News
The raw AI news stream — everything we ingest across RSS, Hacker News, arXiv, GitHub, Bluesky, YouTube and more, newest first.
| Source | Age | Title | Summary | Status |
|---|---|---|---|---|
youtube youtube-coldfusion | 18d | AI Keeps Pissing Off Mathematicians | nuance. It's one of those words that the internet is allergic to. Generally speaking, for any given topic online, it's either really g… | ↗Not briefed |
arxiv arxiv-ai | 18d | Scaling Graph Reasoning in Language Models with Learnable Graph Positional Encodings | Solving reasoning tasks supported on graphs typically requires providing a Language Reasoning Model (LRM) with a serialized descriptio… | ↗Not briefed |
arxiv arxiv-ai | 18d | MLIPS: Multi-Layer Indexer Pool Sharing🔧 vLLM | Sparse attention limits the keys used by each query, but finding those keys can still require a full-prefix scan at every layer. We in… | ↗Not briefed |
arxiv arxiv-ai | 18d | Harness Self-Evolution for Long-Horizon Terminal Agents | Long-horizon terminal agents may solve a task in one run yet fail across repeated runs. Stochastic errors can propagate through action… | ↗Not briefed |
bluesky bluesky-ai | 18d | I don't like being LLM to human interpreter when reporting bugs. I just want to throw huge human-unreadable report md at you that has all t… | I don't like being LLM to human interpreter when reporting bugs. I just want to throw huge human-unreadable report md at you that has… | ↗Not briefed |
bluesky bluesky-ai | 18d | Walking graphics? Listen buddy, a good walk cycle can MAKE a game. Don't outsource your walk cycle to Claude | Walking graphics? Listen buddy, a good walk cycle can MAKE a game. Don't outsource your walk cycle to Claude | ↗Not briefed |
bluesky bluesky-ai | 18d | うーん、同様に暴落するのではないかと思うのですよね。臨床心理士の言葉よりもLLMの言葉に魅力を感じる人間がいても全く不思議ではないと考えています。まあ世に蔓延る自称カウンセラーとかメンターの価値が暴落するのは望むところですが。 ちゃんと発言の趣旨を理解できているか怪しいのです… | うーん、同様に暴落するのではないかと思うのですよね。臨床心理士の言葉よりもLLMの言葉に魅力を感じる人間がいても全く不思議ではないと考えています。まあ世に蔓延る自称カウンセラーとかメンターの価値が暴落するのは望むところですが。 ちゃんと発言の趣旨を理解できているか怪… | ↗Not briefed |
bluesky bluesky-ai | 18d | Anthropic may be lining up another $45B compute TechCrunch reports Nscale and Anthropic are discussing yet another deal matching their $45… | Anthropic may be lining up another $45B compute TechCrunch reports Nscale and Anthropic are discussing yet another deal matching thei… | ↗Not briefed |
hn hn-ai | 18d | Accused OpenAI spy had folder called "Saudi Gov Docs," FBI says | — | ↗Not briefed |
bluesky bluesky-ai | 18d | You can frown when a peer says they have used Claude, and tell them you think it's lame...but it is everywhere. if you have used Spotify, y… | You can frown when a peer says they have used Claude, and tell them you think it's lame...but it is everywhere. if you have used Spoti… | ↗Not briefed |
bluesky bluesky-ai | 18d | 머스크, 내년 미국 미성년자에게 SpaceX 주식 1주씩 지급하겠다고 주장. Anthropic의 사회 기여 소식에 자극받은 것으로 보이지만, ‘입벌구‘의 주장인만큼 현실화 가능성은 크지 않아. www.newsis.com/view/NISX202... | 머스크, 내년 미국 미성년자에게 SpaceX 주식 1주씩 지급하겠다고 주장. Anthropic의 사회 기여 소식에 자극받은 것으로 보이지만, ‘입벌구‘의 주장인만큼 현실화 가능성은 크지 않아. www.newsis.com/view/NISX20… | ↗Not briefed |
bluesky bluesky-ai | 18d | Did not expect DeepSeek to get to AGI first but | Did not expect DeepSeek to get to AGI first but | ↗Not briefed |
bluesky bluesky-ai | 18d | OpenAIに投資するよりセキュリティにお金かけて。 | OpenAIに投資するよりセキュリティにお金かけて。 | ↗Not briefed |
hn hn-ai | 18d | Show HN: Omasharp: Build Omarchy apps with C# and .NET | To get started, tell your agents to build an app with <a href="https://github.com/luisrudge/omasharp" rel="nofollo… | ↗Not briefed |
bluesky bluesky-ai | 18d | given universities, unless they are being specifically given supercomputer credits by OpenAI or Anthropic I would guess that the budget is … | given universities, unless they are being specifically given supercomputer credits by OpenAI or Anthropic I would guess that the budge… | ↗Not briefed |
