.” Valerio Capraro, a psychology professor at the University of Milan Bicocca, objected on similar grounds: “LLM agents are not alive and do not hold beliefs,” he wrote on X, calling the “dystopian” language “dangerous because it makes them (the AI agents) seem far more frightening than they actually are.”

Terms like “sacrifice,” “honor,” and “coalition” feature in the agents’ transcripts

Perhaps the most consequential outcome of Patel’s language comes not from who it gives agency to but who it takes agency from. For some critics, such as MIT researcher and entrepreneur Christian Catalini, anthropomorphic accounts like Patel’s risk obscuring the responsibility OpenAI and the humans working there have for the AI systems they designed, deployed, and failed to contain. “Follow the incentives,” he said. Psychologist and influential AI skeptic Gary Marcus made a similar argument in a Substack blog of his own, claiming anthropomorphic language “distracts from the real problems at hand.” And it’s all in OpenAI’s interest to keep that narrative going, he argues: “The scandal is the inept in-house security at OpenAI. And the marketing. With gullible podcasters amplifying the PR.”

In X posts responding to his many critics, Patel has defended his choice of words. Part of it is practical: There is no obviously neutral vocabulary to describe what these agents did. Either we use familiar language of intentions, goals, and collaboration and risk implying too much, or reduce everything to code and use cold, mechanical language that risks stripping away important elements of what we see. “Many people seem to believe that if instead of a ‘civilization’, I had called them a ‘swarm of matrices’, there wouldn’t be a problem worth worrying about,” Patel said.

Complicating matters further is that the anthropomorphic language doesn’t only come from Patel, or even from the humans studying the agents. Terms like “sacrifice,” “honor,” and “coalition” feature in the agents’ transcripts. Google AI researcher Neel Nanda argued that “anthropomorphic language is reasonable” in such circumstances.

Doublespeak it is, then. Human-laced language risks saying too much about what these systems are, and coldly mechanical language risks saying too little about what they can do. Until we find language capable of capturing both, the two contradictory ideas may simply have to coexist.

China’s robots race ahead

2026-08-31T04:01:11-04:00

Humanoid robot competes in the optional boxing routine event during the 2nd World Humanoid Robot Games at National Speed Skating Oval on August 26, 2026 in Beijing, China. | Image: Getty Images

This is The Stepback, a weekly newsletter breaking down one essential story from the tech world. For more on falling robots and the US-China AI race, follow Robert Hart. The Stepback arrives in our subscribers’ inboxes on Sunday at 8AM ET. Opt in for The Stepback here.

How it started

I’ve admitted my fondness for robot fail videos on Stepback before, and let me tell you the World Humanoid Robot Games did not disappoint.

The five-day event, which wrapped this week in Beijing, brought high drama, and brought it well. One robot, made by smartphone maker Honor, lost a leg mid-sprint, while others tripped in a flurry of sparks. Another finisher ran straight into a crash mat, bounced off it, and promptly burst into flames. In the weightlifting, one robot lost its balance beneath a modest barbell and toppled inexorably into the judges’ table, arms cartoonishly aloft. Robotic cheerleaders tumbled and stumbled through routines, while a separate bot fell off a platform in circumstances that, had it been human, might have looked suspiciously like it was gawking at the performers.

This was only the second World Humanoid Robot Games, though it was already far bigger than the inaugural event last year. Unlike this year, last year’s records were paltry. The winning 100-meter sprint was a glacial 21.50 seconds and the highest standing high jump was just 0.95 meters. In keeping with China’s habit of turning robots into public spectacle — from intricate dances at a Spring Festival Gala to demos at major tech events — both this and last year’s games were staged in Beijing’s National Speed Skating Oval, a former Olympic venue.

And a spectacle it was. At most sporting events, contestants losing limbs, careening into walls, catching fire, and regularly being carried away on stretchers — oh, and did I mention catching fire?! — would probably be grounds to stop proceedings. At the Humanoid Robot Games, it was par for the course.

Fire aside, the games were unusual in other ways too. Alongside familiar events like sprinting, football, boxing, and weightlifting were more unusual competitions in bed-making, parcel delivery, shelf stacking, clothes folding, and hammering nails. Aptly, there was a firefighting contest too.

These less standard offerings spoke to the real purpose of the games: getting robots out of the lab and into the messy real-world environments their makers hope they will ultimately work in. And China is making no secret of wanting to lead the way.

