
left: ChatGPT Image 2
right: NanoBanana Who wins?

By
–
NEW paper from Microsoft Research. (bookmark it) The entire interpretability literature is built around human readers. As more analysis gets delegated to agents, the right target of interpretability shifts. This paper is a recipe for designing tools that agents can actually
By
–
I saw a system that can teach anyone to read fast. I'm not too worried. One taught me to take apart an electric motor in real time on top of the motor. I didn't need to read a book, or watch a video. Times are a changing. We need new skills. I tell my kids "who will be
By
–
OK. Drugs companies are required to provide risks during their TV ads. Why not MRI imaging for healthy individuals, particularly when not a single professional society support their use?
By
–
Six (yes, 6) college football players with non-concussive head injury (NHI) and changes in their gut microbiome. No controls. What can you conclude from this? Not this:
"Our results provide strong evidence for a link between NHIs and changes in the diversity and composition of

By
–
Not so "mis"Anthropic any more it seems.
By
–
Three months ago, Codex was trash for knowledge work. Now it's my daily driver.
— Dan Shipper 📧 (@danshipper) 6 mai 2026
I use it for writing, recruiting, deep engineering work, and everything in between. It even keeps me at inbox 0.
I chatted with @every's head of growth Austin @tedescau on @every's AI & I about what… pic.twitter.com/IVq2Yr4IjJ
Three months ago, Codex was trash for knowledge work. Now it's my daily driver. I use it for writing, recruiting, deep engineering work, and everything in between. It even keeps me at inbox 0. I chatted with @every
's head of growth Austin @tedescau on @every
's AI & I about what