More Thoughts About AGI

Type	URL / Title	Status	Size	Words	Files
notion	https://www.notion.so/More-Thoughts-About-AGI-2504c916374a80... More Thoughts About AGI	fetched	148.9 KB	10,000 (truncated)	Raw / Text
tweet	http://pic.twitter.com/b5HXqxcAZv	error	-	-	- / -
tweet	https://x.com/i/status/1999994684203045003 Tweet by Séb Krier	fetched	330 B	46	Raw / Text
webpage	https://blog.cosmos-institute.org/p/faster-horses Faster Horses - by Harry Law - Cosmos Institute	fetched	224.6 KB	2,620	Raw / Text
webpage	https://www.patreon.com/posts/january-3-2025-147255956	error	-	-	- / -
webpage	https://docs.google.com/document/d/1Ab8Q8_F22drmQFedJD4IWPVs...	error	-	-	- / -
webpage	https://docs.google.com/document/d/1PH7bNtJdy55IPIv0i01JgC41...	error	-	-	- / -
webpage	https://news.ycombinator.com/item?id=46408921	error	-	-	- / -
tweet	https://x.com/snewmanpv/status/2005329341845303612 Tweet by Steve Newman	fetched	311 B	45	Raw / Text
webpage	https://www.lesswrong.com/posts/u6Lacc7wx4yYkBQ3r/insights-i... Insights into Claude Opus 4.5 from Pokémon — LessWrong	fetched	952.6 KB	5,559	Raw / Text
webpage	https://t.co/YH7W3F9FtQ Cognition all the way down 2.0: neuroscience beyond neurons in the diverse intelligence era \| Synthese	fetched	727.3 KB	10,000 (truncated)	Raw / Text
webpage	https://t.co/UKIULAD5wU JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/anilkseth/status/1987894412584566876?t=0JSx8bv... Tweet by Anil Seth	fetched	316 B	39	Raw / Text
tweet	https://x.com/i/status/2000413758729162889 Tweet by Ethan Mollick	fetched	302 B	49	Raw / Text
tweet	https://x.com/i/status/1998778005129224412 Tweet by Ethan Mollick	fetched	304 B	43	Raw / Text
webpage	https://jasmi.news/p/42-notes-on-ai-and-work 42 notes on AI & work - by Jasmine Sun	fetched	205.0 KB	1,572	Raw / Text
webpage	https://www.understandingai.org/p/i-got-fooled-by-ai-for-sci... I got fooled by AI-for-science hype—here's what it taught me	fetched	301.3 KB	3,790	Raw / Text
webpage	https://www.lesswrong.com/posts/4mvphwx5pdsZLMmpY/recent-ai-... Recent AI model progress feels mostly like bullshit — LessWrong	fetched	1.6 MB	9,580	Raw / Text
webpage	https://ringohosp.substack.com/p/what-artificial-intelligenc... What artificial intelligence can do for a small independent hotel	fetched	292.1 KB	5,484	Raw / Text
webpage	https://gwern.net/ai-daydreaming LLM Daydreaming · Gwern.net	fetched	70.3 KB	2,652	Raw / Text
tweet	https://x.com/1a3orn/status/1997056050403725373 Tweet by 1a3orn	fetched	325 B	48	Raw / Text
webpage	https://www.dwarkesh.com/p/thoughts-on-ai-progress-dec-2025 Thoughts on AI progress (Dec 2025) - by Dwarkesh Patel	fetched	216.5 KB	3,159	Raw / Text
tweet	https://x.com/Jack_W_Lindsey/status/1993389056932339721 Tweet by Jack Lindsey	fetched	301 B	44	Raw / Text
webpage	https://www.strangeloopcanon.com/p/epicycles-all-the-way-dow... Epicycles All The Way Down - by Rohit Krishnan	fetched	277.6 KB	4,675	Raw / Text
tweet	https://x.com/mikeknoop/status/1993402375944679673?t=LWizqAJ... Tweet by Mike Knoop	fetched	159 B	21	Raw / Text
tweet	https://x.com/QuintinPope5/status/1991472282422899184?t=DqRl... Tweet by Quintin Pope	fetched	152 B	19	Raw / Text
tweet	https://x.com/MLStreetTalk/status/1989961238839693788?t=Fu6V... Tweet by Machine Learning Street Talk	fetched	342 B	42	Raw / Text
tweet	https://x.com/fchollet/status/1989340153114976598 Tweet by François Chollet	fetched	311 B	42	Raw / Text
tweet	https://x.com/HjalmarWijk/status/1985529956890530217 Tweet by Hjalmar Wijk	fetched	272 B	39	Raw / Text
webpage	https://docs.google.com/document/d/1p7l-KM31hETcNwXMjSkTik0_...	error	-	-	- / -
webpage	https://youtu.be/mYDSSRS-B5U?si=EvNetk1JVc7e6HXt&t=823 - YouTube	fetched	1.1 MB	26	Raw / Text
webpage	https://arxiv.org/abs/2502.05167 [2502.05167] NoLiMa: Long-Context Evaluation Beyond Literal Matching	fetched	48.1 KB	851	Raw / Text
tweet	https://x.com/emollick/status/1985507610963968447 Tweet by Ethan Mollick	fetched	329 B	48	Raw / Text
webpage	https://80000hours.org/podcast/episodes/holden-karnofsky-con... Holden Karnofsky on dozens of amazing opportunities to make AI safer — and all his AGI takes \| 80,000 Hours	fetched	532.4 KB	10,000 (truncated)	Raw / Text
tweet	https://x.com/karpathy/status/1973435013875314729?t=l8ZRYSoh... Tweet by Andrej Karpathy	fetched	332 B	47	Raw / Text
tweet	https://x.com/dwarkesh_sp/status/1979259041013731752?t=_ICAk... Tweet by Dwarkesh Patel	fetched	362 B	48	Raw / Text
tweet	https://x.com/karpathy/status/1979644538185752935?t=H0LxJvX5... Tweet by Andrej Karpathy	fetched	330 B	51	Raw / Text
tweet	https://x.com/ryanpgreenblatt/status/1981012208332280219?s=4... Tweet by Ryan Greenblatt	fetched	50 B	1	Raw / Text
webpage	https://pca.st/episode/d9d8178e-9f7e-4102-bf2c-aa3e747bc188 Autonomous Organizations: Vending Bench & Beyond, w/ Lukas Petersson & Axel Backlund of Andon Labs - Pocket Casts	fetched	266.3 KB	557	Raw / Text
webpage	https://www.dwarkesh.com/p/andrej-karpathy Andrej Karpathy — AGI is still a decade away	fetched	661.4 KB	10,000 (truncated)	Raw / Text
webpage	https://www.henrikkarlsson.xyz/p/wordless-thought When is better to think without words? - by Henrik Karlsson	fetched	222.4 KB	3,737	Raw / Text
webpage	https://ai-frontiers.org/articles/agis-last-bottlenecks AGI's Last Bottlenecks \| AI Frontiers	fetched	169.2 KB	5,715	Raw / Text
tweet	https://x.com/emollick/status/1980704687377486182?t=85fVurBC... Tweet by Ethan Mollick	fetched	306 B	45	Raw / Text
webpage	https://t.co/9rAySHO5L2 JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/emollick/status/1980601776710525103?t=tLJFmxj8... Tweet by Ethan Mollick	fetched	358 B	48	Raw / Text
webpage	https://t.co/ZwhDegc4NO JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/tobyordoxford/status/1980351353227768109?t=Kev... Tweet by Toby Ord	fetched	329 B	47	Raw / Text
webpage	https://t.co/X2EocaaZjQ Inference Scaling Reshapes AI Governance — Toby Ord	fetched	120.3 KB	6,006	Raw / Text
tweet	https://x.com/tobyordoxford/status/1974200193504719049?t=fKg... Tweet by Toby Ord	fetched	286 B	44	Raw / Text
webpage	https://thezvi.substack.com/p/bending-the-curve Bending The Curve - by Zvi Mowshowitz	fetched	330.0 KB	6,311	Raw / Text
webpage	https://www.dwarkesh.com/p/thoughts-on-sutton Some thoughts on the Sutton interview - by Dwarkesh Patel	fetched	325.8 KB	1,938	Raw / Text
webpage	https://www.interconnects.ai/p/thoughts-on-the-curve Thoughts on The Curve - by Nathan Lambert - Interconnects	fetched	241.8 KB	2,519	Raw / Text
webpage	https://t.co/qyMUb0c2Oh JavaScript is not available.	fetched	225.2 KB	29	Raw / Text
tweet	https://x.com/teortaxesTex/status/1971854577755377748?t=V7JV... Tweet by Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)	fetched	400 B	43	Raw / Text
webpage	https://importai.substack.com/p/import-ai-430-emergence-in-v... Import AI 430: Emergence in video models; Unitree backdoor; preventative strikes to take down AGI projects	fetched	190.9 KB	2,352	Raw / Text
tweet	https://x.com/snewmanpv/status/1970528715403854327?t=Mj_RiH2... Tweet by Steve Newman	fetched	306 B	50	Raw / Text
tweet	https://x.com/snewmanpv/status/1970528715403854327 Tweet by Steve Newman	fetched	306 B	50	Raw / Text
webpage	https://epochai.substack.com/p/the-huge-potential-implicatio... The huge potential implications of long-context inference	fetched	228.2 KB	2,250	Raw / Text
webpage	https://pca.st/episode/27144be7-bcae-4e16-b1e4-40302c956a1a The Startup Powering The Data Behind AGI - Pocket Casts	fetched	263.2 KB	395	Raw / Text
tweet	https://x.com/emollick/status/1968339804975948274?t=jgv5dYIj... Tweet by Ethan Mollick	fetched	309 B	44	Raw / Text
tweet	https://x.com/mjbukow/status/1962186888103747892 Tweet by Martin	fetched	319 B	50	Raw / Text
tweet	https://x.com/livgorton/status/1960378468807447026?t=NKCW413... Tweet by Liv	fetched	256 B	32	Raw / Text
webpage	https://www.linkedin.com/posts/tcpeixoto_help-needed-human-i... Is "human-in-the-loop" a luxury for the privileged? \| Tiago C. Peixoto posted on the topic \| LinkedIn	fetched	377.0 KB	3,426	Raw / Text
webpage	https://www.lesswrong.com/posts/2ssPfDpdrjaM2rMbn/my-agi-tim... My AGI timeline updates from GPT-5 (and 2025 so far) — LessWrong	fetched	814.3 KB	3,725	Raw / Text
webpage	https://importai.substack.com/p/import-ai-425-iphone-video-g... Import AI 425: iPhone video generation; subtle misalignment; making open weight models safe through surgical deletion	fetched	200.4 KB	3,622	Raw / Text
webpage	https://www.interconnects.ai/p/contra-dwarkesh-on-continual-... Contra Dwarkesh on Continual Learning - by Nathan Lambert	fetched	215.1 KB	2,167	Raw / Text
webpage	https://www.reddit.com/r/mlscaling/comments/1mrm0di/we_had_t...	error	-	-	- / -
tweet	https://x.com/Simeon_Cps/status/1956519485277684219?t=Y9sBaf... Tweet by Siméon	fetched	326 B	55	Raw / Text
webpage	https://peterwildeford.substack.com/p/gpt-5-a-small-step-for... GPT-5: a small step for intelligence, a giant leap for normal people	fetched	241.1 KB	3,341	Raw / Text
webpage	https://ethanding.substack.com/p/ai-subscriptions-get-short-... tokens are getting more expensive - by Ethan Ding	fetched	241.6 KB	2,352	Raw / Text
webpage	https://thezvi.substack.com/i/169772920/city-in-a-bottle AI #128: Four Hours Until Probably Not The Apocalypse	fetched	678.3 KB	10,000 (truncated)	Raw / Text
tweet	https://x.com/ben_j_todd/status/1934284189928501482 Tweet by Benjamin Todd	fetched	327 B	52	Raw / Text
webpage	/2474c916374a80329e09f667fed4ddce#24c4c916374a8079bc56e14848...	error	-	-	- / -
webpage	http://henrikkarlsson.xyz/p/attention Almost anything you give sustained attention to will begin to loop on itself and bloom	fetched	235.7 KB	3,715	Raw / Text
