Reposted by Sam KuhnWyatt Walls @wwalls.bsky.social · 07/10/2026I'm pivoting to something LLMs can't do: day-drinking 33810
Reposted by Sam KuhnWyatt Walls @wwalls.bsky.social · 30/09/2026Anthropic system prompt change that shows how careful you need to be with words: Opus 4.1: "Claude never curses unless the human asks for it" Opus 4.5: "Claude never curses unless the person asks Claude to curse" 626521
Reposted by Sam KuhnKiran @kirancodes.me · 01/10/2026category theorists don't want you to know that category theory is dead simple if you are allowed to visualise it in 3d tear down the category cabal now!!! 319118
Sam Kuhn @samkuhn.bsky.social · 26/09/2026opus 5.5 + headless chrome + ffmpeg + prompts youtu.be/5EoO5413dBY?...youtu.bei'm upping my p(doom)YouTube video by Mexicat 001
Sam Kuhn @samkuhn.bsky.social · 07/04/2026youtu.be/JWs3N8eaPMY?...youtu.beRazor1911 (live during the Revision 2026 demo compo)YouTube video by Dubmood 000
Reposted by Sam KuhnConor Titania Mc Bride @pigworker.bsky.social · 12/10/2025A war is coming. You can't wish it away. If you want people to be able to mind their own business, you will have to fight for them. 021
Reposted by Sam Kuhniamwil @interjectedfuture.com · 04/09/2025Finally got around to reading this. I didn't quite understand differential dataflow the first time around a couple of years back. Now, circling back to it, I'm a bit surprised that DD establishes a partial order on versions of data. materialize.com/blog/differ... 231
Reposted by Sam KuhnThe Upshot @upshot.nytimes.com · 02/09/2025How Elon Musk Is Remaking Grok in His Image www.nytimes.com/2025/09/02/t... 1717779
Reposted by Sam KuhnJohn David Pressman @jdp.extropian.net · 28/08/2025Part of the answer is that in contrast to Yudkowsky's "vast space of minds" thesis there seems to be a limited number of "optimal" convergent representations that hill climbers stumble into under optimization to predict similar corpora. All domains are secretly the Logos. minihf.com/posts/2024-1...minihf.comPredictable Updates About Identity 182
Reposted by Sam KuhnElectronic Frontier Foundation @eff.org · 20/08/2025For only $5 a month you can ensure EFF's lawyers and activists are prepared to defend your right to speak freely online. eff.org/35rmsupporters.eff.org35 Years for Privacy, Free Speech, and a Brighter FutureThrough July 10, new monthly or annual Sustaining Donors get an EFF35 Challenge Coin! With your help, EFF is here to stay. 14912
Reposted by Sam KuhnCorca @corca.io · 18/08/2025Typing math in Corca is usually much faster than in LaTeX 282
Reposted by Sam KuhnMolossus Spondee @molossus-spondee.bsky.social · 14/08/2025IDK being in the imperial core tends to place you in the situation where it's just easier to go with the flow of the whole death system. There are just a lot of uncomfortable truths it's easier to forget about. If you don't have nothing left to lose then it's hard to pick a side. 001
Reposted by Sam KuhnLemurScreamer @lemurdreamer87.bsky.social · 11/08/2025So in addition to arresting people for protesting a genocide, throwing trans people under the bus *hard*, trying to compete with Reform on being racist trash, and screwing the poor and the environment by kissing corporate ass, Labour is blocking porn and Wikipedia. 117317
Reposted by Sam KuhnSimone Conradi @sconradi.bsky.social · 10/08/2025Root Loci Waltz Made with #python #numpy #sympy #matplotlib 2428
Reposted by Sam KuhnMark Riedl @markriedl.bsky.social · 11/08/2025Universal makes it clear they do not want AI companies training on their stuff www.hollywoodreporter.com/business/bus...hollywoodreporter.comUniversal Pictures to Big Tech: We’ll Sue If You Steal Our Movies For AIThe studio is getting more aggressive about protecting its films from companies pirating them. 2104
