
He Risked Everything To Warn You: No One Is Ready For What's Coming, And The AI Companies Know It!
In this interview, former OpenAI employee Daniel Kokotajlo warns about the risks of current AI development.
Key Points:
- AI companies (OpenAI, Anthropic) are racing toward superintelligence by the end of the decade – with a high chance of loss of control or power concentration.
- Kokotajlo forfeited $2 million by refusing to sign a non-disparagement clause – on principle.
- His forecast “AI 2027” outlines a possible path: automation of AI research, recursive self-improvement, and eventual AI takeover.
- His alternative “AI 2040 Plan A” proposes a regulated, transparent approach with slower progress, international agreements, and a citizens’ dividend (e.g., $25,000 per person initially).
- The technology is uninterpretable; we cannot see inside neural networks – mechanistic interpretability is a hope.
- Personal impact: Kokotajlo decided not to have more children due to uncertainty, and urges the public to wake up and take action.
Conclusion:
AI development is the most important issue of our time. Kokotajlo calls for political regulation, transparency, and public awareness to avert disaster.






