John Bohannon -- see you @ICML!

This is the next ML paper I've seen from Apple after the spicy "Illusion of Reasoning" paper. I find that company's public research output genuinely puzzling. Good work but randomly distributed. arxiv.org/abs/2507.11851

During #ICML2025 last week, we teamed up with @askalphaxiv to host a happy hour + panel on: "The Science Singularity - How Automated ML Research is Changing the ML Landscape" Thanks to @pzakin, @FazlBarez, @tdietterich, and @bohannon_bot for the insights - and everyone who came…

model distillation

My main worry about the big AI safety recommendation is that research keeps accumulating that CoT might **not** be necessary for reasoning after all. If that turns out to be true, then this safety avenue will close. arxiv.org/abs/2507.11473

Yesterday’s panel that I ran at ICML on “The science singularity.” The room was packed so people sat down so others could see. Computer scientists are lovely people. Big thanks to @askalphaxiv @tdietterich @pzakin and @FazlBarez!

I want to announce I am also buying windsurf

Sigh… arxiv.org/abs/2501.06623

Computer Scientist @YejinChoinka uses natural language processing to develop AI systems that can understand language and make inferences about the world. Learn more about the 2022 MacArthur Fellow #MacFellow macfound.org/fellows/class-…

This is a super sophisticated way to probe LLM knowledge: arxiv.org/abs/2507.06952 @keyonV the probe tests alignment to a known world model... but what if the model learned a valid alternative? Any way to reveal the model's implicit world model without assuming one in advance?

Anyone want to jailbreak their Apple Watch...? @naehrdine what is the funnest thing now possible with WatchWitch on an Apple Watch? Anything truly verrückt? arxiv.org/abs/2507.07210