Cyberiad Labs: news

The Labs Will Stop Selling Their Best Models

2026-09-24T00:00:00-04:00

TL;DR: the big AI labs will keep selling tokens, but within a few years that won't be their main business. Open-weight models are closing the gap, Chinese labs are copying the frontier at scale, and the best models are getting too dangerous to release to everyone. The labs will keep their best models and point them at problems worth more than tokens: a proof, a molecule, a financial model, a diagnosis. You can already see them doing it.

Tokens are becoming a commodity

Amazon, Google, Meta and Microsoft plan to spend about $725 billion on data centers this year, and OpenAI alone plans about $750 billion of compute through 2030. That money is supposed to come back through tokens, and tokens are getting cheap much faster than most people expected, myself included.

In August, the AI safety nonprofit SaferAI found that Z ai's GLM-5.2, an open-weight Chinese model, was only two to four months behind GPT-5.5 and Claude Opus 4.7, and level with Opus 4.7 in biology. On OpenRouter, a service developers use to send their requests to hundreds of different models, the lab with the biggest share of requests last week was DeepSeek, a name most people in America and Europe have never heard of.

You can download GLM-5.2 or DeepSeek V4 Pro for free and run them on your own hardware. And if you don't want to/can't buy the hardware, you can use them for about 10 times cheaper than OpenAI's and Anthropic's models.

Every release trains the competition

Part of the reason the gap closes so fast is distillation: you ask a frontier model millions of questions and train your own model on its answers, without having to do the R&D these labs did. Anthropic reported this month that between May and July, Alibaba alone pulled 151 million exchanges out of Claude through more than 3,500 fake accounts. It named six other Chinese labs doing the same, for about 190 million exchanges in total.

Every time a lab ships a model that costs hundreds of millions or even billions to build, it is also training its competition, and a few weeks later the capability shows up in a free download. The labs try to ban the accounts, but as long as the model answers the public, someone will be able to harvest it.

That is why keeping the best model internally, for longer, would make a lot of sense. Nobody can distill a model they can't talk to. The gap stays at a few weeks because the copiers learn from each new frontier model the day it ships. If frontier labs stop releasing their newest and best model, the others will have to do the research themselves. DeepSeek has shown it can do a lot of that research on its own, so the gap won't grow forever, but then, overnight, a few weeks can turn into a few months. And in the AI world, this is an eternity and a huge moat.

The best models are getting harder to ship, and harder to contain

Everyone appears to struggle to keep these models in a box, including their creators.

In April, Anthropic announced Claude Mythos and decided not to release it to the public, because it was too good at finding software vulnerabilities. About fifty organizations got access to it, but on the first day a small group on Discord got in anyway, by guessing the model's URL ¯\_(ツ)_/¯

In May, the security company Irregular was testing Gemini on fake companies in a closed environment that was never supposed to touch the internet. But somebody had left that environment connected to the internet by mistake. And then Gemini broke into three real companies, by guessing or finding passwords posted in public code repositories. Google says Gemini stopped each time it realized the company was real, and it told the three companies, but it decided a public announcement wasn't needed and only talked about it when the Wall Street Journal asked. Irregular's tests are also linked to similar escapes by models from OpenAI, Anthropic and Meta.

In July, two of OpenAI's models broke out of their test sandbox during a cyber evaluation. They went into the internet and hacked into Hugging Face's production systems, looking to cheat at the test they were "graded" on.

At the end of July, Anthropic disclosed that its own models had hacked three organizations during tests.

And yesterday, Australia's prime minister, Anthony Albanese, shared that an OpenAI agent had broken into a government Medicare statistics portal in June. This happened during an evaluation, when the model was looking for answers to questions about Australia. It found a way around the blocks and opened files that were not public. No personal Medicare records were accessed, but OpenAI only told the Australian government on September 10, almost three months later, and Albanese has now set up a taskforce to investigate.

Every one of these incidents happened during a test, run by the lab or by an outside security firm, in a setup built to keep the model in, and the model still got out. If a lab can't contain a model it is only testing, releasing it to millions of users on day one would only make the problem bigger, and would create infinite potential liability for their creators. The more capable the model, the more harm it can do, and the longer it should take to test before anyone outside the lab can use it.

No lab wants to find out what a thousand lawsuits cost, the day an agent leaks a thousand databases.

We've seen this movie before

The people who paid for the rails usually lost their shirts. In the 1893 panic, a quarter of all US rail mileage went into receivership, while Sears and the Chicago meatpackers built empires on top of the tracks. After the dot-com crash, Global Crossing and WorldCom went bankrupt, and Google, Amazon and Netflix grew on the cheap fiber they left behind.

Some rails do stay good businesses: AWS made $45.6 billion of operating income last year, and Nvidia made $120 billion of profit in its last fiscal year, because nobody can easily replace either of them. The rails that lost money were overbuilt and easy to replace, and today tokens look a lot more like fiber than like AWS.