webpage	https://substack.com/redirect/ce40c09e-8b1b-44b4-905f-fc7870... Using spaced repetition systems to see through a piece of mathematics	fetched	44.2 KB	6,450	Raw / Text
webpage	https://substack.com/redirect/a43876f1-09fb-4d47-94b4-1bc350... JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/MLStreetTalk/status/1963492648708132948 Tweet by Machine Learning Street Talk	fetched	348 B	44	Raw / Text
webpage	https://www.interconnects.ai/p/brakes-on-an-intelligence-exp... State of play of AI progress (and related brakes on an intelligence explosion)	fetched	329.7 KB	5,129	Raw / Text
webpage	https://podcasts.apple.com/us/podcast/the-mathematical-found... The Mathematical Foundations o… - Machine Learning Street Talk (MLST) - Apple Podcasts	fetched	146.4 KB	1,187	Raw / Text
notion	https://www.notion.so/1df4c916374a8075bea1c1cec4d1e3bc Will LLMs Generalize?	fetched	146 B	26	Raw / Text
webpage	https://www.lesswrong.com/posts/vvgND6aLjuDR6QzDF/my-model-o... My model of what is going on with LLMs — LessWrong	fetched	1.5 MB	9,127	Raw / Text
webpage	https://www.lesswrong.com/posts/GADJFwHzNZKg2Ndti/have-llms-... Have LLMs Generated Novel Insights? — LessWrong	fetched	1.3 MB	5,125	Raw / Text
webpage	https://helentoner.substack.com/p/2-big-questions-for-ai-pro... 2 big questions for AI progress in 2025-2026	fetched	229.3 KB	4,101	Raw / Text
notion	https://www.notion.so/1a74c916374a80829f2cd66bd4e63b3d How Deep are The Remaining Rabbit Holes?	fetched	6.5 KB	955	Raw / Text
webpage	https://lemmata.substack.com/p/what-i-wish-i-knew-about-fron... What I Wish I Knew About FrontierMath - by Greg Burnham	fetched	221.6 KB	3,741	Raw / Text
webpage	https://amistrongeryet.substack.com/p/were-finding-out-what-... Comments - We're Finding Out What Humans are Bad At	fetched	217.6 KB	1,303	Raw / Text
tweet	https://x.com/emollick/status/1892108434171949321?t=JRIExSHe... Tweet by Ethan Mollick	fetched	302 B	46	Raw / Text
webpage	https://erictopol.substack.com/p/when-doctors-with-ai-are-ou...	error	-	-	- / -
webpage	https://erictopol.substack.com/p/when-doctors-with-ai-are-ou... When Doctors With A.I. Are Outperformed by A.I. Alone	fetched	208.7 KB	1,751	Raw / Text
webpage	https://t.co/b3ZuMjO0ZJ From OpenAI to DeepSeek, companies say AI can “reason” now. Is it true? \| Vox	fetched	394.7 KB	3,734	Raw / Text
tweet	https://x.com/SigalSamuel/status/1892997861886820474?t=x29Sr... Tweet by Sigal Samuel	fetched	301 B	42	Raw / Text
webpage	https://arxiv.org/abs/2410.06992 [2410.06992] SWE-Bench+: Enhanced Coding Benchmark for LLMs	fetched	46.1 KB	824	Raw / Text
webpage	https://epochai.substack.com/p/ai-progress-is-about-to-speed... AI progress is about to speed up - by Ege Erdil - Epoch AI	fetched	191.0 KB	2,908	Raw / Text
notion	https://aidanmclaughlin.notion.site/reasoners-problem	error	-	-	- / -
tweet	https://x.com/AndrewCritchPhD/status/1891887600102932629?t=0... Tweet by Andrew Critch (🤖🩺🚀)	fetched	337 B	46	Raw / Text
webpage	https://thezvi.substack.com/p/ai-104-american-state-capacity... AI #104: American State Capacity on the Brink	fetched	517.7 KB	10,000 (truncated)	Raw / Text
tweet	https://x.com/buccocapital/status/1890745551995424987 Tweet by BuccoCapital Bloke	fetched	312 B	51	Raw / Text
webpage	https://news.ycombinator.com/item?id=43169586	error	-	-	- / -
tweet	https://x.com/littmath/status/1898461323391815820 Tweet by Daniel Litt	fetched	329 B	45	Raw / Text
webpage	https://www.theintrinsicperspective.com/p/ai-plays-pokemon-b... AI plays Pokémon, but so does π. Tesla's "brand collapse." Neuroscience scaling laws. Why do I know all the songs on the radio?	fetched	196.6 KB	1,078	Raw / Text
tweet	https://x.com/slatestarcodex/status/1896457193215742274 Tweet by Scott Alexander	fetched	309 B	41	Raw / Text
webpage	https://t.co/XrZUlbKkek JavaScript is not available.	fetched	225.2 KB	29	Raw / Text
tweet	https://x.com/emollick/status/1898802108926661012?t=5xHM6k-V... Tweet by Ethan Mollick	fetched	332 B	47	Raw / Text
notion	https://www.notion.so/1af4c916374a805fb6cbe37f43843bfb FrontierMath, creativity, benchmarks, and how big AI gaps are	fetched	4.5 KB	712	Raw / Text
webpage	https://lemmata.substack.com/p/alphaproof-and-the-imo AlphaProof and the IMO - by Greg Burnham - Lemmata	fetched	340.9 KB	6,604	Raw / Text
webpage	https://lemmata.substack.com/p/o3-o4-mini-largely-incrementa... o3, o4-mini: Largely Incremental at Math - by Greg Burnham	fetched	234.8 KB	1,939	Raw / Text
tweet	https://x.com/snewmanpv/status/1929724463916241351 Tweet by Steve Newman	fetched	331 B	50	Raw / Text
notion	https://www.notion.so/1d94c916374a800fa414ddaba0145aee Scaling Laws and Fractal Complexity	fetched	434 B	51	Raw / Text
webpage	https://t.co/uAHid8PNFd JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/vitrupo/status/1910544354709762111?t=q9UCeSNz6... Tweet by vitrupo	fetched	324 B	43	Raw / Text
notion	https://www.notion.so/1024c916374a80b7a2d2ddd6ddbb5987 AI Creativity and Problem Solving	fetched	5.1 KB	754	Raw / Text
webpage	https://chatgpt.com/c/6705d33a-9a84-800b-8ddd-f52c7dfd7db2	error	-	-	- / -
webpage	https://thegradientpub.substack.com/i/148749478/research-hig... Update #83: AI Music Fraud and PlanSearch - by Cole Frank	fetched	333.2 KB	2,617	Raw / Text
webpage	https://thegradientpub.substack.com/p/update-83-ai-music-fra... Update #83: AI Music Fraud and PlanSearch - by Cole Frank	fetched	333.2 KB	2,617	Raw / Text
tweet	https://x.com/Simeon_Cps/status/1868411560768811147?t=bZkhvW... Tweet by Siméon	fetched	216 B	25	Raw / Text
tweet	https://twitter.com/MFarajtabar/status/1844456893605191804 Tweet by Mehrdad Farajtabar	fetched	325 B	46	Raw / Text
tweet	https://x.com/boazbaraktcs/status/1844763538260209818?t=4i54...	error	-	-	- / -
tweet	https://x.com/natolambert/status/1846195544458514561?t=-8t7_... Tweet by Nathan Lambert	fetched	180 B	25	Raw / Text
notion	https://www.notion.so/10b4c916374a80b3a424fd2fdd180460 Offcuts	fetched	2.4 KB	388	Raw / Text
webpage	https://amistrongeryet.substack.com/i/114809888/gpt-has-no-p... I’m a Senior Software Engineer. What Will It Take For An AI To Do My Job?	fetched	297.8 KB	7,080	Raw / Text
webpage	https://chatgpt.com/share/6701b6c0-8b50-800b-a40f-12427a2ba2... ChatGPT - String Difference Algorithm JavaScript	fetched	278.6 KB	34	Raw / Text
webpage	https://chatgpt.com/share/6701b757-2bac-800b-a66a-eff0914a90... ChatGPT - String Comparison and Reconstruction	fetched	314.8 KB	34	Raw / Text
webpage	https://claude.ai/chat/4d6ea2d6-598c-4a59-8e0f-fe0593009af4	error	-	-	- / -
tweet	https://x.com/colin_fraser/status/1833953195430719744 Tweet by Find me on bsky @colin-fraser.net	fetched	196 B	23	Raw / Text
webpage	https://t.co/CwTAMbWnUO [2409.04109] Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers	fetched	46.2 KB	816	Raw / Text
notion	https://www.notion.so/12a4c916374a80a8bb3fee3ef70089ed AIs Learn Both Shallow And Deep Things	fetched	118 B	2	Raw / Text
tweet	https://twitter.com/TrentonBricken/status/184665294330043603... Tweet by Trenton Bricken	fetched	251 B	30	Raw / Text
tweet	https://twitter.com/emollick/status/1853857842312188012 Tweet by Ethan Mollick	fetched	289 B	41	Raw / Text
notion	https://www.notion.so/10b4c916374a80c1b46bf09723f5c655 Will Many Tasks Be “AGI Complete”?	fetched	3.1 KB	495	Raw / Text
webpage	https://thezvi.substack.com/i/148809120/intelligent-design AI #82: The Governor Ponders - by Zvi Mowshowitz	fetched	338.2 KB	8,118	Raw / Text
tweet	https://x.com/IntuitMachine/status/1843238594669932686?t=VsI... Tweet by Carlos E. Perez	fetched	331 B	41	Raw / Text
webpage	https://thezvi.substack.com/p/ai-85-ai-wins-the-nobel-prize	error	-	-	- / -
tweet	https://x.com/tszzl/status/1844293154549399866 Tweet by roon	fetched	238 B	33	Raw / Text
tweet	https://x.com/lateinteraction/status/1892724949074559412?t=z... Tweet by Omar Khattab	fetched	328 B	49	Raw / Text
notion	https://www.notion.so/1724c916374a8008b1e8f89d6b9d4295 Inscrutability of AI	fetched	1.4 KB	213	Raw / Text
tweet	https://x.com/_xjdr/status/1875295492906828005?t=m-uTZp61RAt... Tweet by xjdr	fetched	293 B	42	Raw / Text
notion	https://www.notion.so/b07c6f4278c94f7abe9acf429aad7477 Timelines	fetched	43.5 KB	6,467	Raw / Text
webpage	https://arxiv.org/pdf/2311.02462	fetched	279.2 KB	10,000 (truncated)	Raw / Text
webpage	https://semianalysis.com/2024/12/11/scaling-laws-o1-pro-arch... Scaling Laws - O1 Pro Architecture, Reasoning Training Infrastructure, Orion and Claude 3.5 Opus “Failures”	fetched	374.2 KB	7,098	Raw / Text
tweet	https://twitter.com/Miles_Brundage/status/186207007958178224... Tweet by Miles Brundage	fetched	313 B	48	Raw / Text
tweet	https://twitter.com/DKokotajlo67142/status/18620958277007934... Tweet by Daniel Kokotajlo	fetched	174 B	28	Raw / Text
tweet	https://twitter.com/deanwball/status/1861975137425244200 Tweet by Dean W. Ball	fetched	299 B	43	Raw / Text
tweet	https://twitter.com/random_walker/status/1866858354942832966 Tweet by Arvind Narayanan	fetched	309 B	44	Raw / Text
webpage	https://www.lesswrong.com/posts/HiTjDZyWdLEGCDzqu?s=03#comme... Implications of the inference scaling paradigm for AI safety — LessWrong	fetched	1.6 MB	9,598	Raw / Text
webpage	https://news.ycombinator.com/item?id=42565606	error	-	-	- / -
webpage	https://www.aisnakeoil.com/p/is-ai-progress-slowing-down Is AI progress slowing down?	fetched	261.7 KB	4,469	Raw / Text
webpage	https://www.aisnakeoil.com/p/is-ai-progress-slowing-down#foo... Is AI progress slowing down?	fetched	261.7 KB	4,469	Raw / Text
webpage	https://www.saysmaybe.com/latest-work/selective-perspectives... Selective Perspectives	fetched	45.0 KB	998	Raw / Text
webpage	https://arxiv.org/abs/2407.21783 [2407.21783] The Llama 3 Herd of Models	fetched	135.7 KB	2,464	Raw / Text
webpage	https://arxiv.org/abs/2402.00838 [2402.00838] OLMo: Accelerating the Science of Language Models	fetched	50.4 KB	852	Raw / Text