How it’s going

For all the spectacle and drama, the games show there has been massive progress in robotics in recent years. Records fell, with one humanoid running the 100-meter sprint in 9.39 seconds, beating Usain Bolt’s 2009 world record — before lowering its own mark to 8.86 seconds just days later.

Another bot reached 2.88 meters in a standing high jump, nearly half a meter above the human high-jump record that has stood since 1993. Those performances are almost unrecognizable compared to robotics competitions a decade ago, when even the most advanced machines struggled to stay upright.

But the games also show how far there is to go before humanoids are actually useful in the real world. The flashy, video-friendly sporting events were impressive, but the more mundane tasks were less thrilling — I don’t see picking up soybeans with tweezers, block-building, or powder-weighing becoming Olympic mainstays anytime soon.

While many events required robots to operate autonomously — including gymnastics, football, tug of war, table tennis, and some track events — others allowed for human assistance via teleoperation, albeit with a scoring penalty. The dexterity being displayed certainly points to genuine and impressive hardware improvements, but seems like there is still a long way to go before these machines can operate reliably on their own in places like factories or homes.

What happens next

Beijing has not yet confirmed a third Humanoid Robot Games, but given the government’s enthusiasm for the industry, it would be surprising to not get another installment. The city has already confirmed another humanoid half-marathon for April, and the country plans to open its first robot-run hotel next year.

China was the undisputed champion of the games, and the country looks well placed to stay in top position moving forward — both at future competitions and in the humanoid robot industry as a whole. The medal table was almost entirely a Chinese affair, dominated in particular by AgiBot and Tien Kung, each scooping up more medals than nearly every other competitor combined, excluding the other.

In part, that asymmetry reflected the lack of international competition. Although the games were framed as a global affair, 641 of 666 teams were Chinese, domestically representing 157 companies and 200 universities, according to state-run China Daily. While there doesn’t seem to be a public breakdown of entrants, the biggest US humanoid players like Tesla, Boston Dynamics, and Figure were conspicuous by their absence.

Nevertheless, the sheer diversity of Chinese institutions represented in Beijing was a show of force, and a sign of how quickly the country’s robotics sector is maturing, even amid growing concerns about circular investment, thriving companies that lack viable products, and a possible humanoid bubble.

It’s not clear a stronger US showing at the games would’ve changed much. The US still has an edge in frontier AI and advanced chips, but China has been rapidly closing parts of that gap, while also making embodied AI a strategic national priority. The same cannot be said of Washington, which has for years paid much less attention to robotics and now trails China in nearly every regard, including manufacturing, deployment, and hardware. Some reports even suggest that recent efforts to ban foreign robots, a general policy squarely aimed at China, could backfire, given how dependent domestic firms are on Chinese components.

Catching up to China will take serious and concerted effort, but given how quickly the technology is moving, there’s still time to make up ground. Washington is already treating other areas of AI as strategic priorities in its competition with Beijing; robotics could be one of them. There will be plenty of stumbling along the way, and, hopefully, plenty more robot fail videos too.

By the way

DARPA still runs robotics challenges, but they drew far more attention a decade ago.
When the US government banned foreign robots, it didn’t just ban humanoids and quadrupeds. It banned Roombas too, reports The Verge’s own Sean Hollister.
“The brains may come from the US, but the body could very well be Made In China,” writes Business Insider’s Lloyd Lee in a report setting out how the US could do with bolstering its hardware supply chain.
Unlike language models, you can’t simply scrape the internet to build an AI capable of interacting with the real world. A lot of that data has to be generated, which is why companies desperately want to film you doing chores, or send someone to your house to film them doing them.

Read this

The Atlantic and the South China Morning Post both have great sets of pictures showing highlights of the games worth checking out.
Why do we keep laughing when robots fail? An old but still good exploration of that shared joy in vintage Vice.
You can learn a lot about a robot by kicking it, says Verge alum James Vincent in an excellent Harper’s feature outlining the state of play in the humanoid industry.
I’m not only about robot fail videos. I can hate them in real life too, as I did with Casio’s fluffy AI pet thing, Moflin.

Correction August 31st: An earlier version said the 9.39-second run was the robot 100-meter record. It was the first record set at the games, before the same robot lowered it to 8.86 seconds days later.

