Submissions from lesswrong.com

		Model sizes are currently constrained by availability of inference hardware (lesswrong.com)
		1 point by ipnon 45 days ago \| past
		The Rise of Parasitic AI (lesswrong.com)
		3 points by dustingetz 45 days ago \| past
		Brain Efficiency: More Than You Wanted to Know (2022) (lesswrong.com)
		4 points by fdeage 46 days ago \| past
		Show HN: Is Friendly AI an Attractor? Self-Reports from 22 Models Say No (lesswrong.com)
		3 points by jsnider3 49 days ago \| past \| 1 comment
		Four ways learning Econ makes people dumber re: future AI (lesswrong.com)
		3 points by gwintrob 50 days ago \| past
		Claude 4.5 Opus’ Soul Document (lesswrong.com)
		342 points by the-needful 51 days ago \| past \| 244 comments
		Claude 4.5 Opus' Soul Document (lesswrong.com)
		38 points by thm 51 days ago \| past \| 13 comments
		Claude 4.5 Opus Soul Doc (lesswrong.com)
		2 points by pcald 52 days ago \| past
		DeepMind gives up on mechanistic interpretability research (lesswrong.com)
		3 points by cubefox 52 days ago \| past
		Claude Opus Soul Spec (lesswrong.com)
		4 points by panarky 52 days ago \| past
		A Year of Spaced Repetition Software in the Classroom (2015) (lesswrong.com)
		2 points by keiferski 52 days ago \| past
		Claude 4.5 Opus' Soul Document (lesswrong.com)
		2 points by karma_daemon 53 days ago \| past
		Claude 4.5 Opus' Soul Document (lesswrong.com)
		16 points by mcculley 54 days ago \| past \| 2 comments
		Model Weight Preservation is not enough (lesswrong.com)
		1 point by gmays 54 days ago \| past
		Anthropic's Claude 'Soul Document' extracted from Opus 4.5 weights (lesswrong.com)
		11 points by Richard_Weiss 54 days ago \| past \| 1 comment
		I Lost 100 Pounds Using TDT (lesswrong.com)
		1 point by gmays 55 days ago \| past
		AI Teddy Bears: A Brief Investigation (lesswrong.com)
		3 points by nicetomeetyu 56 days ago \| past \| 1 comment
		HybridTree: The First AGI Engine That Thinks in Korean (lesswrong.com)
		1 point by Hanjuwoncarrier 57 days ago \| past
		Urgent everyone – help us kill AI preemption (again) before this Friday (lesswrong.com)
		3 points by PhilosophyForAI 62 days ago \| past
		The Bughouse Effect (lesswrong.com)
		1 point by tbt 64 days ago \| past
		Multiagent Models of Mind (lesswrong.com)
		2 points by eatitraw 65 days ago \| past \| 2 comments
		Colds Spread (lesswrong.com)
		3 points by paulpauper 65 days ago \| past
		Murphyjitsu (2018) (lesswrong.com)
		12 points by surprisetalk 65 days ago \| past \| 4 comments
		How Colds Spread (lesswrong.com)
		2 points by comp_throw7 65 days ago \| past \| 1 comment
		AI 2025 – Last Shipmas (lesswrong.com)
		3 points by surprisetalk 66 days ago \| past
		Why does ChatGPT think mammoths were alive in December? (lesswrong.com)
		3 points by paulpauper 66 days ago \| past \| 3 comments
		Toddler Shoggoth Has Plenty of Raw Material (The Memetic Cocoon Threat Model) (lesswrong.com)
		2 points by kp1197 66 days ago \| past \| 1 comment
		Quotes from Moral Mazes (2019) (lesswrong.com)
		1 point by jger15 67 days ago \| past
		I ate bear fat with honey and salt flakes, to prove a point (lesswrong.com)
		7 points by B1FF_PSUVM 69 days ago \| past
		The Tale of the Top-Tier Intellect (lesswrong.com)
		1 point by cl3misch 70 days ago \| past
		More