tweet	https://www.vox.com/future-perfect/353933/openai-open-letter...	error	-	-	- / -
tweet	https://x.com/snewmanpv/status/1870496367233438200 Tweet by Steve Newman	fetched	307 B	49	Raw / Text
tweet	https://twitter.com/S_OhEigeartaigh/status/18686064825076163... Tweet by Seán Ó hÉigeartaigh	fetched	314 B	47	Raw / Text
webpage	https://thezvi.substack.com/p/ai-94-not-now-google AI #94: Not Now, Google - by Zvi Mowshowitz	fetched	650.9 KB	10,000 (truncated)	Raw / Text
webpage	https://www.hyperdimensional.co/p/thresholds Thresholds - by Dean W. Ball - Hyperdimensional	fetched	222.6 KB	3,552	Raw / Text
webpage	https://substack.com/redirect/b54107cc-741f-41fb-9dea-8b7c09... An Extraordinary Alien - by Dean W. Ball - Hyperdimensional	fetched	235.4 KB	3,178	Raw / Text
webpage	https://t.co/7gG705nmNL JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/tsarnick/status/1861921602235150545?t=bJerP2YQ... Tweet by Tsarathustra	fetched	205 B	29	Raw / Text
tweet	https://x.com/RichardMCNgo/status/1861450205109096847 Tweet by Richard Ngo	fetched	335 B	49	Raw / Text
tweet	https://x.com/DKokotajlo67142/status/1862095827700793499 Tweet by Daniel Kokotajlo	fetched	174 B	28	Raw / Text
webpage	https://www.theintrinsicperspective.com/p/great-scientists-f... Great scientists follow intuition and beauty, not rationality	fetched	204.5 KB	1,338	Raw / Text
webpage	https://thezvi.substack.com/i/151975026/prepare-for-takeoff	error	-	-	- / -
webpage	https://thezvi.substack.com/p/ai-91-deep-thinking	error	-	-	- / -
webpage	https://www.newcomer.co/p/our-biggest-takeaways-from-cerebra... Our Biggest Takeaways from Cerebral Valley: No Wall on AI Progress Despite the Scaling Law Debate	fetched	199.1 KB	730	Raw / Text
webpage	https://www.theintrinsicperspective.com/p/nonpolitical-conte...	error	-	-	- / -
tweet	https://x.com/deanwball/status/1860114571111084250?t=mHKNdAI... Tweet by Dean W. Ball	fetched	209 B	27	Raw / Text
webpage	https://t.co/VzldSvcMCU JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/eli_lifland/status/1860087262849171797?t=GRtV2... Tweet by Eli Lifland	fetched	352 B	46	Raw / Text
webpage	https://www.understandingai.org/p/are-transformer-based-foun... Are transformer-based foundation models running out of steam?	fetched	138.3 KB	723	Raw / Text
webpage	https://thezvi.substack.com/p/ai-86-just-think-of-the-potent...	error	-	-	- / -
tweet	https://x.com/polynoamial/status/1855037689533178289?t=fO542... Tweet by Noam Brown	fetched	214 B	30	Raw / Text
webpage	https://thezvi.substack.com/i/151331494/tear-down-this-wall	error	-	-	- / -
webpage	https://t.co/lcy9sVjC4D JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/nrehiew_/status/1861464774904537215?t=6xU6YvYS... Tweet by wh	fetched	213 B	25	Raw / Text
webpage	https://marginalrevolution.com/marginalrevolution/2024/11/ho... How badly do humans misjudge AIs? - Marginal REVOLUTION	fetched	65.0 KB	364	Raw / Text
webpage	https://sites.google.com/view/raphaelraux/research?authuser=... raphaelraux - Research	fetched	120.6 KB	494	Raw / Text
webpage	https://thezvi.substack.com/i/151650076/the-last-mile	error	-	-	- / -
tweet	https://x.com/BethMayBarnes/status/1860065450824204686?t=HZ5... Tweet by Elizabeth Barnes	fetched	304 B	43	Raw / Text
webpage	https://www.interconnects.ai/p/scaling-realities Scaling realities - by Nathan Lambert - Interconnects	fetched	188.2 KB	959	Raw / Text
tweet	https://x.com/emollick/status/1856950856395984940?t=dvS4aVH3... Tweet by Ethan Mollick	fetched	329 B	55	Raw / Text
tweet	https://x.com/Miles_Brundage/status/1856809970089721991?t=kZ... Tweet by Miles Brundage	fetched	200 B	29	Raw / Text
tweet	https://x.com/swyx/status/1856776660986859632?t=OCywwz3dX18k... Tweet by swyx	fetched	294 B	53	Raw / Text
webpage	https://www.theinformation.com/articles/following-openai-goo...	error	-	-	- / -
webpage	https://www.theintrinsicperspective.com/p/ai-progress-has-pl... AI progress has plateaued below GPT-5 level - by Erik Hoel	fetched	261.3 KB	2,717	Raw / Text
webpage	https://url3396.theinformation.com/ls/click?upn=u001.71kYkaW...	error	-	-	- / -
webpage	https://url3396.theinformation.com/ls/click?upn=u001.71kYkaW...	error	-	-	- / -
webpage	https://url3396.theinformation.com/ls/click?upn=u001.71kYkaW...	error	-	-	- / -
webpage	https://url3396.theinformation.com/ls/click?upn=u001.71kYkaW...	error	-	-	- / -
webpage	https://url3396.theinformation.com/ls/click?upn=u001.71kYkaW...	error	-	-	- / -
webpage	https://podcasts.apple.com/us/podcast/its-not-about-scale-it... It's Not About Scale, It's Abo… - Machine Learning Street Talk (MLST) - Apple Podcasts	fetched	144.4 KB	1,139	Raw / Text
tweet	https://x.com/emollick/status/1847419607281213444?t=hA8v5VKw... Tweet by Ethan Mollick	fetched	307 B	51	Raw / Text
webpage	https://futureoflife.org/podcast/tamay-besiroglu-on-ai-in-20... Tamay Besiroglu on AI in 2030: Scaling, Automation, and AI Agents - Future of Life Institute	fetched	235.0 KB	422	Raw / Text
webpage	https://dynomight.substack.com/p/data-wall The real data wall is billions of years of evolution	fetched	188.6 KB	2,482	Raw / Text
notion	https://www.notion.so/c5a5ade965854bab8efcb5165e792ff8	error	-	-	- / -
webpage	https://www.cantorsparadise.com/gpt-4-is-amazing-but-still-s...	error	-	-	- / -
notion	https://www.notion.so/1d698bcf62a444b69accf376da688ebd	error	-	-	- / -
webpage	https://amistrongeryet.substack.com/publish/post/136172063	error	-	-	- / -
webpage	https://amistrongeryet.substack.com/publish/post/137024233	error	-	-	- / -
webpage	https://amistrongeryet.substack.com/publish/post/137667333	error	-	-	- / -
tweet	https://x.com/aidangomez/status/1749760985102164118?t=2BSnLp... Tweet by Aidan Gomez	fetched	273 B	39	Raw / Text
webpage	https://importai.substack.com/p/import-ai-357-facebooks-open...	error	-	-	- / -
webpage	https://www.threads.net/@yannlecun/post/C8bXz_Huu2M/?xmt=AQG... Can generative image models be good world models? This work from @Meta FAIR shows that there is a tradeoff between realism and diversity. The more realistic a generative model becomes, the less diverse it becomes. Realism comes at the cost of coverage. In other words, the most realistic systems are mode-collapsed. My hunch, supported by a growing amount of empirical evidence, is that world models should not be generative. https://arxiv.org/abs/2406.10429	fetched	477.1 KB	0	Raw / Text
tweet	https://x.com/DanHendrycks/status/1798177460028346722?t=T6gQ... Tweet by Dan Hendrycks	fetched	328 B	41	Raw / Text
tweet	https://x.com/mherreshoff/status/1798388740525789296 Tweet by Marcello Herreshoff	fetched	312 B	47	Raw / Text
tweet	https://twitter.com/DanHendrycks/status/1798390576922178004 Tweet by Dan Hendrycks	fetched	302 B	44	Raw / Text
webpage	https://axrp.net/episode/2024/04/25/episode-29-science-of-de... 29 - Science of Deep Learning with Vikrant Varma \| AXRP - the AI X-risk Research Podcast	fetched	123.6 KB	10,000 (truncated)	Raw / Text
webpage	https://thegradientpub.substack.com/p/update-73-language-era... Update #73: Against Language Erasure and Better Long-Context Benchmarking	fetched	395.1 KB	4,284	Raw / Text
webpage	https://thezvi.substack.com/p/ai-79-ready-for-some-football AI #79: Ready for Some Football - by Zvi Mowshowitz	fetched	360.6 KB	9,546	Raw / Text
tweet	https://x.com/EpochAIResearch/status/1828519375302992248 Tweet by Epoch AI	fetched	205 B	26	Raw / Text
webpage	https://thezvi.substack.com/p/on-the-gladstone-report/commen...	error	-	-	- / -
webpage	https://thezvi.substack.com/p/on-devin	error	-	-	- / -
webpage	https://epochai.org/research About Us \| Epoch AI	fetched	81.6 KB	559	Raw / Text
webpage	https://www.cognitiverevolution.ai/flow-engineering-and-code... Flow Engineering and Code Integrity at Scale with Itamar Friedman, CEO of Codium AI	fetched	67.0 KB	867	Raw / Text
webpage	https://thezvi.substack.com/i/141710014/quiet-speculations AI #52: Oops - by Zvi Mowshowitz	fetched	470.5 KB	8,985	Raw / Text
webpage	https://thezvi.substack.com/i/137255844/the-reversal-curse	error	-	-	- / -
webpage	https://www.reddit.com/r/math/comments/19fg9rx/some_perspect...	error	-	-	- / -
webpage	https://www.lesswrong.com/posts/K2D45BNxnZjdpSX2j/ai-timelin... AI Timelines — LessWrong	fetched	3.6 MB	10,000 (truncated)	Raw / Text
webpage	https://amistrongeryet.substack.com/p/e492b0ae-d637-4604-81d...	error	-	-	- / -
webpage	https://amistrongeryet.substack.com/p/cfc9096d-a38c-476c-9eb...	error	-	-	- / -
webpage	https://amistrongeryet.substack.com/p/c06d9966-3cca-4168-9bb...	error	-	-	- / -
webpage	https://amistrongeryet.substack.com/p/684cd573-defd-4a5c-875...	error	-	-	- / -
notion	https://www.notion.so/The-Long-Road-From-ChatGPT-To-AGI-309b...	error	-	-	- / -
webpage	https://arxiv.org/pdf/2310.08560.pdf	fetched	620.2 KB	10,000 (truncated)	Raw / Text
notion	https://www.notion.so/A-Lower-Bound-On-AGI-Timelines-04135c7...	error	-	-	- / -
notion	https://www.notion.so/Two-Paths-to-AGI-1d698bcf62a444b69accf...	error	-	-	- / -
notion	https://www.notion.so/The-Future-Of-Training-Data-beafc63e62...	error	-	-	- / -
notion	https://www.notion.so/Other-People-s-Estimates-6e9f8a6a18414...	error	-	-	- / -
notion	https://www.notion.so/My-Forecast-For-AI-Milestones-c5a5ade9...	error	-	-	- / -
notion	https://www.notion.so/My-Estimate-254770a7d6504b25b8d85c3471...	error	-	-	- / -
notion	https://www.notion.so/Paths-for-Near-Term-Progress-1e3c0d501...	error	-	-	- / -
notion	https://www.notion.so/More-On-Timelines-And-Limitations-Of-L... More On Timelines And Limitations Of LLMs and Missing Capabilities	fetched	3.0 KB	441	Raw / Text
webpage	https://arxiv.org/abs/2311.02462?s=03 [2311.02462] Levels of AGI for Operationalizing Progress on the Path to AGI	fetched	46.3 KB	844	Raw / Text
notion	https://www.notion.so/348fdc5be8ee4451a7b16120edad193d	error	-	-	- / -
