links 9/3/26: https://roamresearch.com/#/app/srcpublic/page/09-03-2026

• • https://alexanderwales.tumblr.com/post/826667577260605440/sometimes-i-talk-to-the-llms-and-get-a-sentence EEGs only detect electrical charge because pyramidal neurons all face the same way, so the charges add macroscopically
  • https://warontherocks.com/stop-using-ai-start-commanding-it/ not sure what to make of this. it's basically a pitch for "the military should use swarms of open-source AI agents". lots of jargon i'm not familiar with, agnostic on whether this is in fact a good idea.
  • https://www.lesswrong.com/posts/Bs8geGyWEitYvCzys/pauseai-has-officially-disendorsed-pauseai-us drama in the pause-AI world
  • https://milan.cvitkovic.net/china_notes_q3_2026/ Milan Cvitkovic joins the chorus of people saying Chinese biotech is legitimately impressive
  • https://asteriskmag.substack.com/p/we-already-know-what-a-world-withou Clara Collier says that if people didn't have to work, they'd spend their time on gossip and petty drama, drawing examples from history and ethnography. sounds plausible.
  • https://www.lesswrong.com/posts/cr5pyW7Mzm33p4AvN/ai-safety-acculturation-is-neglected it's a good question: how could Successful Professionals (in non-STEM office jobs) communicate the norms of their culture or how to make a good impression on them? I notice a distinct lack of such people chiming in to offer suggestions.