The Information

Art by Clark Miller

Interpretability Is the Key to Physical AI, Says World Mechanics CEO Sonia Joseph

By Rocket Drew · Oct 19, 2026 5:39pm PDT

Two of the hardest problems in AI right now are creating models that understand the physical world—so called world models—and making sense of the internal workings of AI models—otherwise known as mechanistic interpretability. Typically, these are treated as two separate problems: researchers working on world models are often... Two of the hardest problems in AI right now are creating models that understand the physical...

Researchers Need to Learn How AI Makes Decisions, Says Goodfire’s Tom McGrath

By Rocket Drew · Oct 12, 2026 5:43pm PDT · 2 comments

To understand the decision-making of AI, developers often look at the text that models produce as they reason through problems. But these “chains of thought” are increasingly hard to read, which is prompting AI developers to consider other approaches. One alternative, known as “mechanistic interpretability,” seeks to explain the inner workings... To understand the decision-making of AI, developers often look at the text that models produce as...

Aligning AI With Human Goals Might Be Impossible, Says AI Prof. Stuart Russell

By Rocket Drew · Oct 5, 2026 2:53pm PDT · 1 comment

Last week, OpenAI scrapped the model it had slated to release as GPT-6.1 Astra due to test results showing that it was deceptive and otherwise misaligned. “I think it’s about time,” said Stuart Russell, the computer science professor at the University of California, Berkeley, who co-authored the leading textbook on AI. Last week, OpenAI scrapped the model it had slated to release as GPT-6.1 Astra due to test...

Recursive Self-Improvement Won’t Be Winner-Takes-All, Say Founders of Ricursive

By Rocket Drew · Sep 28, 2026 3:30pm PDT

Recursive Self-Improvement Won’t Be Winner-Takes-All, Say Founders of Ricursive

By Rocket Drew · Sep 28, 2026 3:30pm PDT

Recursive self-improvement, the idea that AI models could get so good at research and development that they train themselves, without a need for human researchers, has captivated Silicon Valley. Some in the industry have suggested that the first company to achieve RSI will gain all the benefits: after all, once one company manages to cross that... Recursive self-improvement, the idea that AI models could get so good at research and development...

World Models Need More Intelligence, Says Luma CEO Amit Jain

By Rocket Drew · Sep 21, 2026 3:46pm PDT · 1 comment

The AI boom has been built largely on progress in language models, but a different approach has gained enormous attention in the industry this year: world models, which are designed to generate accurate predictions about the physical world, rather than words and numbers. I sat down with Amit Jain, co-founder and CEO of Luma AI, to discuss the... The AI boom has been built largely on progress in language models, but a different approach has...

OpenAI’s ‘Top Priority’ for AI Agents is Automating AI Research, Says Noam Brown

By Rocket Drew · Sep 14, 2026 1:08pm PDT

OpenAI’s latest AI model, GPT-6 Astra, notched improvements at a variety of professional tasks, from designing video games to transcribing sheet music. But OpenAI’s main goal when training new AI models is making them better at AI research and development, OpenAI research scientist Noam Brown told me. “The number one priority is recursive... OpenAI’s latest AI model, GPT-6 Astra, notched improvements at a variety of professional tasks,...