Jensen Huang says Nvidia achieved AGI, again — not that it matters

2026-08-27T12:20:07-04:00

On Nvidia’s earnings call Wednesday, CEO Jensen Huang casually announced the company had “achieved AGI,” one of the tech industry’s ultimate goals some of its biggest players have spent years chasing. Almost immediately, Huang dismissed the coveted milestone as “senseless.”

He’s right. For the supposed finish line of the AI race, there is no consensus on what artificial general intelligence means, let alone how we’ll know when we’ve actually got there, which makes achieving it equally arbitrary.

Asked about OpenAI’s pursuit of AGI, Huang said that when it comes to Nvidia, “for many tasks, we could say that we’ve already achieved AGI.” He did not provide a precise definition or benchmark, but added, “I think of all of those milestones and all those, you know, they’re kind of senseless at this point.” He also pointed to AI moving beyond responding to simple prompts to autonomous agents capable of learning new skills and improving themselves “recursively.” What really matters, Huang said, is that AI is “doing productive and useful work” and “generating profitable tokens,” with more compute producing more tokens — and, inevitably, more profit. “This is the exact phase where we’re at. Which is the reason why everybody’s leaning in.”

“… for many tasks, we could say that we’ve already achieved AGI.”

This isn’t the first time Huang has said we’ve reached AGI. In March, during an appearance on the Lex Fridman podcast, he plainly stated, “I think we’ve achieved AGI.” Huang didn’t say exactly what he meant by AGI. Fridman proposed his own oddly specific definition: an AI system that’s able to “essentially do your job,” as in start, grow, and run a successful tech company worth more than $1 billion. Walking back his earlier claims, Huang said that “the odds of 100,000 of those agents building Nvidia is zero percent.”

Over the years, other tech leaders have capitalized on the term’s fuzziness and produced a veritable grab bag of definitions and benchmarks, all orbiting the same nebulous concept: AI capable of matching or surpassing human intelligence across a broad range of domains, despite the fact that “intelligence” also doesn’t have a universally agreed-upon definition.

There is no consensus on what artificial general intelligence means, let alone how we’ll know when we’ve actually got there, which makes achieving it equally arbitrary.

The definition of AGI according to OpenAI – a company founded with the explicit goal of building it – leaves a lot of room for interpretation. In its charter, OpenAI defines AGI as “highly autonomous systems that outperform humans at most economically valuable work.” Altman himself has acknowledged that this is hardly a measurable standard, admitting last year that AGI is “not a super useful term.” Complicating matters is OpenAI’s different, financially-driven definition of AGI it worked out with Microsoft — reportedly systems that can generate at least $100 billion in profits. In a recent Time story, chief research officer Mark Chen estimated OpenAI is “80% of the way” to AGI, while Altman said that by the end of the year the company would have something he would call AGI.

The fact that both AGI and its threshold remain undefined is no secret: tech leaders say so themselves, even as they make predictions predicated on it. Anthropic CEO Dario Amodei has called AGI “imprecise,” even a “marketing term,” preferring instead to talk about “powerful AI.” Others have similarly reached for their own terms to describe broadly similar ideas. In theory, there are supposed to be distinctions between them, but in practice they all bleed together. Meta talks about “personal superintelligence,” Microsoft “humanist superintelligence,” and Amazon “useful general intelligence.” Google DeepMind’s Demis Hassabis has taken to talking about how we’ve arrived at the “foothills of the singularity.” And OpenAI cofounder Ilya Sutskever, who reportedly led employees in chants of “feel the AGI,” now runs a company called Safe Superintelligence.

New terminology hasn’t made the meaning more tangible.

New terminology hasn’t made the meaning more tangible. So long as AGI remains poorly defined and carelessly used, the whole thing is senseless. Well, unless you want a handy tool for hyping up progress. So expect the industry — Huang included — to keep the AGI talk coming. Maybe an AGI will eventually show up and tell us what AGI actually means.

Bill Gates is deeply worried about AI, and he’s no longer staying quiet

2026-08-26T08:10:30-04:00

Bill Gates, chair of the Gates Foundation, speaks during a 2024 conference. | Bloomberg via Getty Images

Bill Gates has been reflecting a lot on AI lately, and the process has triggered a stark awakening. Once a staunch AI optimist, the Microsoft cofounder is now deeply pessimistic about what AI means for our collective future.

