Good read on model safety, but doesn't feel easy when you put it side by side with Anthropic’s report on CoT faithfulness —not just because CoT monitorability is fragile, but also because efforts to make CoT more faithful didn’t really move the needle. And then there’s Coconut (continuous latent space reasoning), which doesn’t give human-readable CoT at all. Seems like some reductionist approaches—like the deeper behavioral analysis Goodfire does—are still essential
This space is where we’ll be sharing sharp, concise insights on the fast-evolving world of AI, robotics, and deep tech. From frontier research to emerging use cases, expect regular posts that explore where technology is heading and what it means for the systems we live in. Whether it’s breakthroughs in machine intelligence, the ethics of automation, or the shifting interface between human and machine, the Intelligence Feed exists to keep signal high and noise low. Thanks for joining us at the start. More to come soon.