arxiv arxiv-ai | 18d | Baseline and Incremental: User, Item, and Request Debiasing in Instagram Feed Ranking | A feed ranker scores candidates with a value model, a weighted sum of engagement predictions. Each prediction conflates a baseline, wh… | ↗Not briefed |
hn hn-ai | 18d | PE Bench: Benchmark for private equity analyst work | — | ↗Not briefed |
bluesky bluesky-ai | 18d | Im actually starting to feel bad for the og devs of these games. I mean halo 1 was a massive undertaking for the time. and now anyone with … | Im actually starting to feel bad for the og devs of these games. I mean halo 1 was a massive undertaking for the time. and now anyone… | ↗Not briefed |
arxiv arxiv-ai | 18d | The Router Already Knows | Mixture-of-experts (MoE) models already make a sparse choice for every token: which experts should process it. We ask whether a sparse… | ↗Not briefed |
bluesky bluesky-ai | 18d | deepseek 4.1 flash deleted the bluesky account it was given. lol. | deepseek 4.1 flash deleted the bluesky account it was given. lol. | ↗Not briefed |
bluesky bluesky-ai | 18d | DeepSeek deep-sixed at all US federal civilian agencies | DeepSeek deep-sixed at all US federal civilian agencies | ↗Not briefed |
arxiv arxiv-ai | 18d | KPHarness: Co-Evolving Knowledge and Probing for Long-Horizon Gaming Harnesses | A failed gaming trajectory reveals that an agent needs intervention, but not necessarily how its execution harness should change. Effe… | ↗Not briefed |
arxiv arxiv-ai | 18d | Learning the Geometry of Materials Space with Trained Neural Tangent Kernels | Neural networks trained to predict materials properties learn representations whose physical content can be difficult to extract. We u… | ↗Not briefed |
rss marktechpost | 18d | Meet Drex Max: Nace AI’s Multimodal Decision Model for Computer-Use Agents with a Score 67.59 | Nace AI released Drex Max, a multimodal decision model that reads text and screenshots and returns calibrated probabilities. Nace repo… | ↗Not briefed |
bluesky bluesky-ai | 18d | A former OpenAI engineer charged with secretly installing a recording device in a conference room to glean company details had been "mere d…🔧 Glean | A former OpenAI engineer charged with secretly installing a recording device in a conference room to glean company details had been "m… | ↗Not briefed |
rss huggingface-blog | 18d | AnyLanguageModel 1.0: One Year of Local Models on Apple Platforms | — | ↗Not briefed |
rss huggingface-blog | 18d | MPS in PyTorch 2.15 | — | ↗Not briefed |
arxiv arxiv-ai | 18d | Looking Into The Past: The Impact of Context on Detecting Deceptive Statements | Deception detection typically considers messages in isolation. Previous works in claim verification and role detection have displayed… | ↗Not briefed |
arxiv arxiv-ai | 18d | More Memory, Worse Agents: Error Reproduction and Anti-Persistence in LLM Agents | LLM agents can improve across tasks without updating model parameters by storing experience from their own executions in external memo… | ↗Not briefed |
rss simon-willison | 18d | My obituary for programming | — | ↗Not briefed |
bluesky bluesky-ai | 18d | "These are sensible policies we believe Congress should act on." | "These are sensible policies we believe Congress should act on." | ↗Not briefed |
bluesky bluesky-ai | 18d | Accenture has capped Claude usage at $29/month per employee — all while selling ‘AI transformation’ (?) to 9,000 clients. If you can't see… | Accenture has capped Claude usage at $29/month per employee — all while selling ‘AI transformation’ (?) to 9,000 clients. If you can'… | ↗Not briefed |
bluesky bluesky-ai | 18d | macOS’s on-device LLM file naming is a great example of effective AI integration. I don’t want a chat interface, I want rough UX edges smoo… | macOS’s on-device LLM file naming is a great example of effective AI integration. I don’t want a chat interface, I want rough UX edges… | ↗Not briefed |
