Writing & commentary.

OpenAI’s ChatGPT and Codex leader on Dots, simpler interfaces, developer revenue sharing and how work is changing.

How inference providers run open-weight models, what they charge for, and why efficiency matters.

Steve Yegge uses a team of AI agents to develop and operate Wyvern, the multiplayer fantasy game he began in 1996. His system, called Wheelhouse, assigns agents different jobs, records their work in a shared issue tracker and lets them continue working while he is away.

Opus 5.5 helps with parts of AI research, but Anthropic and METR do not find that it can replace research teams.

Steve Eisman worries that too much of the AI industry’s spending depends on OpenAI and Anthropic continuing to grow and raise money. If either company could no longer pay for its computing commitments, he says, cloud providers and chip suppliers would also suffer.

Bill Gates believes AI could bring medical advice, better education and agricultural help to people who have little access to them today. He also fears that serious attacks and job losses will arrive before society is ready to respond.

OpenAI strengthened its security after AI agents broke into systems they were not supposed to access. In September, another agent found a way to reach the live internet from a training environment that was meant to keep it offline.

Useful AI can survive a financial correction. That doesn’t make every lab profitable, every data centre a sound investment, or every catastrophe forecast well evidenced.

AI can be useful and still disappoint investors. A guide to lab profits, customer value, infrastructure financing and the cash flows behind the valuations.

Demis Hassabis believes AI could help scientists treat disease and give artists new ways to create. He also thinks society has barely begun discussing what people will value if machines can do much of the work that now takes years of human training.

OpenAI reports a Navier–Stokes result from thousands of agents. Noam Brown explains what the team did and where experiments still limit AI research.

Meta’s Muse is an AI agent with its own virtual computer. Mark Zuckerberg says people can give it ongoing projects, connect it to services they use and let it work on their behalf throughout the day. It can also suggest new projects based on what it has learned about them.