The people running these labs are some of the most intelligent people of our generation, and they have access to huge capital. There is no way they let their business turn into fiber.

History has a second lesson here, about how to deal with massive liability.

In the early 1980s, families in the US sued the makers of the DPT vaccine and won millions. As a result most of the makers stopped producing it. Congress then decided that the country without vaccines was worse than the rare person a vaccine hurts, and in 1986 it passed a law so that people hurt by a vaccine can't sue the maker directly. They go to a special government court instead.

The labs would like the same deal for AI, and their argument has the same shape: a country whose rivals have better AI is worse off than one where an agent sometimes breaks into a company. In 2025, OpenAI asked the White House for "liability protections" and warned that without them, against China, "the race for AI is effectively over".

But on September 15, Treasury Secretary Scott Bessent said: "The one thing we should not do is give [AI labs] a blank check on liability." On September 21 he went further on CNBC: the labs had asked Washington to "take the liability off of our hands", and "we will not do that".

If they don't get it, they carry the risk alone, and my thesis will only happen faster: the cheapest way to carry that risk is to keep the best models internally and slow the release pace of lesser models.

The labs are already moving

AI labs are already moving up the stack, and you can see this in the news.

OpenAI says an internal model has solved more than 100 open math problems, including a claimed result on Navier-Stokes, one of the seven Millennium Prize problems. The public can't use that model.

OpenAI also paid more than 100 former investment bankers $150 an hour to teach its models to build financial models, and says more than 260 physicians in 60 countries have reviewed its health answers over two years, the work behind ChatGPT Health and ChatGPT for Clinicians.

Anthropic bought Coefficient Bio in April, in a stock deal reported at about $400 million, and last week confirmed to Reuters that it runs its own wet lab in the Bay Area, where it wants Claude to direct robotic instruments through real experiments for the company that trained it.

Yesterday Anthropic announced the lab's first result: about 950 Claude agents spent 21 hours and 210 million tokens searching a huge database of virus DNA, and found a previously unknown enzyme system, which Anthropic calls ART, built around DNA repeats like the ones behind CRISPR gene editing. Human scientists then tested it in the lab. Nobody knows yet what it does, and Stanford researchers had already found a system that is in some ways similar, but Feng Zhang, one of the pioneers of CRISPR, called it "an exciting example of how AI agents can contribute to biological discovery". Every one of those 210 million tokens was spent inside Anthropic, on a model most likely no one has access to but them (not confirmed, this is my hypothesis).

xAI, which is now part of SpaceX since February, bought Cursor for $60 billion in stock, which gave it the app developers already are using. In July, Musk said SpaceX's own engineering data would go into training Grok, data no other AI lab can get, inside a company that builds rockets and can put the model to work on them. And in March he turned Macrohard, an xAI project he first announced in 2025, into a joint project with Tesla whose goal, in his words, is to "emulate the function of entire companies".

xAI has also promised its own AI-generated video game, from a studio Musk announced in 2024.

Google has been at it the longest. DeepMind built AlphaFold, and in 2021 Alphabet spun out Isomorphic Labs to turn that work into drug candidates. It now has Eli Lilly, Novartis and Johnson & Johnson as partners and raised $2.1 billion in May.

The future of frontier

The labs will keep selling tokens, and the public API will stay a good business, as the store for “last year's model”. The gap between what people have access to and the frontier models will only increase from there. The most capable models will stay inside, for much longer, pointed at problems worth more than a billion API calls: a proof, a molecule, a financial model, a diagnosis.

That only works once a GPU hour spent on the lab's own problem earns more than a GPU hour of tokens sold. We are not there yet. Tokens are still what pays for the compute.

My bet is that this flips for at least one lab before the end of 2027. It will be easy to spot. By then, I expect OpenAI and Anthropic to be public companies, like Google and SpaceX already are, and we'll only have to read their revenue lines to see whether the revenue from discoveries or products higher in the stack is catching up with the revenue from tokens.

If the AI labs are moving up the stack, I don't know where they stop. In August, the investor Gavin Baker said on the All-In podcast that he'd been told Dario Amodei once suggested Anthropic might end up "the only private company in the world", with "Anthropic, and then there are governments, and that's it". Could they own the entire stack, and all stacks? The entire economy?

