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 |
|---|---|---|---|---|
hn hn-ai | 1mo | Reverse-Engineering Jane Street's ASIC | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Tasteful Agent: Measuring and Improving Taste in Long-Horizon Tasks | LLM agents increasingly work on long-horizon tasks, and the decisions they make along the way, such as which hypothesis to test or whi… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Latest Exact Match Attention | We introduce latest exact match attention (LEMA), an attention variant for transformers where queries and keys are binarized and each… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Reply to comments arXiv:2512.07881 and arXiv:2601.06104 on quantum structure in human and AI-generated language | We reply to the comments by M. Sienicki and K. Sienicki (arXiv:2512.07881) and by K. Sienicki (arXiv:2601.06104) on our work on quantu… | ↗Not briefed |
hn hn-ai | 1mo | Physics Problem Finder | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Isn’t it also conceivable that people from other countries don’t actually want to work for/collaborate with OpenAI, in particular unpaid? (… | Isn’t it also conceivable that people from other countries don’t actually want to work for/collaborate with OpenAI, in particular unpa… | ↗Not briefed |
bluesky bluesky-ai | 1mo | “Staff at Anthropic, OpenAI and Google DeepMind have also publicly complained of similar issues, with many top researchers quitting due to … | “Staff at Anthropic, OpenAI and Google DeepMind have also publicly complained of similar issues, with many top researchers quitting du… | ↗Not briefed |
bluesky bluesky-ai | 1mo | “Staff at world-leading AI companies and institutions are increasingly suffering from the mental strain of working on systems they fear cou… | “Staff at world-leading AI companies and institutions are increasingly suffering from the mental strain of working on systems they fea… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evaluating Accuracy and Probabilistic Reliability of Zero-Shot Time Series Foundation Models | Time Series Foundation Models (TSFMs) promise a paradigm shift toward zero-shot forecasting by eliminating task-specific training. How… | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLMs sincerely don't interest me. Not because they're magic. I know precisely how they work, I'm competent w/ the math involved. I have low… | LLMs sincerely don't interest me. Not because they're magic. I know precisely how they work, I'm competent w/ the math involved. I hav… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Lightweight Plastic-Memory Framework for Graph Few-Shot Class-Incremental Learning | Graph Incremental Learning has garnered increasing attention as dynamic graph data continues to emerge across diverse fields. Conventi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | まあ現状の技術だとLLMでしかないからなんともね せいぜい墓に液晶モニターくっつけて定型文を繰り返させときゃそれでええんちゃうの | まあ現状の技術だとLLMでしかないからなんともね せいぜい墓に液晶モニターくっつけて定型文を繰り返させときゃそれでええんちゃうの | ↗Not briefed |
arxiv arxiv-ai | 1mo | Statistical Gains from Looped Estimation under Parameter Budgets | Growing memory demands in artificial intelligence motivate learning with fewer trainable parameters. We ask whether a looped estimator… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Disentangling Heterogeneous Traffic Dynamics for Multi-Step Traffic Forecasting via Adaptive Spectral Decomposition | Accurate multi-step traffic forecasting remains challenging because observed traffic signals contain heterogeneous temporal dynamics w… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Graded Representation Theory of Equivariant Neural Networks | Nonlinear activations can create equivariant interactions between irreducible representations that linear maps cannot. We use the Gaus… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Video-HopChain: Multi-Hop Questions and Confidence-Gated Exploration for Video Reasoning Models | HopChain has shown on still images that multi-hop data synthesis improves vision-language reasoning, because long chain-of-thought rea… | ↗Not briefed |
bluesky bluesky-ai | 1mo | “This mass assault could have been prevented had OpenAI warned [law enforcement] or the school,” the lawsuit said. It added the company was…🔧 ChatGPT | “This mass assault could have been prevented had OpenAI warned [law enforcement] or the school,” the lawsuit said. It added the compan… | ↗Not briefed |
