Submissions from lesswrong.com

		Pacing the Frontier is not the actual goal for AI labs (lesswrong.com)
		82 points by brlewis 1 day ago \| rank \| 92 comments
		Watch AI materials-science and bioscience abilities closely (lesswrong.com)
		40 points by joozio 15 days ago \| rank \| 49 comments
		Astra and Fable still hack on simple variants of alignment evals from 2025 (lesswrong.com)
		479 points by Levitating 16 days ago \| rank \| 235 comments
		How My Students Think About AI (lesswrong.com)
		50 points by paulpauper 19 days ago \| rank \| 10 comments
		What is nueralese and why is it bad (lesswrong.com)
		77 points by tristanMatthias 26 days ago \| rank \| 6 comments
		How concerned should we be about Astra's recurrent architecture? (lesswrong.com)
		151 points by yurivish 26 days ago \| rank \| 128 comments
		The Library of Scott Alexandria (lesswrong.com)
		4 points by benatkin 27 days ago \| rank \| 1 comment
		LLMs are (still) mostly powered by imitative learning, not RL (lesswrong.com)
		3 points by wslh 39 days ago \| rank \| discuss
		Recursive Middle Manager Hell (lesswrong.com)
		5 points by rzk 42 days ago \| rank \| discuss
		Misaligned AIs could use killer robots to take over (lesswrong.com)
		7 points by x312 49 days ago \| rank \| 4 comments
		Don't Build Mindreading (lesswrong.com)
		21 points by paulpauper 50 days ago \| rank \| 14 comments
		What I did in the hedonium shockwave, by Emma, age six and a half (lesswrong.com)
		11 points by paulpauper 51 days ago \| rank \| 2 comments
		How Go Players Disempower Themselves to AI (lesswrong.com)
		3 points by zetalyrae 58 days ago \| rank \| 1 comment
		Is Mythos good at cyber because it kept hacking Anthropics sandboxes in training (lesswrong.com)
		5 points by lukaspetersson 62 days ago \| rank \| discuss
		Unless Its Governance Changes, Anthropic Is Untrustworthy (2025) (lesswrong.com)
		25 points by thepasch 63 days ago \| rank \| 1 comment
		Why I Left Google DeepMind (lesswrong.com)
		192 points by eatitraw 64 days ago \| rank \| 57 comments
		OpenAI's myopia keeps causing alignment problems (lesswrong.com)
		3 points by joozio 64 days ago \| rank \| discuss
		An OpenAI model left notes about how to evade containment; we need more details (lesswrong.com)
		17 points by joozio 65 days ago \| rank \| 10 comments