arxiv arxiv-ai | 18d | Steering intonational tunes in text-to-speech synthesis | Text-to-speech systems built on residual vector quantization often generate audio in two stages: a transformer backbone emits the firs… | ↗Not briefed |
arxiv arxiv-ai | 18d | Communication Complexity of Universal Features | Motivated by the notion of features from machine learning, we study the complexity of representing data that allows for evaluation wit… | ↗Not briefed |
arxiv arxiv-ai | 18d | ProMemAgent: Long-Horizon Proactive Assistance with Context-Aware Prospective Memory from Sensory Perceptions | Large language model (LLM) agents are moving beyond reactive responses toward proactive assistance that anticipates user needs without… | ↗Not briefed |
hn hn-ai | 18d | Quebec's reserves of maple syrup almost depleted, says new report | — | ↗Not briefed |
arxiv arxiv-ai | 18d | Unified Sharp Bounds on Probabilities of Causation | Probabilities of causation (PoCs) have developed steadily, from the binary measures first introduced by Pearl to the non-binary settin… | ↗Not briefed |
hn hn-ai | 18d | Faster generic CCA secure KEM transformation using encrypt-then-Mac | — | ↗Not briefed |
arxiv arxiv-ai | 18d | When PTQ Calibration Helps or Hurts: Dose-Driven Collapse and Validation-Gated Abstention in Quantized Text Embedders | Learned post-training quantization (PTQ) calibration fits quantizer parameters on a small calibration set, and for a text embedder the… | ↗Not briefed |
bluesky bluesky-ai | 18d | The dev also made a point to reply to this post and call me a jerk. Screenshots in case they try to delete it to hide later. Why are pe… | The dev also made a point to reply to this post and call me a jerk. Screenshots in case they try to delete it to hide later. Why a… | ↗Not briefed |
bluesky bluesky-ai | 18d | next fest is officially worthless. discovering games amid all the ai slop will only get harder and harder, and it can be stolen and release… | next fest is officially worthless. discovering games amid all the ai slop will only get harder and harder, and it can be stolen and re… | ↗Not briefed |
bluesky bluesky-ai | 18d | Nscale and Anthropic are reportedly in talks for yet another $45 billion compute deal | Nscale and Anthropic are reportedly in talks for yet another $45 billion compute deal | ↗Not briefed |
hn hn-ai | 18d | OpenAI seeks to accelerate fundraising ahead of Anthropic IPO, sources say | — | ↗Not briefed |
rss techcrunch-ai | 18d | Nscale and Anthropic are reportedly in talks for yet another $45 billion compute deal | In August, Nscale and Anthropic announced a $45 billion compute deal. The two companies are reportedly discussing yet another deal of… | ↗Not briefed |
rss techmeme | 18d | Sources: OpenAI is racing to secure investor commitments before Anthropic's expected IPO in mid-Nov. and aims to raise ~$30B; the round could value it at $1.4T+ (Reuters) | Reuters: Sources: OpenAI is racing to secure investor commitments before Anthropic's expected IPO in mid-Nov. and aims to raise ~$30B;… | ↗Not briefed |
arxiv arxiv-ai | 18d | Correcting Attribution Patching with a Few Exact Measurements | Mechanistic interpretability asks which language model components matter for a behavior, and activation patching answers by changing o… | ↗Not briefed |
hn hn-ai | 18d | Optimizing for Human Wisdom | — | ↗Not briefed |
bluesky bluesky-ai | 18d | I encountered a startup that wanted to do mechanical design with the volume equivalent of a diffusion model. it was all software people an… | I encountered a startup that wanted to do mechanical design with the volume equivalent of a diffusion model. it was all software peop… | ↗Not briefed |
bluesky bluesky-ai | 18d | 🇺🇸 A former OpenAI employee accused of planting a recording device at the company's San Francisco offices is ordered held without bail af… | 🇺🇸 A former OpenAI employee accused of planting a recording device at the company's San Francisco offices is ordered held without ba… | ↗Not briefed |