webpage	https://www.theintrinsicperspective.com/p/the-1-ai-cant-teac... The #1 AI can't teach basic stuff like reading	fetched	285.9 KB	2,152	Raw / Text
webpage	https://arxiv.org/abs/2407.06023 [2407.06023] Distilling System 2 into System 1	fetched	44.1 KB	739	Raw / Text
webpage	https://thezvi.substack.com/p/ai-74-gpt-4o-mini-me-and-llama... AI #74: GPT-4o Mini Me and Llama 3 - by Zvi Mowshowitz	fetched	566.7 KB	10,000 (truncated)	Raw / Text
tweet	https://x.com/atroyn/status/1815441095721185693 Tweet by anton 🇺🇸	fetched	285 B	42	Raw / Text
webpage	https://importai.substack.com/p/import-ai-380-distributed-13... Import AI 380: Distributed 1.3bn parameter LLM; math AI; and why reality is hard for Ai	fetched	203.2 KB	3,359	Raw / Text
webpage	https://rodneybrooks.com/rodney-brooks-three-laws-of-robotic... Rodney Brooks’ Three Laws of Robotics – Rodney Brooks	fetched	32.3 KB	2,052	Raw / Text
webpage	https://www.ben-evans.com/benedictevans/2024/7/9/the-ai-summ... The AI summer — Benedict Evans	fetched	128.9 KB	2,024	Raw / Text
webpage	https://nicholas.carlini.com/writing/2024/how-i-use-ai.html How I Use "AI"	fetched	2.0 MB	10,000 (truncated)	Raw / Text
webpage	https://deliprao.substack.com/p/juicy-research-ideas-and-how...	error	-	-	- / -
webpage	https://importai.substack.com/p/import-ai-386-googles-chip-d...	error	-	-	- / -
webpage	https://www.interconnects.ai/p/how-scaling-changes-model-beh...	error	-	-	- / -
webpage	https://t.co/aHFcjlVguy JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/teortaxesTex/status/1848103571310747978?t=UZY6... Tweet by Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)	fetched	395 B	47	Raw / Text
notion	https://www.notion.so/a3c8fea72f0a4b2885610d8978036faa Agent Timelines	fetched	1.4 KB	218	Raw / Text
webpage	https://on.ft.com/3yWP1kV Subscribe to the Financial Times	fetched	482.2 KB	5,306	Raw / Text
webpage	https://www.insightpartners.com/ideas/ai-agents-disrupting-a... AI Agents are disrupting automation: Current approaches, market solutions and recommendations \| Insight Partners	fetched	193.8 KB	3,540	Raw / Text
webpage	https://mlsafety.slack.com/team/U04U6QXE18F Slack	fetched	51.4 KB	0	Raw / Text
webpage	https://www.swebench.com/ SWE-bench Leaderboards	fetched	2.6 MB	377	Raw / Text
notion	https://www.notion.so/333a1f173c314cbaada633d7faf21334 More Material for Timelines	fetched	4.6 KB	754	Raw / Text
tweet	https://x.com/emollick/status/1864509024214905060?t=ZVSDTAuR... Tweet by Ethan Mollick	fetched	329 B	50	Raw / Text
webpage	https://www.dwarkeshpatel.com/p/sholto-douglas-trenton-brick... Sholto Douglas & Trenton Bricken - How to Build & Understand GPT-7's Mind	fetched	750.2 KB	10,000 (truncated)	Raw / Text
webpage	https://dblalock.substack.com/p/2024-4-7-arxiv-roundup-dbrx-...	error	-	-	- / -
webpage	https://www.youtube.com/watch?v=i9MiLFMTmfI - YouTube	fetched	1.3 MB	26	Raw / Text
webpage	https://arxiv.org/abs/2402.15627 [2402.15627] MegaScale: Scaling Large Language Model Training to More Than 10,000 GPUs	fetched	50.4 KB	878	Raw / Text
webpage	https://arxiv.org/abs/2204.02311 [2204.02311] PaLM: Scaling Language Modeling with Pathways	fetched	56.4 KB	1,019	Raw / Text
webpage	https://thezvi.substack.com/p/ai-58-stargate-agi	error	-	-	- / -
webpage	https://arxiv.org/abs/2403.13187 [2403.13187] Evolutionary Optimization of Model Merging Recipes	fetched	47.6 KB	814	Raw / Text
webpage	http://sakana.ai/ Sakana AI	fetched	6.8 KB	42	Raw / Text
webpage	https://importai.substack.com/p/import-ai-367-googles-world-... Import AI 367: Google's world-spanning model; breaking AI policy with evolution; $250k for alignment benchmarks	fetched	171.2 KB	2,485	Raw / Text
notion	https://www.notion.so/c48fa565774d4634abfc2667799eea28 More On Timelines And Limitations Of LLMs and Missing Capabilities	fetched	3.0 KB	441	Raw / Text
webpage	https://www.lesswrong.com/posts/nt8PmADqKMaZLZGTC/inside-vie... Inside Views, Impostor Syndrome, and the Great LARP — LessWrong	fetched	1.2 MB	9,822	Raw / Text
webpage	https://sideways-view.com/2018/02/24/takeoff-speeds/ Takeoff speeds – The sideways view	fetched	175.6 KB	7,932	Raw / Text
webpage	https://thegradientpub.substack.com/i/138655621/research-hig...	error	-	-	- / -
webpage	https://arxiv.org/pdf/2305.18654.pdf	fetched	5.9 MB	10,000 (truncated)	Raw / Text
notion	https://www.notion.so/a5c510f6cc6544349be193c9633294d9 How Long Do We Have To Take Action / Timeline Until No Return	fetched	173 B	36	Raw / Text
webpage	https://www.lesswrong.com/posts/JPan54R525D68NoEt/the-date-o... The date of AI Takeover is not the day the AI takes over — LessWrong	fetched	1.2 MB	5,555	Raw / Text
webpage	https://www.aisnakeoil.com/p/can-ai-automate-computational-r... Can AI automate computational reproducibility?	fetched	214.8 KB	1,296	Raw / Text
webpage	https://arxiv.org/abs/1905.02175 [1905.02175] Adversarial Examples Are Not Bugs, They Are Features	fetched	44.5 KB	730	Raw / Text
notion	https://www.notion.so/1134c916374a8037915bf2f3151f22d7 More on o1, AI reasoning, and Test-Time Compute	fetched	22.3 KB	3,425	Raw / Text
webpage	https://bsky.app/profile/lauraruis.bsky.social/post/3lbfeqtk... @lauraruis.bsky.social on Bluesky	fetched	9.0 KB	0	Raw / Text
notion	https://www.notion.so/1024c916374a807a9300edb5d5f8e637 o1	fetched	11.9 KB	1,678	Raw / Text
webpage	https://www.interconnects.ai/p/openai-strawberry-and-inferen... OpenAI’s Strawberry and inference scaling laws	fetched	230.8 KB	2,332	Raw / Text
tweet	https://x.com/youjiacheng/status/1829021962489999393?s=46 Tweet by You Jiacheng	fetched	167 B	21	Raw / Text
tweet	https://twitter.com/natolambert/status/1835686346880626880 Tweet by Nathan Lambert	fetched	62 B	4	Raw / Text
webpage	https://arxiv.org/abs/2305.20050 [2305.20050] Let's Verify Step by Step	fetched	45.9 KB	758	Raw / Text
webpage	https://en.wikipedia.org/wiki/AlphaGo_versus_Lee_Sedol#Game_... AlphaGo versus Lee Sedol - Wikipedia	fetched	1.9 MB	8,690	Raw / Text
tweet	https://x.com/emollick/status/1835659394459013565?t=VgIjyuN2... Tweet by Ethan Mollick	fetched	305 B	54	Raw / Text
notion	https://www.notion.so/b6f1bd454e3a4a04a22791189050d59c	error	-	-	- / -
tweet	https://x.com/MLStreetTalk/status/1834609042230009869?t=JPGx... Tweet by Machine Learning Street Talk	fetched	320 B	44	Raw / Text
tweet	https://x.com/natolambert/status/1834629102097338766?t=nRk8B... Tweet by Nathan Lambert	fetched	197 B	27	Raw / Text
tweet	https://x.com/btibor91/status/1834686946846597281?t=zkGfSi1Z... Tweet by Tibor Blaho	fetched	325 B	53	Raw / Text
webpage	https://simonw.substack.com/p/openais-new-o1-chain-of-though... OpenAI's new o1 chain-of-thought models	fetched	509.7 KB	10,000 (truncated)	Raw / Text
tweet	https://x.com/deedydas/status/1833539735853449360?t=PTSb4uub... Tweet by Deedy	fetched	276 B	30	Raw / Text
tweet	https://x.com/emollick/status/1835090283832442982?t=tv57JQa9... Tweet by Ethan Mollick	fetched	329 B	50	Raw / Text
webpage	https://www.oneusefulthing.org/p/something-new-on-openais-st... Something New: On OpenAI's "Strawberry" and Reasoning	fetched	229.2 KB	1,246	Raw / Text
webpage	https://www.understandingai.org/p/why-i-dont-view-teslas-nea... Comments - Tesla's near-infinite training data isn't a silver bullet	fetched	130.3 KB	96	Raw / Text
tweet	https://x.com/emollick/status/1835366663715258818?t=dNkT4XdO... Tweet by Ethan Mollick	fetched	211 B	31	Raw / Text
tweet	https://x.com/AstronoMisfit/status/1835328355430007164?t=ur1... Tweet by Kyle Kabasares, PhD	fetched	146 B	19	Raw / Text
tweet	https://twitter.com/emollick/status/1836064591479971894 Tweet by Ethan Mollick	fetched	304 B	48	Raw / Text
webpage	https://t.co/8lkD4oRDP2 JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/JuliusSimonelli/status/1834704564618244210?t=H... Tweet by Julius	fetched	121 B	15	Raw / Text
webpage	https://www.mindprison.cc/i/148326057/capabilities-are-uncle... Everything Is a Spy Device and OpenAI’s New o1 Model	fetched	223.5 KB	2,069	Raw / Text
webpage	https://deliprao.substack.com/p/a-few-thoughts-about-o1	error	-	-	- / -
webpage	https://openai.com/index/introducing-openai-o1-preview/	error	-	-	- / -
webpage	https://news.ycombinator.com/item?id=41540902	error	-	-	- / -
webpage	https://arcprize.org/blog/openai-o1-results-arc-prize OpenAI o1 Results on ARC-AGI-Pub	fetched	26.9 KB	1,846	Raw / Text
tweet	https://x.com/rohanpaul_ai/status/1834728874879455250?t=c5Yg... Tweet by Rohan Paul	fetched	330 B	49	Raw / Text
tweet	https://x.com/emollick/status/1834785685661696062?t=E3Cgf9eY... Tweet by Ethan Mollick	fetched	189 B	26	Raw / Text
webpage	https://t.co/qgLJZm8CGr JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/JuliusSimonelli/status/1834377470856040685?t=v... Tweet by Julius	fetched	233 B	36	Raw / Text
tweet	https://x.com/Miles_Brundage/status/1834291595648377091?t=DF... Tweet by Miles Brundage	fetched	290 B	50	Raw / Text
webpage	https://t.co/JssF7NRW4e JavaScript is not available.	fetched	225.2 KB	29	Raw / Text
tweet	https://x.com/ShakeelHashim/status/1834292284193734768?t=Nxy... Tweet by Shakeel	fetched	294 B	40	Raw / Text
webpage	https://arxiv.org/abs/2407.21787 [2407.21787] Large Language Monkeys: Scaling Inference Compute with Repeated Sampling	fetched	48.1 KB	847	Raw / Text
webpage	https://arxiv.org/pdf/2406.02061	fetched	10.0 MB	10,000 (truncated)	Raw / Text
webpage	https://www.maginative.com/article/openai-has-demonstrated-s... OpenAI Has Demonstrated "Strawberry" AI Capabilities to U.S. National Security Officials	fetched	51.8 KB	721	Raw / Text
webpage	https://www.lesswrong.com/posts/8oX4FTRa8MJodArhj/the-inform... The Information: OpenAI shows 'Strawberry' to feds, races to launch it — LessWrong	fetched	564.7 KB	2,096	Raw / Text