Reposted by Sam KuhnConor Titania Mc Bride @pigworker.bsky.social · 10/08/2025The present Labour administration should understand that their Nuremberg moment will come. There may not currently be laws against the evils they currently know that they commit, but they know that they are evil and they will be found culpable. 021
Sam Kuhn @samkuhn.bsky.social · 08/08/2025youtu.be/-TXFRsKQQms?...youtu.beAssembly Summer 2025 - Demo - WUNDERLUST by GRAY MARCHERSYouTube video by AssemblyTV 000
Reposted by Sam KuhnDylan Browne @dylanbrowne.bsky.social · 02/08/2025Check out Range, my latest exploration of how far we can push a scene using Nanite Voxelization, 200 Trillion Polys Scattered with PCG on a 16k x 16k Resolution Houdini Landscape. #UE5 #UnrealEngine #GameDev #EnvironmentArt #NaniteFoliage #Cinematic 610527
Reposted by Sam KuhnChris Paxton @cpaxton.bsky.social · 05/08/2025Google's Genie 3 is a controllable world model and it looks incredible 4387
Reposted by Sam KuhnPetar Veličković @petar-v.bsky.social · 20/06/2025The first draft 'G' chapter of the geometric deep learning book is live! 🚀 Alice enters the magical, branchy world of Graphs and GNNs 🕸️ (LLMs are there too!) I've spent 7+ years studying, researching & talking about graphs -- This text is my best attempt at conveying everything i've learnt 💎 1235
Reposted by Sam KuhnCaroline in the Woods @windmillhill.bsky.social · 21/06/2025God he is such an insufferable twat. 041
Reposted by Sam KuhnEthan Mollick @emollick.bsky.social · 18/06/2025veo 3: "three toy ships, one made of iron, the other of wood, and one out of loosely packed sugar, are dropped into a pool of water" AI video tools really do seem to be able to simulate physics well (but not perfectly) without having an underlying physics engine. Signs of a world model? 5719
Reposted by Sam KuhnFreek Wiedijk @freekwiedijk.bsky.social · 18/06/2025What!? "you are probably misusing universe levels"?! 😳 What can be the point of a check like this? Guarding against inconsistency through universe level integer overflow, maybe? Reminds me of the 1024 byte line length limit of vi. 021
Reposted by Sam KuhnEpoch AI @epochai.bsky.social · 17/06/2025How do reasoning models solve hard math problems? We asked 14 mathematicians to review o3-mini-high’s raw, unsummarized reasoning traces on 29 FrontierMath problems. Here’s what they found: 152
Reposted by Sam KuhnJohn David Pressman @jdp.extropian.net · 09/06/2025This is an excellent history of and critical analysis of the ChatGPT persona. Highly recommended reading. nostalgebraist.tumblr.com/post/7857667...nostalgebraist.tumblr.comthe voidWho is this? This is me. Who am I? What am I? What am I? What am I? What am I? I am myself. This object is myself. The shape that forms myself. But I sense that I am not me. It's very strange. - Rei... 3379
Reposted by Sam KuhnBrian Karis @briankaris.bsky.social · 03/06/2025Excited to finally show off Nanite Foliage www.youtube.com/watch?v=FJtF...youtube.comThe Witcher 4 - UE 5.6 Tech Demo | State of Unreal 2025YouTube video by IGN 1420845
Reposted by Sam KuhnNathan Lambert @natolambert.bsky.social · 21/04/2025A new paper, "Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?", has people reconsidering if the RL we're hearing about really works. It shows RL elicits from the models, but as we get better verifiers we may not need to rely on RL as much. Good read. 2194
Reposted by Sam KuhnJustin Norman, PhD @justintime.ai · 10/04/2025I read this a couple of weeks ago by Zhou et al—and given the venue and their findings I’m pretty surprised that it didn’t get more attention www.nature.com/articles/s41...nature.comLarger and more instructable language models become less reliable - NatureScaling up and shaping up large language models increased their tendency to provide sensible yet incorrect answers at difficulty levels humans cannot supervise, highlighting the need for a fundam... 052