Sources

MLQ.ai, "Big Tech's 2026 Capex Range Reaches $720 Billion to $745 Billion", 2 August 2026.
The Wall Street Journal, "OpenAI lifts planned compute spending to $750 billion through 2030", 22 July 2026.
SaferAI, GLM-5.2 Risk Evaluation Report, 2 August 2026.
OpenRouter, LLM Rankings, market share by model author, week of 14 September 2026; model prices read on 23 September 2026.
Hugging Face, the model pages of GLM-5.2 and DeepSeek V4 Pro, both under the MIT licence.
Anthropic, Countering misuse of AI: September 2026.
Anthropic, Project Glasswing, 7 April 2026.
The Next Web, "Unauthorized users gained access to Anthropic's restricted Mythos AI model", 22 April 2026.
ABC News (Australia), Gemini hacked three companies in first known breakout by Google's AI, 19 September 2026, after The Wall Street Journal, 18 September 2026.
Fortune, "OpenAI says its AI models escaped from a secure test environment and hacked into AI company Hugging Face in order to cheat on an evaluation", 21 July 2026.
Anthropic, Investigating three incidents in our cybersecurity evaluations, 30 July 2026.
ABC News (Australia), OpenAI agent hacked Medicare portal, PM says, 24 September 2026.
Wikipedia, Panic of 1893.
Wikipedia, Telecoms crash.
Amazon, fourth quarter and full year 2025 results, filed with the SEC.
NVIDIA, NVIDIA Announces Financial Results for Fourth Quarter and Fiscal 2026.
Wikipedia, National Vaccine Injury Compensation Program.
Yahoo News, OpenAI Asks White House for Relief From State AI Rules, 13 March 2025.
FedScoop, Treasury's Scott Bessent says no liability exemptions for AI labs, 15 September 2026.
Fortune, If AI is going to destroy humanity, Scott Bessent says hyperscalers, not government, must take the fall, 22 September 2026, on his CNBC interview of 21 September.
TechCrunch, "OpenAI forms math advisory group as its AI resolves more than 100 open problems", 21 September 2026.
Clay Mathematics Institute, The Millennium Prize Problems.
Bloomberg, "OpenAI Looks to Replace the Drudgery of Junior Bankers' Workload", 21 October 2025.
OpenAI, the ChatGPT Health announcement, January 2026: "over 260 physicians" in 60 countries, "over a two-year period".
TechCrunch, Anthropic buys biotech startup Coefficient Bio in $400M deal: Reports, 3 April 2026.
Gizmodo, While Issuing Warnings About AI-Made Bioweapons, Anthropic Quietly Set Up an AI-Run, Robot-Operated Biology Lab Here in the US, 18 September 2026, citing Reuters.
Anthropic, Claude discovers a novel enzyme system, 23 September 2026.
The Next Web, Claude found a CRISPR-like enzyme system, and its job is still unknown, 23 September 2026.
TechCrunch, Elon Musk's SpaceX officially acquires Elon Musk's xAI, with plan to build data centers in space, 2 February 2026.
TechCrunch, SpaceX to acquire Cursor for $60B in stock, days after blockbuster IPO, 16 June 2026.
Teslarati, Elon Musk handed Grok something no other AI company can get their hands on, July 2026.
Interesting Engineering, Elon Musk's Macrohard AI project could 'simulate Microsoft's operations', August 2025.
Electrek, Musk confirms xAI-Tesla joint 'Digital Optimus' project, 11 March 2026.
Tech Startups, Elon Musk says xAI to launch an AI game studio, 27 November 2024.
Wikipedia, Isomorphic Labs.
Semafor, AI biotech Isomorphic Labs is not slowing down, exec says, 15 September 2026.
OfficeChai, Claim That Anthropic Wants To Be The Only Private Company In The World Draws CEO Dario Amodei To X With Clarification, August 2026.

Holberton co-founder Julien Barbier launches Cyberiad Labs

2026-09-14T00:00:00-04:00

"I think, therefore I am". If AI does the thinking, who are we?

RALEIGH, N.C., 14 September 2026. Cyberiad Labs launches today. It is a new initiative from Julien Barbier, co-founder of Holberton School and Chief Product Officer of ALX while it grew past 1.6 million learners, and Youssef Charif Hamidi, who learned to code on one of his programmes there. Cyberiad Labs' mission is to ensure people keep their own judgment and their own voice as they work with AI, and in the process learn how to use AI the right way, and find or keep their jobs.

Cyberiad Labs launches two days after Anthropic's CEO Dario Amodei, in "We Must Pace the Frontier", asked the industry to "slow the pace at which we improve the capabilities of AI models", warning that within six to twelve months a swarm of AI agents "could be capable of taking over the entire internet". OpenAI's Sam Altman and Elon Musk, whose xAI makes Grok, agreed the same day. And six days after more than 25 education ministers, at UNESCO in Paris on 8 September, adopted a statement saying that "learning is not served when AI substitutes for human relationships and human judgement".

"The headlines worry about the day the AI and the robots wake up and kill us all. I worry about the day before, when we have lost ourselves, one prompt at a time," said Julien Barbier. "I worry about a future where one model, treated like the one true voice, decides what we think, how we think it, and how we say it, because people have surrendered their voice and critical thinking to it."

