Latest publications
Jack Clark on AI's Future: An Essay on Positive Singularity
Journalist Jack Clark publishes an essay and fictional story on how humanity should prepare for AI development and plan for a positive outcome of this process.
Google, China, and the British AI institute: how models are learning to break down, hack, and jam
Three studies showed where frontier AI is heading: Google's models break down under pressure, cyber agents are advancing quickly, and China is training AI for electronic warfare.
The economics of AGI: humans will become verifiers as machines take over labor
Researchers from MIT, WashU, and UCLA presented a model of the future economy in which AI performs most labor and the human role is reduced to verifying results. What does that mean for the labor market?
Nuclear LLMs, Chinese benchmarks, and the politics of measurement: highlights from Import AI 446
The new edition of the influential Import AI digest raises three key topics: the use of nuclear energy for data centers, an ambitious Chinese benchmark for AI, and a simple but effective approach to regulation through me
Import AI: 2 GW Data Centers and Control Over Superintelligence
The new issue of Import AI examines the construction of 2 GW data centers, regulatory risks, and ways to counter a hypothetical superintelligence.
Forecasts for the emergence of superintelligence and the saving “human factor” in the AI era
We break down the latest issue of Import AI: when to expect superintelligence, how neural networks solve the hardest mathematical problems, and why people are still willing to pay extra for real human interaction.
End of chalk era: AI starts proving theorems for humans
For a long time, we lived with a comforting belief that mathematics was the last bastion of pure human intelligence.