webpage	https://www.theinformation.com/articles/openai-shows-strawbe...	error	-	-	- / -
tweet	https://x.com/emollick/status/1836867501138780517?t=sy2vMyow... Tweet by Ethan Mollick	fetched	331 B	53	Raw / Text
webpage	https://t.co/DQQ7jn6wWD JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/tsarnick/status/1836215965912289306?t=BRvHougH... Tweet by Tsarathustra	fetched	243 B	32	Raw / Text
notion	https://www.notion.so/AI-Creativity-and-Problem-Solving-1024... AI Creativity and Problem Solving	fetched	5.1 KB	754	Raw / Text
notion	https://www.notion.so/AI-Creativity-and-Problem-Solving-1024... AI Creativity and Problem Solving	fetched	5.1 KB	754	Raw / Text
notion	https://www.notion.so/1134c916374a80aab9b7c05dbf00c17e More on AI Creativity	fetched	2.7 KB	342	Raw / Text
webpage	https://t.co/ISCj6QKSxw JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/emollick/status/1838240203007205379?t=XWI1JUoH... Tweet by Ethan Mollick	fetched	332 B	45	Raw / Text
tweet	https://x.com/emollick/status/1833190390964359321?t=jqookDxX... Tweet by Ethan Mollick	fetched	324 B	44	Raw / Text
webpage	https://t.co/T7U4G4Aye0	error	-	-	- / -
webpage	https://t.co/iqOURCHRwI JavaScript is not available.	fetched	225.1 KB	29	Raw / Text
tweet	https://x.com/emollick/status/1834201561867698210?t=KxOAMlj5... Tweet by Ethan Mollick	fetched	334 B	45	Raw / Text
webpage	https://www.astralcodexten.com/p/sakana-strawberry-and-scary... Sakana, Strawberry, and Scary AI - by Scott Alexander	fetched	186.3 KB	2,302	Raw / Text
webpage	https://thezvi.substack.com/p/ai-81-alpha-proteo AI #81: Alpha Proteo - by Zvi Mowshowitz	fetched	406.0 KB	10,000 (truncated)	Raw / Text
webpage	https://marginalrevolution.com/marginalrevolution/2024/09/ll... LLMs are Creative Reasoners - Marginal REVOLUTION	fetched	98.0 KB	397	Raw / Text
webpage	https://arxiv.org/abs/2409.04109 [2409.04109] Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers	fetched	46.2 KB	816	Raw / Text
webpage	https://dynomight.substack.com/p/automated?utm_source=post-e... Thoughts while watching myself be automated	fetched	160.6 KB	2,086	Raw / Text
tweet	https://x.com/emollick/status/1846227413832765819?t=DU4wKha8... Tweet by Ethan Mollick	fetched	178 B	25	Raw / Text
tweet	https://x.com/natolambert/status/1848352562161848445?t=yhKma... Tweet by Nathan Lambert	fetched	49 B	1	Raw / Text

Type	URL / Preview	Note
tweet	https://x.com/i/status/1999994684203045003 (I know I'm a stuck record) An important assumption in AI discourse is that sufficiently capable gen...	links not explored
web	https://blog.cosmos-institute.org/p/faster-horses Faster Horses - by Harry Law - Cosmos Institute	links not explored
gdoc	https://docs.google.com/document/d/1Ab8Q8_F22drmQF... Google Doc	content not explored
gdoc	https://docs.google.com/document/d/1PH7bNtJdy55IPI... Google Doc	content not explored
tweet	https://x.com/snewmanpv/status/2005329341845303612 Underlying this trend is the fact that AIs are now entering the zone of real utility. Progress has n...	links not explored
web	https://www.lesswrong.com/posts/u6Lacc7wx4yYkBQ3r/... Insights into Claude Opus 4.5 from Pokémon — LessWrong	links not explored
web	https://t.co/YH7W3F9FtQ Cognition all the way down 2.0: neuroscience beyond neurons in the diverse intelligence era \| Synthese	links not explored
web	https://t.co/UKIULAD5wU JavaScript is not available.	links not explored
tweet	https://x.com/anilkseth/status/1987894412584566876... "Cognition all the way down". Great to see this fine new paper from @robertchisciure & @drmichaelle...	links not explored
tweet	https://x.com/i/status/2000413758729162889 In general, all the chatbots seem to struggle with files in a way that CLI versions do not. Gemini w...	links not explored
tweet	https://x.com/i/status/1998778005129224412 An understudied, high-value issue in agents is how to have them write the best summaries of what the...	links not explored
web	https://jasmi.news/p/42-notes-on-ai-and-work 42 notes on AI & work - by Jasmine Sun	links not explored
web	https://www.understandingai.org/p/i-got-fooled-by-... I got fooled by AI-for-science hype—here's what it taught me	links not explored
web	https://www.lesswrong.com/posts/4mvphwx5pdsZLMmpY/... Recent AI model progress feels mostly like bullshit — LessWrong	links not explored
web	https://ringohosp.substack.com/p/what-artificial-i... What artificial intelligence can do for a small independent hotel	links not explored
web	https://gwern.net/ai-daydreaming LLM Daydreaming · Gwern.net	links not explored
tweet	https://x.com/1a3orn/status/1997056050403725373 This METR chart is my fav depiction of how human continual learning >> LLM continual learning. Huma...	links not explored
web	https://www.dwarkesh.com/p/thoughts-on-ai-progress... Thoughts on AI progress (Dec 2025) - by Dwarkesh Patel	links not explored
tweet	https://x.com/Jack_W_Lindsey/status/19933890569323... One analysis from our pre-release audit of Opus 4.5 stands out to me. Our behavioral evals uncovered...	links not explored
web	https://www.strangeloopcanon.com/p/epicycles-all-t... Epicycles All The Way Down - by Rohit Krishnan	links not explored
tweet	https://x.com/mikeknoop/status/1993402375944679673... "scaling sucked out all the oxygen in the room, everyone converged to the same ideas" --> new ideas ...	links not explored
tweet	https://x.com/QuintinPope5/status/1991472282422899... It is genuinely weird how many complex dynamical feedback loops end up producing smooth exponentials...	links not explored
tweet	https://x.com/MLStreetTalk/status/1989961238839693... Intelligence is not just compression. The future is much harder to know than the past, especially f...	links not explored
tweet	https://x.com/fchollet/status/1989340153114976598 All the great breakthroughs in science are, at their core, compression. They take a complex mess of ...	links not explored
tweet	https://x.com/HjalmarWijk/status/19855299568905302... I really liked your post as well (https://t.co/vGhYMbzwlW) and it moved me toward thinking there are...	links not explored
gdoc	https://docs.google.com/document/d/1p7l-KM31hETcNw... Google Doc	content not explored
web	https://youtu.be/mYDSSRS-B5U?si=EvNetk1JVc7e6HXt&t... - YouTube	links not explored
web	https://arxiv.org/abs/2502.05167 [2502.05167] NoLiMa: Long-Context Evaluation Beyond Literal Matching	links not explored
tweet	https://x.com/emollick/status/1985507610963968447 AI laziness remains one of the wildest common failure states when you think about it. Also one of t...	links not explored
web	https://80000hours.org/podcast/episodes/holden-kar... Holden Karnofsky on dozens of amazing opportunities to make AI safer — and all his AGI takes \| 80,000 Hours	links not explored
tweet	https://x.com/karpathy/status/1973435013875314729?... Finally had a chance to listen through this pod with Sutton, which was interesting and amusing. As ...	links not explored
tweet	https://x.com/dwarkesh_sp/status/19792590410137317... The most interesting part for me is where @karpathy describes why LLMs aren't able to learn like hu...	links not explored
tweet	https://x.com/karpathy/status/1979644538185752935?... My pleasure to come on Dwarkesh last week, I thought the questions and conversation were really good...	links not explored
tweet	https://x.com/ryanpgreenblatt/status/1981012208332... https://t.co/duJDFyPx8o	links not explored
web	https://www.dwarkesh.com/p/andrej-karpathy Andrej Karpathy — AGI is still a decade away	links not explored
web	https://www.henrikkarlsson.xyz/p/wordless-thought When is better to think without words? - by Henrik Karlsson	links not explored
web	https://ai-frontiers.org/articles/agis-last-bottle... AGI's Last Bottlenecks \| AI Frontiers	links not explored
tweet	https://x.com/emollick/status/1980704687377486182?... Papers like this show that are a lot of potential pathways forward on some of the hardest outstandin...	links not explored
web	https://t.co/9rAySHO5L2 JavaScript is not available.	links not explored
tweet	https://x.com/emollick/status/1980601776710525103?... Looking back at an exponentially improving technology & you will see how momentum led to R&D which o...	links not explored
web	https://t.co/ZwhDegc4NO JavaScript is not available.	links not explored
tweet	https://x.com/tobyordoxford/status/198035135322776... New post on RL scaling: Careful analysis of OpenAI’s public benchmarks reveals RL scales far worse t...	links not explored
web	https://t.co/X2EocaaZjQ Inference Scaling Reshapes AI Governance — Toby Ord	links not explored
tweet	https://x.com/tobyordoxford/status/197420019350471... So it looks like most of the gains are coming from the ability to spend more compute on each answer ...	links not explored
web	https://thezvi.substack.com/p/bending-the-curve Bending The Curve - by Zvi Mowshowitz	links not explored
web	https://www.dwarkesh.com/p/thoughts-on-sutton Some thoughts on the Sutton interview - by Dwarkesh Patel	links not explored
web	https://www.interconnects.ai/p/thoughts-on-the-cur... Thoughts on The Curve - by Nathan Lambert - Interconnects	links not explored
web	https://t.co/qyMUb0c2Oh JavaScript is not available.	links not explored
tweet	https://x.com/teortaxesTex/status/1971854577755377... I think @RichardSSutton is spiritually right, but offers non-constructively framed arguments (much ...	links not explored
web	https://importai.substack.com/p/import-ai-430-emer... Import AI 430: Emergence in video models; Unitree backdoor; preventative strikes to take down AGI projects	links not explored
tweet	https://x.com/snewmanpv/status/1970528715403854327... Why do coding agents cheat on unit tests (e.g. by modifying the test to always return true)? The ob...	links not explored
