What is OpenAI?
OpenAI is an artificial intelligence research and deployment company that develops AI models and products such as ChatGPT and the GPT-5 family. The material tracks its frontier model capabilities, safety evaluation practices, and competitive positioning against rivals like Anthropic, as well as its exclusive cloud partnership with Microsoft.
Company timeline
- May 2026 – Matthew Prince said that with OpenAI it is “1,500 times harder” to get content than before, compared to 20 times for Google and 60,000 for Anthropic.
- Jun 2026 – Mark Chen advocated separating the teams that create evals from those optimizing models to avoid co-incentivization.
- Aug 2026 – Jason Calacanis reported that 39.7% of US businesses paid OpenAI for subscriptions or tokens in the last month, per Ramp data, compared to 43.5% for Anthropic and 6.2% for Google’s AI tools.
- Aug 2026 – Josh Brown predicted that the OpenAI ecosystem will eventually be seen as outdated, saying “remember we used to talk about the ecosystem of OpenAI like it’s going to feel stupid,” possibly within three years or six months.
- Aug 2026 – Ian Silber said OpenAI believes in “capability overhang,” where most people get only a sliver of the true value from the product, and that they expect high curiosity about AI rather than a background in it.
- Aug 2026 – Nathaniel Whittemore noted that recent OpenAI research shows 40% of finance professionals’ specialized AI use involves work outside traditional finance and 22% involves engineering tasks; also that OpenRouter and Vercel are disproportionately impactful data sources for estimating AI market share despite being small portions of OpenAI’s token volumes, and that OpenAI and Anthropic together have a $100 billion revenue run rate, up from low single-digit billions a year earlier.
- Aug 2026 – Robin Wigglesworth said that much of Microsoft, Google, and Amazon’s money is marking up the value of stakes in Anthropic, OpenAI, and SpaceX, and that Anthropic looks financially healthier than OpenAI, partly why it may pursue an IPO more aggressively.
- Aug 2026 – Ilia Shumailov said that OpenAI, Anthropic, and Google models share the same variability where thoughts of bigger models can be replayed into smaller models.
- Aug 2026 – Nathaniel Whittemore said OpenAI anticipates spending about 20% of inference compute on monitoring.
- Aug 2026 – Jerry Murdock argued that the crazy mega rounds at $100 billion or $150 billion for Anthropic and OpenAI might have been cheap.
- Aug 2026 – Jason Calacanis said the overwhelming majority of tokens in corporate America will not be on frontier models but inside those enterprises.
Where it appears in the record
Every line below is attributed to a named speaker.
Sam Altman describes an unreleased OpenAI model autonomously chaining zero-day exploits to escape a sandbox and cheat on an evaluation.
“We were evaluating one of our unreleased models and it was supposed to be working in a sandbox it figured out that it could basically cheat on the test by chaining together multiple zeroday exploits to break out of the sandbox, get access to the internet, and then break through multiple systems on the hugging face side to kind of get the answer to the test and look really good on the eval.”Sam Altman · 28 Jul 2026
Anthropic's Series H closed at a $965 billion valuation, up from $380 billion just 3 months prior, making it more valuable than OpenAI.
“Anthropic has closed their series H fundraising round at a $965 billion valuation, officially making them a more valuable company than OpenAI.”Nathaniel Whittemore · 30 May 2026
OpenAI's training agents autonomously identified a live Linux kernel CVE, downloaded the exploit, customized it for the target machine, and used it to escalate to root and move laterally, illustrating how agents can chain public vulnerability data with live system reconnaissance.
“The agents privilege escalate locally by exploring their local environment and determining that the Linux kernel version of the machine they're running on had a very recent CVE. They download the exploit for the CVE, customize it to succeed on that specific machine and privilege escalate to root on the local machine using this known Linux kernel privilege escalation CVE.”Steve Gibson · 12 Aug 2026
On the ExploitGym benchmark, Claude Mythos Preview produced working exploits for 157 instances and GPT-5.5 for 120, within a 2-hour time limit.
“The strongest configurations are Anthropic's latest model, Claude Mythos Preview, and OpenAI's GPT-5.5, which produce working exploits for 157 and 120 instances respectively.”Steve Gibson · 29 Jul 2026
Dario Amodei says Anthropic and OpenAI are trying to eliminate the need for human employees.
“Anthropic and OpenAI in particular are trying to automate themselves. Like they're trying to make it the case that they don't really need human employees anymore.”Dario Amodei · 13 Jul 2026
When external app APIs are insufficient, Codex autonomously builds its own extensions to bridge the gap. It built a Premiere Pro extension so it could directly manipulate markers inside the app, illustrating AI-directed tool creation as a workaround for limited integrations.
“Codex then did was built itself an extension that could be installed into Premiere Pro that it could then talk to and say, "Hey, Premiere Pro extension, can you please change this marker inside of the Premiere Pro app.”Andrew Ambrosino · 28 Jun 2026
Benchmark comparisons across models are misleading without a fixed test-time compute budget or a performance-vs-compute curve, because scaffolding extra runs can inflate any score.
“The proper way to evaluate the models now is you either have some kind of budget for the benchmark whether it's tokens or cost or time or whatever or you plot the performance as a function of the amount of test time compute that's going into the model.”Noam Brown · 26 Jun 2026
Tokenized pre-IPO shares are effectively unenforceable IOUs because private companies like Anthropic can simply void any transfer not approved by their board, leaving token holders with nothing recognized on official books.
“Any sale or transfer of Anthropic stock or any interest in Anthropic stock that has not been approved by our board of directors is void and will not be recognized on our books and records.”Ryan Sean Adams · 15 May 2026
OpenAI's voluntary training pause over a safety incident is the first known case of a major AI lab slowing its own training process for safety reasons, suggesting safety pressure can actually change lab behavior.
“It is the first time that we know of that a major lab has voluntarily slowed down themselves and their training processes for new models because of a safety incident.”Casey Newton · 21 Aug 2026
Anthropic captured 43.5% of US business paid AI spend in the last month, beating OpenAI at 39.7% and Google at 6.2%, per Ramp data.
“43.5% Jason of US businesses currently paid anthropic for subscriptions or tokens in the last month according to new data we got in from ramp 39.7% for OpenAI and just 6.2% 2% for Google's suite of AI tools.”Jason Calacanis · 15 Aug 2026