bluesky bluesky-ai | 1mo | “The review team concluded that there was a “credible specific risk of harm to others”. OpenAI’s reviewers recommended a referral to the p… | “The review team concluded that there was a “credible specific risk of harm to others”. OpenAI’s reviewers recommended a referral to… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "O'Mega Jazz Remix" (Official Music Video) We've run the Psy-Lounge custom AI model to generate a sax-filled remix of our hybrid public do… | "O'Mega Jazz Remix" (Official Music Video) We've run the Psy-Lounge custom AI model to generate a sax-filled remix of our hybrid publ… | ↗Not briefed |
bluesky bluesky-ai | 1mo | What would actually help? Mandatory safety testing before deployment. Compulsory incident reporting. International coordination on model… | What would actually help? Mandatory safety testing before deployment. Compulsory incident reporting. International coordination on… | ↗Not briefed |
bluesky bluesky-ai | 1mo | There's an applied epistemology paper to write here arguing that AI is really good at identifying unidentified routes from collective doxas… | There's an applied epistemology paper to write here arguing that AI is really good at identifying unidentified routes from collective… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Towards Omni-dimensional GUI Agent Navigation with Masked Trajectory Prediction | Graphical User Interface (GUI) Agents autonomously interact with software to fulfill user requests, where GUI navigation stands out as… | ↗Not briefed |
hn hn-ai | 1mo | Move On – export your AI chats, file GDPR deletion, escalate if ignored | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI gadgets have a privacy problem Tech giants such as Meta and OpenAI are racing to build AI devices that can see and hear the world. But … | AI gadgets have a privacy problem Tech giants such as Meta and OpenAI are racing to build AI devices that can see and hear the world.… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Neurosymbolic Action Model Learning under Partial Observability | AI planning studies how an agent can reach a goal by executing a sequence of actions. To plan correctly, the agent needs an action mod… | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: Is Astra's writing starting to bother anyone else? | As someone who pretty much only uses GPT models (due to how generous the limits are), I always felt like I’d dodged a bullet when peop… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I don't personally mind that generative IA would settle open conjectures en masse (if it did not destroy the biosphere in the process, say)… | I don't personally mind that generative IA would settle open conjectures en masse (if it did not destroy the biosphere in the process,… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Limits of Simulated Societies: How Post-Training and Survey Fine-Tuning Erase Cross-Cultural Variance🔧 Consensus | Using large language models (LLMs) to simulate diverse human populations has the potential to transform many aspects of computational… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Minimal Recurrent Behavioral Memory for Imitation under Partial Observability | What is the least recurrent memory needed to reproduce a specified expert under partial observability? The instantaneous requirement i… | ↗Not briefed |
hn hn-ai | 1mo | Where Improvement Becomes the Problem | — | ↗Not briefed |
hn hn-ai | 1mo | Bk2204/talk-rust-in-Git: GitHub repository with SHA-256 commit IDs | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Syndrome, Synergy, and Safety: Structured Reasoning and Knowledge-Driven Alignment for TCM Prescription Generation🔧 GPT-5 | Applying large language models to Traditional Chinese Medicine (TCM) prescription generation reveals three clinically critical gaps: m… | ↗Not briefed |
bluesky bluesky-ai | 1mo | British Columbia sues OpenAI and Sam Altman over Tumbler Ridge mass school shooting | British Columbia sues OpenAI and Sam Altman over Tumbler Ridge mass school shooting | ↗Not briefed |
bluesky bluesky-ai | 1mo | Staff at world-leading AI companies and institutions are increasingly suffering from the mental strain of working on systems they fear coul… | Staff at world-leading AI companies and institutions are increasingly suffering from the mental strain of working on systems they fear… | ↗Not briefed |