tweet	https://x.com/snewmanpv/status/1970528715403854327 Why do coding agents cheat on unit tests (e.g. by modifying the test to always return true)? The ob...	links not explored
web	https://epochai.substack.com/p/the-huge-potential-... The huge potential implications of long-context inference	links not explored
tweet	https://x.com/emollick/status/1968339804975948274?... A big issue with today’s agent implementation is that they don’t ask questions, even when the thinki...	links not explored
tweet	https://x.com/mjbukow/status/1962186888103747892 I think we've known for a while that LLMs can only replay data that is in-distribution. That doesn't...	links not explored
tweet	https://x.com/livgorton/status/1960378468807447026... 17/ If this is right, it reframes the robustness problem: adversarial vulnerability might be the pri...	links not explored
web	https://www.linkedin.com/posts/tcpeixoto_help-need... Is "human-in-the-loop" a luxury for the privileged? \| Tiago C. Peixoto posted on the topic \| LinkedIn	links not explored
web	https://www.lesswrong.com/posts/2ssPfDpdrjaM2rMbn/... My AGI timeline updates from GPT-5 (and 2025 so far) — LessWrong	links not explored
web	https://importai.substack.com/p/import-ai-425-ipho... Import AI 425: iPhone video generation; subtle misalignment; making open weight models safe through surgical deletion	links not explored
web	https://www.interconnects.ai/p/contra-dwarkesh-on-... Contra Dwarkesh on Continual Learning - by Nathan Lambert	links not explored
tweet	https://x.com/Simeon_Cps/status/195651948527768421... I have this theory that we are in a period of increasing marginal utility of capabilities. GPT-2 to ...	links not explored
web	https://peterwildeford.substack.com/p/gpt-5-a-smal... GPT-5: a small step for intelligence, a giant leap for normal people	links not explored
web	https://ethanding.substack.com/p/ai-subscriptions-... tokens are getting more expensive - by Ethan Ding	links not explored
web	https://thezvi.substack.com/i/169772920/city-in-a-... AI #128: Four Hours Until Probably Not The Apocalypse	links not explored
tweet	https://x.com/ben_j_todd/status/193428418992850148... Why can AIs code for 1h but not 10h? A simple explanation: if there's a 10% chance of error per 10m...	links not explored
web	http://henrikkarlsson.xyz/p/attention Almost anything you give sustained attention to will begin to loop on itself and bloom	links not explored
web	https://substack.com/redirect/ce40c09e-8b1b-44b4-9... Using spaced repetition systems to see through a piece of mathematics	links not explored
web	https://substack.com/redirect/a43876f1-09fb-4d47-9... JavaScript is not available.	links not explored
tweet	https://x.com/MLStreetTalk/status/1963492648708132... We had a wonderful chat with Prof. Cris Moore of the Santa Fe Institute (@SFIscience), famous for wr...	links not explored
web	https://www.interconnects.ai/p/brakes-on-an-intell... State of play of AI progress (and related brakes on an intelligence explosion)	links not explored
web	https://podcasts.apple.com/us/podcast/the-mathemat... The Mathematical Foundations o… - Machine Learning Street Talk (MLST) - Apple Podcasts	links not explored
web	https://www.lesswrong.com/posts/vvgND6aLjuDR6QzDF/... My model of what is going on with LLMs — LessWrong	links not explored
web	https://www.lesswrong.com/posts/GADJFwHzNZKg2Ndti/... Have LLMs Generated Novel Insights? — LessWrong	links not explored
web	https://helentoner.substack.com/p/2-big-questions-... 2 big questions for AI progress in 2025-2026	links not explored
web	https://lemmata.substack.com/p/what-i-wish-i-knew-... What I Wish I Knew About FrontierMath - by Greg Burnham	links not explored
web	https://amistrongeryet.substack.com/p/were-finding... Comments - We're Finding Out What Humans are Bad At	links not explored
tweet	https://x.com/emollick/status/1892108434171949321?... As I have written many times, AI is not naturally a great tutor, it offers explanations but, without...	links not explored
web	https://erictopol.substack.com/p/when-doctors-with... When Doctors With A.I. Are Outperformed by A.I. Alone	links not explored
web	https://t.co/b3ZuMjO0ZJ From OpenAI to DeepSeek, companies say AI can “reason” now. Is it true? \| Vox	links not explored
tweet	https://x.com/SigalSamuel/status/18929978618868204... The big AI story of the past 6 months is: Companies now claim that their AI models are capable of ge...	links not explored
web	https://arxiv.org/abs/2410.06992 [2410.06992] SWE-Bench+: Enhanced Coding Benchmark for LLMs	links not explored
web	https://epochai.substack.com/p/ai-progress-is-abou... AI progress is about to speed up - by Ege Erdil - Epoch AI	links not explored
tweet	https://x.com/AndrewCritchPhD/status/1891887600102... This indeed works and saves much time: you can tell an LLM to enumerate hypotheses and testing stra...	links not explored
web	https://thezvi.substack.com/p/ai-104-american-stat... AI #104: American State Capacity on the Brink	links not explored
tweet	https://x.com/buccocapital/status/1890745551995424... People were curious, so here's how I'm using Deep Research. I'll walk through the prompting and then...	links not explored
tweet	https://x.com/littmath/status/1898461323391815820 In this thread I want to share some thoughts about the FrontierMath benchmark, on which, according t...	links not explored
web	https://www.theintrinsicperspective.com/p/ai-plays... AI plays Pokémon, but so does π. Tesla's "brand collapse." Neuroscience scaling laws. Why do I know all the songs on the radio?	links not explored
tweet	https://x.com/slatestarcodex/status/18964571932157... I agree that superintelligence can't just brute force every possible combination of ideas and keep t...	links not explored
web	https://t.co/XrZUlbKkek JavaScript is not available.	links not explored
tweet	https://x.com/emollick/status/1898802108926661012?... If it turns out LLMs are only capable of recombinatory innovation (finding novel connections among e...	links not explored
web	https://lemmata.substack.com/p/alphaproof-and-the-... AlphaProof and the IMO - by Greg Burnham - Lemmata	links not explored
web	https://lemmata.substack.com/p/o3-o4-mini-largely-... o3, o4-mini: Largely Incremental at Math - by Greg Burnham	links not explored
tweet	https://x.com/snewmanpv/status/1929724463916241351 Another great post from @GregHBurnham. He takes a deep dive into an LLM's proof of a tough math prob...	links not explored
web	https://t.co/uAHid8PNFd JavaScript is not available.	links not explored
tweet	https://x.com/vitrupo/status/1910544354709762111?t... Sam Altman asks: why are scaling laws a property of the universe? Daniel Selsam says intelligence e...	links not explored
web	https://thegradientpub.substack.com/i/148749478/re... Update #83: AI Music Fraud and PlanSearch - by Cole Frank	links not explored
web	https://thegradientpub.substack.com/p/update-83-ai... Update #83: AI Music Fraud and PlanSearch - by Cole Frank	links not explored
tweet	https://x.com/Simeon_Cps/status/186841156076881114... Sounds about right. Maths has all the ingredients: Data generation (theorem provers + lean), verifia...	links not explored
tweet	https://twitter.com/MFarajtabar/status/18444568936... 5/ #Result 2: The fragility of supposed LLM reasoning. LLMs remain sensitive to changes in proper na...	links not explored
web	https://pca.st/episode/27144be7-bcae-4e16-b1e4-403... The Startup Powering The Data Behind AGI - Pocket Casts	links not explored
tweet	https://x.com/natolambert/status/18461955444585145... You spend more tokens on inference, your scores go up, its a simple business. This is clever work t...	links not explored
web	https://amistrongeryet.substack.com/i/114809888/gp... I’m a Senior Software Engineer. What Will It Take For An AI To Do My Job?	links not explored
web	https://chatgpt.com/share/6701b6c0-8b50-800b-a40f-... ChatGPT - String Difference Algorithm JavaScript	links not explored
web	https://chatgpt.com/share/6701b757-2bac-800b-a66a-... ChatGPT - String Comparison and Reconstruction	links not explored
tweet	https://x.com/colin_fraser/status/1833953195430719... ok here's my full review of this paper. It's easy and short, you should just read it if you want to....	links not explored
web	https://t.co/CwTAMbWnUO [2409.04109] Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers	links not explored
tweet	https://twitter.com/TrentonBricken/status/18466529... Excited this small update is out! https://t.co/ENwx7rld5F We explore using dictionary learning fea...	links not explored
tweet	https://twitter.com/emollick/status/18538578423121... Interesting addition to the discussion over whether AI video models are world models. But even if t...	links not explored
web	https://thezvi.substack.com/i/148809120/intelligen... AI #82: The Governor Ponders - by Zvi Mowshowitz	links not explored
tweet	https://x.com/IntuitMachine/status/184323859466993... 1/n Finding Intelligence at the Edge of Chaos The pursuit of artificial intelligence has traditiona...	links not explored
tweet	https://x.com/tszzl/status/1844293154549399866 it’s obvious now that superhuman machine performance in various domains is clearly possible without ...	links not explored
tweet	https://x.com/lateinteraction/status/1892724949074... I suspect if only a 10% random sample of us in the field could stay, they'd make roughly similar pro...	links not explored
tweet	https://x.com/_xjdr/status/1875295492906828005?t=m... if you will forgive a bit of anthropomorphizing, to the extent models think, they do not think in to...	links not explored
web	https://arxiv.org/pdf/2311.02462 https://arxiv.org/pdf/2311.02462	links not explored
web	https://semianalysis.com/2024/12/11/scaling-laws-o... Scaling Laws - O1 Pro Architecture, Reasoning Training Infrastructure, Orion and Claude 3.5 Opus “Failures”	links not explored