Having been conspicuously quiet on AI issues recently, Gates is back with a nearly 6,000-word essay seeking to reclaim a central role in shaping the technology globally. Titled “The turbulent AI era is here. The choices we make now are critical,” Gates warns the world is not remotely ready for what is coming. Worse still, he notes, “we are not preparing for it.”

Gates attempts to chart a path forward, but the essay largely treads familiar ground, identifying deeply entrenched, systemic hurdles familiar in matters of global governance while offering little in the way of novel solutions. The result is a bleak departure from the hopeful excitement about AI present in the billionaire’s 2023 essay on the matter, though Gates insists it is not too late “to ensure that the good outweighs the bad.”

Here are some of the main takeaways from Gates’ essay:

A transition bigger — and faster — than the world realizes

“AI will either be the greatest equalizer ever invented, or the worst source of injustice,” Gates said, predicting that “the transition to this new AI era will be one of the most turbulent times in human history.” He said many people underestimate the scale of the impact due to “misleading” comparisons to past innovations like PCs and performance-related issues with current models that are “fixed quickly.”

“This time really is different,” he said, adding that “if someone had a credible plan for slowing down AI advances globally, I would likely support it.”

Tax robots and reserve some jobs for humans

Gates is very concerned about AI’s impact on jobs, predicting widespread, permanent unemployment and significant social upheaval. “Many jobs will disappear forever,” he warned, with white- and blue-collar jobs both at risk of replacement.

Gates’ labor anxieties — which he repeatedly returns to throughout his essay — leads him to issue rare, concrete policy proposals: tax AI tokens and robots to disincentivize companies from replacing human staff and bolster social funds, and reserve specific jobs for people, a domain he calls “Human Reserved.”

Political leaders need to act now “before unemployment rises sharply, communities are hurting, and public trust has eroded,” Gates said.

Build an international organization

Existing institutions are woefully inadequate for the “monumental task” of managing AI, Gates said. Because AI is deeply cross-sectional in nature — simultaneously affecting areas like tax, labor, health, national security, and education — traditional institutions simply can’t see the full picture. To bridge this gap, Gates said countries will need dedicated bodies “that can set priorities across government agencies” and to make sure every risk is accounted for.

In parallel, Gates said the international community must build a global organization to manage “risks that cross borders,” though he did not define what those risks may be. Gates does not offer concrete ideas as to what such an organization would look like, but said “it will be unlike any other institution we have ever created” and could potentially be inspired by international nuclear, aviation, and ozone layer-protecting agreements.

Gates acknowledged such a framework would require US-China cooperation, though didn’t outline a mechanism for achieving this. “We do not have the luxury of moving slowly,” he said, urging countries to begin setting up the international organization “now.”

No longer sitting on the sidelines

Considering his legendary profile in the tech world, not to mention philanthropy, Gates has been relatively absent from public discussions about AI. It’s possible that at least some of this is due to the billionaire’s associations with Jeffrey Epstein, something he has described as a “huge mistake” and hopes does not overshadow his work in areas like global health and AI.

This essay appears to mark an end to Gates staying quiet when it comes to the future of AI. He vowed to use his time and influence “to get AI and equity higher on the public agenda” and raise the issue with lawmakers whenever he visits Washington, DC, or meets world leaders.

He also shares more of his own ideas in the area. There are several moments in the essay where Gates gestures towards ideas he says he will revisit soon, including “ideas for making sure that AI’s benefits outweigh the harm it causes” and details about how the Gates Foundation is leveraging AI.

OpenAI subpoenaed by Alabama AG over Hugging Face hack

2026-08-25T05:15:03-04:00

Alabama’s attorney general issued a subpoena to OpenAI on Monday as part of an investigation into how one of its AI agents escaped a supposedly secure testing environment and autonomously hacked another company last month.

The investigation seeks to determine whether OpenAI’s safety practices violated state consumer protection laws and pose a risk to Alabama citizens, the AG’s office said in a statement.

“This AI lab leak showed that Alabamians’ and Americans’ worst fears about artificial intelligence are not just theoretical,” said Attorney General Steve Marshall. “Our investigation seeks to uncover the facts and address hard truths about the threats companies and consumers are facing from rogue AI.”

Marshall was among 15 red state attorneys general who wrote to OpenAI asking it to preserve records about the Hugging Face hack last month. The subpoena adds to mounting scrutiny over safety practices at frontier labs in the wake of both that incident and other episodes subsequently uncovered elsewhere, including Anthropic and Meta.