Every story across every category, newest first. Each card links to the original publisher; daily-brief posts open as editorial pages.
An interesting way to take Noam at his word in regards to always keeping a constant inference budget for any eval reporting -is that open models…
RT xjdrwe've been running GLM 5.2 in bf16 and in fp8 (experts and kvcache only, attention is always bf16) and have recorded virtually 0 measurable quality…
Article URL: https://ipcrawl.com/ Comments URL: https://news.ycombinator.com/item?id=48700834 Points: 21 # Comments: 10
Article URL: https://www.lloydslist.com/LL1157680/Ships-keep-moving-through-Hormuz-despite-strike-and-suspension-of-IMO-exit-strategy Comments URL: https://news.ycombinator.com/item?id=48700805 Points: 41 # Comments: 23
RT Alexander DoriaI mean just HF out, ModelScope blocked and it’s over. You won’t seed GLM.Naithan Jones: “THeY cAnT bAn OpEn sOuRcE bRO hOw cAn tHeY…
Repo link and results - https://github.com/Abhinand20/MathFormer Task: Given a factorized expression like (7-3*z)*(-5*z-9), predict the expanded form -> 15*z*2-8*z-63 Key takeaway: A tiny (4M param) seq2seq…
OpenAI announced GPT-5.6 but a terrible thing has happened
RT Matthew Zirwas, MDThe take that frontier models aren't ready for use in medicine is dead wrong.I've been using Opus 4.x in my clinical workflow every…
RT yung macro 宏观年少传奇So just to summarize, the official US foreign policy posture toward Europe is:1. Overall, it is an intention of this admin to punish…
Speculative decoding speeds up LLM generation by using a small "drafter" model to predict several tokens ahead of the main model. The main model then verifies…
Maraget Atwood, the storied author of The Handmaid's Tale and The Blind Assassin, was interviewed as part of the Babell Literary and Cultural Festival in Porto,…
RT alphaXivSakana Fugu Technical ReportInstead of training one larger model, Sakana AI trains an orchestrator that reads each query and dynamically routes or composes GPT-5.5, Gemini-3.1-Pro,…
RT alphaXivSakana Fugu Technical ReportInstead of training one larger model, Sakana AI trains an orchestrator that reads each query and dynamically routes or composes GPT-5.5, Gemini-3.1-Pro,…
RT Harrison KinsleyDario's args:"Opensource you can see the source, here you cannot see inside the model"- yes you can that's literally the open weights part btw.-…
Han barbarity. Children in Europe could have eaten that dog^W AC…🇨🇳XuZhenqing徐祯卿: ✨🇨🇳A Chinese content creator installed an air conditioner in his dog’s living space to keep…
This is how Freedom Wins.Petri Kuittinen: European Union joins US-led Pax Silica alliance, but at what cost? Will Europeans have AI or chip sovereignty or will…
Grok and DeepSeek engaging in an acausal values handshakethis was predicted for AGIs… Yud winningSiberian fox🔸: this says a lot about society
Article URL: https://www.wiseowl.com/articles/a2fpga-videx-01-the-card-that-made-the-apple-ii-serious/ Comments URL: https://news.ycombinator.com/item?id=48700317 Points: 4 # Comments: 0
when you (a EuroRICH) secretly scavenge and install a Chynese AC to not fucking die from stroke during a heat wave, and your neighbor has gotta…
RT Allie HoweTHE GREAT LOOPS DEBATE is on@ianlivingstone and @GeoffreyHuntley take on @dexhorthy and a surprise guest to answerDo loops produce slop?Do we have the AI…
«Gains come from models failing on different questions, not from adding more models.»Haven't read it yet but sounds right. Mixture of Models, where all models are…
Tried creating a quick 3d game with it, after 3 prompts, it got me this(checkvideo). If I compare this with qwen3.5-35b-a3b, it was not able to…
Too many times I hear people whine about not being ble to run SOTA models or claim it would require $50k, or $100k. https://www.ebay.com/itm/398079051468 Epcy Motherboard…
RT X FreezexAI models now support Zero Data Retention (ZDR) on OpenRouterDevelopers can now use:• Grok 4.3• Grok 4.20• Grok Build 0.1with Zero Data Retention enabledThis…