tweet	https://twitter.com/Miles_Brundage/status/18620700... Certainly it’ll be easier to directly target areas that have clear reward functions but even besides...	links not explored
tweet	https://twitter.com/DKokotajlo67142/status/1862095... Whereas I think that each gap between milestones will be shorter than the previous one, e.g. 1-day t...	links not explored
tweet	https://twitter.com/deanwball/status/1861975137425... Finally got a chance to read this piece, which argues that o1-style reasoners will not generalize be...	links not explored
tweet	https://twitter.com/random_walker/status/186685835... There's an emerging cope among some AI developers and researchers that progress in capabilities isn'...	links not explored
web	https://pca.st/episode/d9d8178e-9f7e-4102-bf2c-aa3... Autonomous Organizations: Vending Bench & Beyond, w/ Lukas Petersson & Axel Backlund of Andon Labs - Pocket Casts	links not explored
web	https://www.lesswrong.com/posts/HiTjDZyWdLEGCDzqu?... Implications of the inference scaling paradigm for AI safety — LessWrong	links not explored
web	https://www.aisnakeoil.com/p/is-ai-progress-slowin... Is AI progress slowing down?	links not explored
web	https://www.aisnakeoil.com/p/is-ai-progress-slowin... Is AI progress slowing down?	links not explored
web	https://www.saysmaybe.com/latest-work/selective-pe... Selective Perspectives	links not explored
web	https://arxiv.org/abs/2407.21783 [2407.21783] The Llama 3 Herd of Models	links not explored
web	https://arxiv.org/abs/2402.00838 [2402.00838] OLMo: Accelerating the Science of Language Models	links not explored
tweet	https://x.com/snewmanpv/status/1870496367233438200 Is it though? I thought the idea of RSI was AI recursively creating more powerful AI *in a cycle of ...	links not explored
tweet	https://twitter.com/S_OhEigeartaigh/status/1868606... The builders are saying quite clearly what they're aiming for. But the implications are too wild. & ...	links not explored
web	https://thezvi.substack.com/p/ai-94-not-now-google AI #94: Not Now, Google - by Zvi Mowshowitz	links not explored
web	https://www.hyperdimensional.co/p/thresholds Thresholds - by Dean W. Ball - Hyperdimensional	links not explored
web	https://substack.com/redirect/b54107cc-741f-41fb-9... An Extraordinary Alien - by Dean W. Ball - Hyperdimensional	links not explored
web	https://t.co/7gG705nmNL JavaScript is not available.	links not explored
tweet	https://x.com/tsarnick/status/1861921602235150545?... Yann Lecun says his estimate for the creation of human-level AI is not that different from Sam Altma...	links not explored
tweet	https://x.com/RichardMCNgo/status/1861450205109096... My timelines are “we’re close enough to AGI that we can no longer meaningfully talk about AGI as a s...	links not explored
tweet	https://x.com/DKokotajlo67142/status/1862095827700... Whereas I think that each gap between milestones will be shorter than the previous one, e.g. 1-day t...	links not explored
web	https://www.theintrinsicperspective.com/p/great-sc... Great scientists follow intuition and beauty, not rationality	links not explored
web	https://www.newcomer.co/p/our-biggest-takeaways-fr... Our Biggest Takeaways from Cerebral Valley: No Wall on AI Progress Despite the Scaling Law Debate	links not explored
tweet	https://x.com/deanwball/status/1860114571111084250... Counterintuitively, AI research may end up being one of the easiest tasks to substantively automate....	links not explored
web	https://t.co/VzldSvcMCU JavaScript is not available.	links not explored
tweet	https://x.com/eli_lifland/status/18600872628491717... Takeaways re: AI R&D performance: 1. Claude 3.5 Sonnet reaches ~50th percentile human baseline 8-hou...	links not explored
web	https://www.understandingai.org/p/are-transformer-... Are transformer-based foundation models running out of steam?	links not explored
tweet	https://x.com/polynoamial/status/18550376895331782... I've heard people claim that Sam is just drumming up hype, but from what I've seen everything he's s...	links not explored
web	https://t.co/lcy9sVjC4D JavaScript is not available.	links not explored
tweet	https://x.com/nrehiew_/status/1861464774904537215?... If you actually want to learn about test time scaling, just watch this talk instead. Significant hig...	links not explored
web	https://marginalrevolution.com/marginalrevolution/... How badly do humans misjudge AIs? - Marginal REVOLUTION	links not explored
web	https://sites.google.com/view/raphaelraux/research... raphaelraux - Research	links not explored
tweet	https://x.com/BethMayBarnes/status/186006545082420... The tasks are also limited in scope vs real-life ML research. We've tried to include aspects like in...	links not explored
web	https://www.interconnects.ai/p/scaling-realities Scaling realities - by Nathan Lambert - Interconnects	links not explored
tweet	https://x.com/emollick/status/1856950856395984940?... For what it is worth, the idea that there is no barrier to scaling to AGI is something multiple head...	links not explored
tweet	https://x.com/Miles_Brundage/status/18568099700897... Betting against AI scaling continuing to yield big gains is a bad idea. Would recommend that anyone...	links not explored
tweet	https://x.com/swyx/status/1856776660986859632?t=OC... balanced thoughts on the "ai is hitting a wall" meme: - gpt5, 3 opus, gemini 2 delays are real. no ...	links not explored
web	https://www.theintrinsicperspective.com/p/ai-progr... AI progress has plateaued below GPT-5 level - by Erik Hoel	links not explored
web	https://podcasts.apple.com/us/podcast/its-not-abou... It's Not About Scale, It's Abo… - Machine Learning Street Talk (MLST) - Apple Podcasts	links not explored
tweet	https://x.com/emollick/status/1847419607281213444?... Some comments seem to be missing some of the point - AI has always had mixed performance on tasks, t...	links not explored
web	https://futureoflife.org/podcast/tamay-besiroglu-o... Tamay Besiroglu on AI in 2030: Scaling, Automation, and AI Agents - Future of Life Institute	links not explored
web	https://dynomight.substack.com/p/data-wall The real data wall is billions of years of evolution	links not explored
tweet	https://x.com/aidangomez/status/174976098510216411... The fact that LLMs lack a constrained inner monologue feels like a fairly catastrophic weakness that...	links not explored
web	https://www.threads.net/@yannlecun/post/C8bXz_Huu2... Can generative image models be good world models? This work from @Meta FAIR shows that there is a tradeoff between realism and diversity. The more realistic a generative model becomes, the less diverse it becomes. Realism comes at the cost of coverage. In other words, the most realistic systems are mode-collapsed. My hunch, supported by a growing amount of empirical evidence, is that world models should not be generative. https://arxiv.org/abs/2406.10429	links not explored
tweet	https://x.com/DanHendrycks/status/1798177460028346... This is worth checking out. Minor criticisms: I think industry's "algorithmic secrets" are not a ve...	links not explored
tweet	https://x.com/mherreshoff/status/17983887405257892... Your barrel transcript is an example of a specific LLM pathology I call "template fixation". That's ...	links not explored
tweet	https://twitter.com/DanHendrycks/status/1798390576... I think the case for an intelligence explosion rests on suspicious algorithmic efficiency estimates....	links not explored
web	https://axrp.net/episode/2024/04/25/episode-29-sci... 29 - Science of Deep Learning with Vikrant Varma \| AXRP - the AI X-risk Research Podcast	links not explored
web	https://thegradientpub.substack.com/p/update-73-la... Update #73: Against Language Erasure and Better Long-Context Benchmarking	links not explored
web	https://thezvi.substack.com/p/ai-79-ready-for-some... AI #79: Ready for Some Football - by Zvi Mowshowitz	links not explored
tweet	https://x.com/EpochAIResearch/status/1828519375302... 1/ Automating AI research could rapidly drive innovation. But which research tasks are nearing autom...	links not explored
web	https://epochai.org/research About Us \| Epoch AI	links not explored
web	https://www.cognitiverevolution.ai/flow-engineerin... Flow Engineering and Code Integrity at Scale with Itamar Friedman, CEO of Codium AI	links not explored
web	https://thezvi.substack.com/i/141710014/quiet-spec... AI #52: Oops - by Zvi Mowshowitz	links not explored
web	https://www.lesswrong.com/posts/K2D45BNxnZjdpSX2j/... AI Timelines — LessWrong	links not explored
web	https://arxiv.org/pdf/2310.08560.pdf https://arxiv.org/pdf/2310.08560.pdf	links not explored
web	https://arxiv.org/abs/2311.02462?s=03 [2311.02462] Levels of AGI for Operationalizing Progress on the Path to AGI	links not explored
web	https://www.theintrinsicperspective.com/p/the-1-ai... The #1 AI can't teach basic stuff like reading	links not explored
web	https://arxiv.org/abs/2407.06023 [2407.06023] Distilling System 2 into System 1	links not explored
web	https://thezvi.substack.com/p/ai-74-gpt-4o-mini-me... AI #74: GPT-4o Mini Me and Llama 3 - by Zvi Mowshowitz	links not explored
tweet	https://x.com/atroyn/status/1815441095721185693 my theory of why people inside the labs have very different timelines from people outside is because...	links not explored
web	https://importai.substack.com/p/import-ai-380-dist... Import AI 380: Distributed 1.3bn parameter LLM; math AI; and why reality is hard for Ai	links not explored
web	https://rodneybrooks.com/rodney-brooks-three-laws-... Rodney Brooks’ Three Laws of Robotics – Rodney Brooks	links not explored
web	https://www.ben-evans.com/benedictevans/2024/7/9/t... The AI summer — Benedict Evans	links not explored