hn hn-ai | 1mo | A Summer of AI Optimization | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | #girl #girls #model #bikini #teen #aigirl #aigirls #girlmodel #aiart #ai #feet #aifeet #feetlover | #girl #girls #model #bikini #teen #aigirl #aigirls #girlmodel #aiart #ai #feet #aifeet #feetlover | ↗Not briefed |
bluesky bluesky-ai | 1mo | My own husband, after listen to me prattle on for about 20 minutes about the history of datura, took a drag of his Marlboro and said “it lo… | My own husband, after listen to me prattle on for about 20 minutes about the history of datura, took a drag of his Marlboro and said “… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Modular Norm RandOpt: Population-Efficient Ensembling through Architecture-Aware Perturbations🔧 Gemma 3 | RandOpt samples weight-perturbed language models and ensembles top-ranked candidates through plurality voting, but its global perturba… | ↗Not briefed |
rss latent-space | 1mo | [AINews] Xiaomi MiMo-V2.6-Pro 1T-A42B: the new top Open Weights model, trained for $3M | crowning a new Chinese frontier lab | ↗Not briefed |
bluesky bluesky-ai | 1mo | It’s part of a wider phenomenon across leading AI companies: staff at Anthropic, OpenAI and Google DeepMind have also publicly complained o… | It’s part of a wider phenomenon across leading AI companies: staff at Anthropic, OpenAI and Google DeepMind have also publicly complai… | ↗Not briefed |
bluesky bluesky-ai | 1mo | EXC: AI staff complain of mental toll over fears of threat of society Multiple staff at Britain’s AI Security Institute (AISI) have been … | EXC: AI staff complain of mental toll over fears of threat of society Multiple staff at Britain’s AI Security Institute (AISI) have… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI refused to share Tumbler Ridge shooter’s chat history, B.C. attorney general says while filing California lawsuit | OpenAI refused to share Tumbler Ridge shooter’s chat history, B.C. attorney general says while filing California lawsuit | ↗Not briefed |
bluesky bluesky-ai | 1mo | You would not BELIEVE how badly a certain type of person who has made LLM enthusiasm their entire personality needs everyone they encounter… | You would not BELIEVE how badly a certain type of person who has made LLM enthusiasm their entire personality needs everyone they enco… | ↗Not briefed |
bluesky bluesky-ai | 1mo | What I do for $dayjob pretty often: "So we have this thing that uses [OpenAI usually, sometimes Vertex, etc] and really all it does is for… | What I do for $dayjob pretty often: "So we have this thing that uses [OpenAI usually, sometimes Vertex, etc] and really all it does i… | ↗Not briefed |
hn hn-ai | 1mo | GeoRisk Trace – An evidence-gated agent that abstains on unsupported risk paths | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I have decided to write a book on my experience with Claude, Gemini and Chat GPT. I have a working title: "Plausible Nonsense: The Role of… | I have decided to write a book on my experience with Claude, Gemini and Chat GPT. I have a working title: "Plausible Nonsense: The Ro… | ↗Not briefed |
rss techmeme | 1mo | Dutch nanoimprint lithography startup Morphotonics raised €40M from 3M Ventures, EIC Fund, and others, and plans to expand into data center optical components (TechCrunch) | TechCrunch: Dutch nanoimprint lithography startup Morphotonics raised €40M from 3M Ventures, EIC Fund, and others, and plans to expand… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It’s a bit funny that Harvey, a legal AI company that offered seat based pricing, saw its gross margin go from 50% to negative 50% as custo… | It’s a bit funny that Harvey, a legal AI company that offered seat based pricing, saw its gross margin go from 50% to negative 50% as… | ↗Not briefed |
arxiv arxiv-ai | 1mo | OmniFysics-Nano-V2 Technical Report: Understanding the Physical World Across Modalities🔧 Modal | Omni-modal models have expanded multimodal interaction across vision, audio, speech, and language. However, their training is predomin… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 何回もこんな話したくないんですが bioに「AI学習禁止」と書いて xfolioで画像や文章を公開しながら ChatGPTやGeminiやClaudeやCopilotやGrokやPerplexityを 触ったり名前を付けたりお礼を言ったりするのをやめていただきたいです 特に…🔧 ChatGPT | 何回もこんな話したくないんですが bioに「AI学習禁止」と書いて xfolioで画像や文章を公開しながら ChatGPTやGeminiやClaudeやCopilotやGrokやPerplexityを 触ったり名前を付けたりお礼を言ったりするのをやめていただきたいで… | ↗Not briefed |