• Why he believes the AI industry’s safety narratives are rationalisations
• The emails that reveal why AI CEOs are racing each other • What changed inside OpenAI after ChatGPT launched
• The secret clause that would have cost him 2 million dollars
• Why he doesn’t believe Sam Altman’s explanation
• How OpenAI and Anthropic are trying to automate themselves
• The moment in AI 2027 when the AIs stop taking orders
A former OpenAI researcher who refused to sign away his right to criticise the company says the founding stories behind the biggest AI labs are rationalisations, and that what these companies say in public differs entirely from what happens inside. Daniel Kokotajlo, one of the world’s leading AI forecasters and lead author of AI 2027, explains what he saw at OpenAI and why he walked away.