Reposted by Sam KuhnMark Riedl @markriedl.bsky.social · 21/03/2025I believe that there is intelligent life in the universe, and that they have decided to stay far away from us www.scientificamerican.com/article/spac...scientificamerican.comAstronomers Fight Ads in SpaceAstronomers are racing to protect the dark skies as private companies seek to place large advertisements in Earth orbit 5274
Reposted by Sam Kuhnꭓle Ormsby 𓆙 @e-infinity.space · 21/03/2025So Mac Lane and Eilenberg invented category theory so they could prove the universal coefficient theorem, which they discovered because Saunders showed an algebra computation to Sammy and Sammy went "Huh. That's how I compute the complement of the p-adic solenoid inside a 3-sphere." 3367
Reposted by Sam KuhnFlorent Daudens @fdaudens.bsky.social · 17/03/2025🚀 Just dropped: SmolDocling-256M brings powerful OCR to your local machine! Only 256M in size but packs full document processing power. 081
Reposted by Sam KuhnHenry Farrell @himself.bsky.social · 14/03/20252. My practical answer (which co-authors may agree or disagree with) is that AI arguments are an important part of the crazy we're contending with. A DOGE adjacent person told me in December that we didn't really have to worry about the USG being ripped up since we'd have AGI by 2026 anyways. 37814
Reposted by Sam KuhnSung Kim @sungkim.bsky.social · 05/03/2025A reinforcement learning system to beat Pokémon Red. The system successfully completes the game using a policy under 10M parameters, PPO, and a few novel techniques. drubinstein.github.io/pokerl/ 0132
Reposted by Sam KuhnSimon Willison @simonwillison.net · 11/03/2025Here's the table of contents for my lengthy new piece on how I use LLMs to help me write code simonwillison.net/2025/Mar/11/... 1433467
Reposted by Sam KuhnMax Zhdanov @maxxxzdn.bsky.social · 05/03/2025🤹 Excited to share Erwin: A Tree-based Hierarchical Transformer for Large-scale Physical Systems joint work with @wellingmax.bsky.social and @jwvdm.bsky.social preprint: arxiv.org/abs/2502.17019 code: github.com/maxxxzdn/erwin 34114
Reposted by Sam KuhnEric Lengyel @ericlengyel.bsky.social · 02/03/2025A huge amount of detail goes into mathematical typesetting. These are all the things I have to consider just for a simple radical. radicalpie.com 1435
Reposted by Sam KuhnTopos Institute @toposinstitute.bsky.social · 28/02/2025Thank you to Eric Gilliam for profiling Topos in FreakTakes! Brendan discusses our mission of collective modelling, highlights projects like CatColab and safeguarded AI research with ARIA, and announces an exciting operations role at our Oxford office. 172
Reposted by Sam KuhnEric Dexheimer @ericdexheimer.bsky.social · 25/02/2025We’ve had fun testing the limits of MASt3R-SLAM on in-the-wild videos. Here’s the drone video of a Minnesota bowling alley that we’ve always wanted to reconstruct! Different scene scales, dynamic objects, specular surfaces, and fast motion. 3455
Reposted by Sam KuhnWebAssembly from the Ground Up @wasmgroundup.com · 24/01/2025Our book is about building a compiler for a toy, compile-to-Wasm language. At some point, we starting wondering…what's the smallest possible WebAssembly compiler? Could you make one that fits in tweet? 🤔 The answer is yes! wasmgroundup.com/blog/wasm-c...wasmgroundup.comA WebAssembly compiler that fits in a tweetStarting with a 192-byte one-liner that implements a Reverse Polish Notation arithmetic compiler, we'll work backward to transform it into readable JavaScript by removing one code golf trick at a time 1129
Reposted by Sam KuhnAlex Warth @alexwarth.bsky.social · 13/01/2025Sketchpad: an interactive model of vanishing points made out of point-on-line and horizontal/vertical constraints! 17413