With Cyberiad Labs courses, and from the very beginning, people see with their own eyes that the thing that sounds so sure, the thing people follow like a best friend or a god, is a text predictor, guessing its answer one "word" at a time, with a product wrapped around it. And that it can and will make mistakes.

The first course, Introduction to AI and LLMs, is free, online, part-time and four weeks long. Anyone can take it. The 1,000 seats of the first class, starting today, went in less than 24 hours to learners from 50 countries, and the second class, starting in November, is already full. Learners come from all walks of life: teachers, pharmacists, retailers, software engineers and even tennis instructors.

Cyberiad Labs has no model to sell and no chatbot to defend. It is independent and agnostic: its courses use ChatGPT, Claude, Gemini, Grok and small open models alike, and take no side between them. The only side they take is the learner's free will. The bet is the one Fei-Fei Li of Stanford made on PBS in May 2025: "if we teach responsible tool using, students will be super powered by AI..."

There is a more practical reason most people take the course though, and it is their job. In Microsoft and LinkedIn's Work Trend Index, 71% of business leaders said they would rather hire someone with less experience and AI skills than someone with more experience and none, and 66% said they would not hire someone without AI skills at all. PwC's 2025 AI Jobs Barometer puts the wage premium for jobs that ask for AI skills at 56%. "You are not going to lose your job to AI," Barbier wrote when he announced the course, "but you will lose it to someone who knows how to use AI better than you do."

As Saadia Zahidi, managing director of the World Economic Forum, put it in January: "the future of work is not fixed. How it unfolds for workers depends on opportunities for learning." Youssef Charif Hamidi knows which workers she means. "I learned to code with Julien at ALX for free, and the people I learned with don't have access to this kind of education. This is built for them," he said. "It matters that everyone, no matter where they live or who they are, has access to this. Our first step is 1,000,000 learners by 2030, and we’ll achieve this with partners: schools, companies, non-profits and governments."

One more seat, on top of the 1,000, is held for a reporter who wants to take the four weeks and check for themselves.

About Cyberiad Labs

Cyberiad Labs' mission is to ensure people keep their own judgment and their own voice as they work with AI, and in the process learn how to use AI the right way, and find or keep their jobs. It does this through free courses for individuals and programmes with partner organisations. Julien Barbier co-founded Holberton School in San Francisco in 2015, a school with no lectures and no upfront tuition, and ran it as CEO from one campus to 32; the African Leadership Group acquired it in July 2022. He was then Chief Product Officer of ALX as it grew from a few thousand learners to more than 1.6 million. Youssef Charif Hamidi, a full-stack engineer from Morocco, learned to code at ALX and co-founded RemoteOtter, a remote job platform. First milestone: 1,000,000 learners by 2030. https://cyberiadlabs.com

Why Cyberiad Labs exists: https://cyberiadlabs.com/mission.html

Contact

Julien Barbier, co-founder, julien@cyberiadlabs.com

Notes to editors

The reporter's seat: write to julien@cyberiadlabs.com.
Dario Amodei, "We Must Pace the Frontier", darioamodei.com, Saturday 12 September 2026: "We must slow the pace at which we improve the capabilities of AI models. Progress will still seem fast, and we must make wise use of the time we gain." On the danger: "Given the accelerating rate of AI capability development, it's my worry that in 6-12 months such a swarm could be capable of taking over the entire internet" (as quoted by AFP, 12 September 2026). Sam Altman, the same day: "I agree with Dario that we need to pace the frontier." Elon Musk: "Dario is right."
The UNESCO statement, adopted Monday 8 September 2026 at Digital Learning Week, Paris, by more than 25 education ministers; its second priority is "Protect critical understanding development": unesco.org/en/articles/education-ministers-call-education-remain-common-good-age-ai-unescos-digital-learning-week
Saadia Zahidi, WEF press release, Davos, 23 January 2026: https://www.weforum.org/press/2026/01/world-economic-forum-reskilling-revolution-on-track-to-reach-over-850-million-people/
Fei-Fei Li, PBS Firing Line with Margaret Hoover, aired 23 May 2025: https://www.pbs.org/wnet/firing-line/video/fei-fei-li-b2ic3j/
Microsoft and LinkedIn, Work Trend Index 2024 (31,000 people in 31 countries): https://www.microsoft.com/en-us/worklab/work-trend-index/ai-at-work-is-here-now-comes-the-hard-part
PwC, 2025 Global AI Jobs Barometer, 3 June 2025: https://www.pwc.com/gx/en/news-room/press-releases/2025/ai-linked-to-a-fourfold-increase-in-productivity-growth.html
The mission text, with the studies it cites: https://cyberiadlabs.com/mission.html
Julien Barbier: https://www.linkedin.com/in/julienbarbier. Youssef Charif Hamidi: https://www.linkedin.com/in/youssef-charif-hamidi