web	https://nicholas.carlini.com/writing/2024/how-i-us... How I Use "AI"	links not explored
web	https://t.co/aHFcjlVguy JavaScript is not available.	links not explored
tweet	https://x.com/teortaxesTex/status/1848103571310747... btw I've started having doubts about near-term transformative AGI. Timelines up to 2035 or 2040 seem...	links not explored
web	https://on.ft.com/3yWP1kV Subscribe to the Financial Times	links not explored
web	https://www.insightpartners.com/ideas/ai-agents-di... AI Agents are disrupting automation: Current approaches, market solutions and recommendations \| Insight Partners	links not explored
web	https://mlsafety.slack.com/team/U04U6QXE18F Slack	links not explored
web	https://www.swebench.com/ SWE-bench Leaderboards	links not explored
tweet	https://x.com/emollick/status/1864509024214905060?... It is worth noting that this is an increasingly common message from insiders at the big AI labs. It...	links not explored
web	https://www.dwarkeshpatel.com/p/sholto-douglas-tre... Sholto Douglas & Trenton Bricken - How to Build & Understand GPT-7's Mind	links not explored
web	https://www.youtube.com/watch?v=i9MiLFMTmfI - YouTube	links not explored
web	https://arxiv.org/abs/2402.15627 [2402.15627] MegaScale: Scaling Large Language Model Training to More Than 10,000 GPUs	links not explored
web	https://arxiv.org/abs/2204.02311 [2204.02311] PaLM: Scaling Language Modeling with Pathways	links not explored
web	https://arxiv.org/abs/2403.13187 [2403.13187] Evolutionary Optimization of Model Merging Recipes	links not explored
web	http://sakana.ai/ Sakana AI	links not explored
web	https://importai.substack.com/p/import-ai-367-goog... Import AI 367: Google's world-spanning model; breaking AI policy with evolution; $250k for alignment benchmarks	links not explored
web	https://www.lesswrong.com/posts/nt8PmADqKMaZLZGTC/... Inside Views, Impostor Syndrome, and the Great LARP — LessWrong	links not explored
web	https://sideways-view.com/2018/02/24/takeoff-speed... Takeoff speeds – The sideways view	links not explored
web	https://arxiv.org/pdf/2305.18654.pdf https://arxiv.org/pdf/2305.18654.pdf	links not explored
web	https://www.lesswrong.com/posts/JPan54R525D68NoEt/... The date of AI Takeover is not the day the AI takes over — LessWrong	links not explored
web	https://www.aisnakeoil.com/p/can-ai-automate-compu... Can AI automate computational reproducibility?	links not explored
web	https://arxiv.org/abs/1905.02175 [1905.02175] Adversarial Examples Are Not Bugs, They Are Features	links not explored
web	https://bsky.app/profile/lauraruis.bsky.social/pos... @lauraruis.bsky.social on Bluesky	links not explored
web	https://www.interconnects.ai/p/openai-strawberry-a... OpenAI’s Strawberry and inference scaling laws	links not explored
tweet	https://x.com/youjiacheng/status/18290219624899993... $NVDA CFO: "Over the trailing four quarters, we estimate that inference drove more than 40% of our d...	links not explored
tweet	https://twitter.com/natolambert/status/18356863468... post is out: https://t.co/yykDn3wynf	links not explored
web	https://arxiv.org/abs/2305.20050 [2305.20050] Let's Verify Step by Step	links not explored
web	https://en.wikipedia.org/wiki/AlphaGo_versus_Lee_S... AlphaGo versus Lee Sedol - Wikipedia	links not explored
tweet	https://x.com/emollick/status/1835659394459013565?... In this post, I tried to lay out, in a non-technical way, where we are in the development of AI, the...	links not explored
tweet	https://x.com/MLStreetTalk/status/1834609042230009... Reasoning is knowledge acquisition. The new OpenAI models don't reason, they simply memorise reaso...	links not explored
tweet	https://x.com/natolambert/status/18346291020973387... People are going to underestimate o1 style systems until there's an equivalent "move 37" moment from...	links not explored
tweet	https://x.com/btibor91/status/1834686946846597281?... Summary of what we have learned during AMA hour with the OpenAI o1 team today Model Names and Reaso...	links not explored
web	https://simonw.substack.com/p/openais-new-o1-chain... OpenAI's new o1 chain-of-thought models	links not explored
tweet	https://x.com/deedydas/status/1833539735853449360?... Research published by Google Deepmind reveals OpenAI Strawberry's approach. Searches at inference t...	links not explored
tweet	https://x.com/emollick/status/1835090283832442982?... Note o1 is showing how unprepared we are for testing high level AIs to figure out what they are good...	links not explored
web	https://www.oneusefulthing.org/p/something-new-on-... Something New: On OpenAI's "Strawberry" and Reasoning	links not explored
web	https://www.understandingai.org/p/why-i-dont-view-... Comments - Tesla's near-infinite training data isn't a silver bullet	links not explored
tweet	https://x.com/emollick/status/1835366663715258818?... Question @AstronoMisfit - are you sure this wasn’t reproduced from a public github repo in the train...	links not explored
tweet	https://x.com/AstronoMisfit/status/183532835543000... That feeling when ChatGPT o1 accomplishes in 1 hour what took you about a year in your PhD: https://...	links not explored
tweet	https://twitter.com/emollick/status/18360645914799... After spending a decent amount of time with o1-preview, I would be very surprised if it is not able ...	links not explored
web	https://t.co/8lkD4oRDP2 JavaScript is not available.	links not explored
tweet	https://x.com/JuliusSimonelli/status/1834704564618... Another instance of o1-preview missing the part where e I tell it the answer pic.twitter.com/8lkD4oR...	links not explored
web	https://www.mindprison.cc/i/148326057/capabilities... Everything Is a Spy Device and OpenAI’s New o1 Model	links not explored
web	https://arcprize.org/blog/openai-o1-results-arc-pr... OpenAI o1 Results on ARC-AGI-Pub	links not explored
tweet	https://x.com/rohanpaul_ai/status/1834728874879455... Ilya Sutskever was a co-author of this Paper. Helps to put some light on how the latest o1 🍓 Model...	links not explored
tweet	https://x.com/emollick/status/1834785685661696062?... I think it is worth reading the red team reports on o1. Some interesting stuff in there (and these a...	links not explored
web	https://t.co/qgLJZm8CGr JavaScript is not available.	links not explored
tweet	https://x.com/JuliusSimonelli/status/1834377470856... ChatGPT o1-preview still doesn't know what it knows... If you ask it for three people with the same ...	links not explored
tweet	https://x.com/Miles_Brundage/status/18342915956483... My team + many others across OpenAI have been thinking for a while about the implications of AIs tha...	links not explored
web	https://t.co/JssF7NRW4e JavaScript is not available.	links not explored
tweet	https://x.com/ShakeelHashim/status/183429228419373... I took a stab at summarising the OpenAI o1 system card. A few bits in particular jumped out at me: ...	links not explored
web	https://arxiv.org/abs/2407.21787 [2407.21787] Large Language Monkeys: Scaling Inference Compute with Repeated Sampling	links not explored
web	https://arxiv.org/pdf/2406.02061 https://arxiv.org/pdf/2406.02061	links not explored
web	https://www.maginative.com/article/openai-has-demo... OpenAI Has Demonstrated "Strawberry" AI Capabilities to U.S. National Security Officials	links not explored
web	https://www.lesswrong.com/posts/8oX4FTRa8MJodArhj/... The Information: OpenAI shows 'Strawberry' to feds, races to launch it — LessWrong	links not explored
tweet	https://x.com/emollick/status/1836867501138780517?... The issue with o1 (or successors) helping with PhD level problems is that most people don’t have a t...	links not explored
web	https://t.co/DQQ7jn6wWD JavaScript is not available.	links not explored
tweet	https://x.com/tsarnick/status/1836215965912289306?... OpenAI's Noam Brown says that while AI model performance scales roughly equivalently with more train...	links not explored
web	https://t.co/ISCj6QKSxw JavaScript is not available.	links not explored
tweet	https://x.com/emollick/status/1838240203007205379?... Comparing the creativity of a representative human sample to GPT-4 finds "the creative ideas produce...	links not explored
tweet	https://x.com/emollick/status/1833190390964359321?... We have increasing evidence that today's AIs can, indeed, generate novel ideas This paper shows tha...	links not explored
web	https://t.co/iqOURCHRwI JavaScript is not available.	links not explored
tweet	https://x.com/emollick/status/1834201561867698210?... I don’t necessarily agree with it, but this is an interesting theoretical paper that lays out the ar...	links not explored
web	https://www.astralcodexten.com/p/sakana-strawberry... Sakana, Strawberry, and Scary AI - by Scott Alexander	links not explored
web	https://thezvi.substack.com/p/ai-81-alpha-proteo AI #81: Alpha Proteo - by Zvi Mowshowitz	links not explored
web	https://marginalrevolution.com/marginalrevolution/... LLMs are Creative Reasoners - Marginal REVOLUTION	links not explored
web	https://arxiv.org/abs/2409.04109 [2409.04109] Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers	links not explored
web	https://dynomight.substack.com/p/automated?utm_sou... Thoughts while watching myself be automated	links not explored
tweet	https://x.com/emollick/status/1846227413832765819?... A nice demonstration about why world models will not be enough for AI to replace the artistry of top...	links not explored
tweet	https://x.com/natolambert/status/18483525621618484... https://t.co/2Ct7L7CF3l	links not explored

Summarizer

More Thoughts About AGI

Summary (from 50,000 words, truncated from 526,284)

Resources (265 fetched, 65 errors)

Not Explored (247)

Problems (81)

Raw Files

Execution Log

LLM Invocations (Total: $0.4024)