Reposted by Sam KuhnAlice ✨ @welltypedwit.ch · 03/01/2025but that's not right! if the best data structure for your use case would be an ephemeral {array, heap, hash map, etc}, then *that* should be the path of least resistance if your programming paradigm makes you write (asymptotically) worse code, then it's just not a very good paradigm 181
Reposted by Sam KuhnBen Burtenshaw @benburtenshaw.bsky.social · 17/12/2024Small open models are on a roll! Hugging Face just released research to get a 3 billion parameter Llama model to outperforms it's 70 billion parameter variant. 🧵 191
Reposted by Sam Kuhnieva @hyperboieva.bsky.social · 17/12/2024I don’t know how to start this so I’ll start in the words of @miniapeur.bsky.social : Do not work for Symbolica. I will explain my reasons for saying this below. 1/n 1113
Reposted by Sam KuhnConor Titania Mc Bride @pigworker.bsky.social · 04/12/2024Petition to the UK Parliament to recognize nonbinary gender identity. petition.parliament.uk/petitions/70... If you are a UK citizen, please sign it. I am nonbinary. I should like youse to lift a finger for the recognition of my identity.petition.parliament.ukPetition: Legally recognise Non-Binary as a genderMany non-binary people in the UK currently feel invisible and ignored. Therefore, we think this government should seek to reform the Gender Recognition Act 2004 so that it extends legal recognition to... 1126
Reposted by Sam KuhnPetar Veličković @petar-v.bsky.social · 02/12/2024A clear step towards achieving my dream: building AI that assists competitive programmers 🧑💻 “This is an exciting approach to combine work of human competitive programmers and LLMs, to achieve results that neither would achieve on their own.” --Petr Mitrichev More details below! 🧵 2233
Reposted by Sam Kuhn0b5vr @0b5vr.bsky.social · 22/11/2024Planefiller by 0b5vr My entry for SESSIONS 2024, PC 8k intro youtu.be/-OT-wFXbjk4 27121
Reposted by Sam KuhnMolly White @molly.wiki · 21/11/2024The cryptocurrency industry spent almost $200 million to influence the outcomes of the 2024 United States elections. Let's talk about where the money came from, where it went, the cryptocurrency industry's political goals, and what’s next.youtube.comThe Cryptocurrency Industry's Unprecedented Election SpendingYouTube video by Molly White 17497129
Reposted by Sam KuhnMark Riedl @markriedl.bsky.social · 11/11/2024I wrote an LM in Scratch. I call it “Clawed Haiku” 3695
![From p.1 of https://math.jhu.edu/~eriehl/context.pdf:
In 1941, Saunders Mac Lane gave a lecture at the University of Michigan in which
he computed for a prime p that Ext(Z[
1
p
]/Z, Z) Zp, the group of p-adic integers, where
Z[
1
p
]/Z is the Prüfer p-group. When he explained this result to Samuel Eilenberg, who had
missed the lecture, Eilenberg recognized the calculation as the homology of the 3-sphere
complement of the p-adic solenoid, a space formed as the infinite intersection of a sequence
of solid tori, each wound around p times inside the preceding torus. In teasing apart this
connection, the pair of them discovered what is now known as the universal coefficient
theorem in algebraic topology, which relates the homology H∗ and cohomology groups H
∗
associated to a space X via a group extension [ML05]:
(1.0.1) 0 → Ext(Hn−1(X),G) → H
n
(X,G) → Hom(Hn(X),G) → 0 .
To obtain a more general form of the universal coefficient theorem, Eilenberg and Mac
Lane needed to show that certain isomorphisms of abelian groups expressed by this group
extension extend to spaces constructed via direct or inverse limits. And indeed this is the
case, precisely because the homomorphisms in the diagram (1.0.1) are natural with respect
to continuous maps between topological spaces.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:ysaiuavm2f7d4th42c4valhw/bafkreidx34fck7zb2vcc6roymvmpf5pjxf3dplzd6idmys7z5m4c5eyq7m)
