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 |
|---|---|---|---|---|
arxiv arxiv-ai | 1mo | Geometric Moment Contraction for Stochastic Nesterov Acceleration | We study geometric moment contraction (GMC) of the constant-parameter stochastic Nesterov recursion \[ Y_k=Θ_k+β(Θ_k-Θ_{k-1}),\qquad Θ… | ↗Not briefed |
hn hn-ai | 1mo | Entropy-Based Guided Collaboration in Heterogeneous LLM Multi-Agent Systems | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yeah! can’t agree more An LLM to me is some combination of: - video game character - pet - search engine All of which I already anthrop… | Yeah! can’t agree more An LLM to me is some combination of: - video game character - pet - search engine All of which I already an… | ↗Not briefed |
hn hn-ai | 1mo | Tell HN: AGI Is Here | Whether by your strict definition or not AGI is here is immaterial. For 90%+ of humanity AGI is here <i>today</i>, they act like it do… | ↗Not briefed |
bluesky bluesky-ai | 1mo | openai.com/index/openai... | openai.com/index/openai... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond Approved Actions: Runtime Validation of Persistent Outcomes in Agent Workflows | Large language model agents increasingly act on software systems, no longer merely generating text but also changing databases and onl… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Buried further down: “Outside experts said that such claims jumped far ahead of the evidence the team has gathered so far." “But outside … | Buried further down: “Outside experts said that such claims jumped far ahead of the evidence the team has gathered so far." “But out… | ↗Not briefed |
arxiv arxiv-ai | 1mo | UniAR: A Unified Framework for Autism Recognition Enhanced by Multi-View Prompt Learning | Autism Spectrum Disorder (ASD) is a complex neurodevelopmental disorder for which early and accurate diagnosis is critical to improvin… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Lol Lmao even. www.404media.co/people-train... | Lol Lmao even. www.404media.co/people-train... | ↗Not briefed |
hn hn-ai | 1mo | U.S. Rebukes Australia over Proposed Limits on Social Media Algorithms | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Wayne has worked on some incredible films and series, including The Fifth Element. I don't know how he feels about the new art book using g… | Wayne has worked on some incredible films and series, including The Fifth Element. I don't know how he feels about the new art book us… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Not to mention the fact that the LLM-pushers are eugenicists while we work in special education and disability rehab, so any money we give … | Not to mention the fact that the LLM-pushers are eugenicists while we work in special education and disability rehab, so any money we… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 💪🏼 | 💪🏼 | ↗Not briefed |
arxiv arxiv-ai | 1mo | Why Alzheimer's Speech Screening Fails to Generalize: Bridging the Deployment Gap via Cross-Corpus Evidence Anchoring | Speech-based screening is a promising, non-invasive approach for detecting Alzheimer's disease and related cognitive risks. However, m… | ↗Not briefed |
bluesky bluesky-ai | 1mo | it’s hard to beat a good em dash. we can’t let the llm take that from us | it’s hard to beat a good em dash. we can’t let the llm take that from us | ↗Not briefed |
arxiv arxiv-ai | 1mo | Softmax Reparameterization for Output-Head Quantization | Large vocabularies make output heads a substantial inference cost in small language models. We propose softmax reparameterization, a p… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Trump’s administration requested OpenAI and Anthropic withhold new AI models from the UK’s flagship testing agency without prior review by … | Trump’s administration requested OpenAI and Anthropic withhold new AI models from the UK’s flagship testing agency without prior revie… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Unfortunately this "expert slop", where expert is a self-perceived title, carries a corporate framing that damages interactions. Others be… | "Unfortunately this "expert slop", where expert is a self-perceived title, carries a corporate framing that damages interactions. Othe… | ↗Not briefed |
rss techcrunch-ai | 1mo | Last 24 hours to save up to $200 on TechCrunch Disrupt 2026. Reason 5/5 to attend: Leave further ahead. | Last 24 hours to save up to $200 on your TechCrunch Disrupt 2026 pass. Leave the event further in your startup's trajectory than where… | ↗Not briefed |
hn hn-ai | 1mo | ReAnchor, a Jev Optimizer in DSPy | — | ↗Not briefed |
hn hn-ai | 1mo | A Jevlike using an escalation model to route between a classifier and an LLM | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | G2MAF: Test-Time Gradient Guidance for Multi-Agent Flow Policies | Offline multi-agent reinforcement learning (MARL) learns cooperative policies from fixed datasets without further environment interact… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Cannot believe this place is still getting softball coverage, or that people are paying 55k a year. | Cannot believe this place is still getting softball coverage, or that people are paying 55k a year. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Just gave a colleague an opinion blast on the LLM they’re subscribed to for writing IEP goals for us. | Just gave a colleague an opinion blast on the LLM they’re subscribed to for writing IEP goals for us. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Autorizando al LLM a que ordene la carpeta que se llama "Ordenar", que contiene 98346 archivos incluido otra carpeta que se llama "Ordenar"… | Autorizando al LLM a que ordene la carpeta que se llama "Ordenar", que contiene 98346 archivos incluido otra carpeta que se llama "Ord… | ↗Not briefed |
hn hn-ai | 1mo | The US stood alone in dismissing AI safety concerns at the UN | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Resource-Optimized and Energy-Aware Agentic AI Framework Anchored on Blockchain for Secure Software Supply Chains | This paper proposes a blockchain-backed agentic security framework designed to safeguard the complete software development lifecycle (… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I need to write this up but I think this is a really, really important point for understanding why openai and anthropic in particular are d… | I need to write this up but I think this is a really, really important point for understanding why openai and anthropic in particular… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The flip side of "LLM were a surprise" is that in 2010 natural language and text gen people were a side show of journalists and civic tech … | The flip side of "LLM were a surprise" is that in 2010 natural language and text gen people were a side show of journalists and civic… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MA-WAM: Multi-Agent World-Action Model for Test-Time Planning | Multi-agent cooperative tasks require different agents to execute a joint action simultaneously, and each agent's action affects both… | ↗Not briefed |
rss techmeme | 1mo | Sources: Tesla ramped up Optimus production to several hundred units per week but faces hurdles with its hands, automation equipment, and supplier constraints (The Information) | The Information: Sources: Tesla ramped up Optimus production to several hundred units per week but faces hurdles with its hands, autom… | ↗Not briefed |
hn hn-ai | 1mo | Benchmarking Jev, Laya, and five open models under three stresses | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Cognitive Skills in the Age of AI: Computing Students and Experts Perceptions | AI is becoming increasingly integrated into daily workflows, especially in computing. We are gradually shifting towards an AI-rich fut… | ↗Not briefed |
bluesky bluesky-ai | 1mo | and so a lot of what you've seen since is this weird race where openai and anthropic are not precisely trying to make their models _better_… | and so a lot of what you've seen since is this weird race where openai and anthropic are not precisely trying to make their models _be… | ↗Not briefed |
hn hn-ai | 1mo | The state of SIMD in Rust in 2026 | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | is it possible to start embedding viruses & killcodes into whatever we post online so that whenever an LLM inevitably crawls over it, it po… | is it possible to start embedding viruses & killcodes into whatever we post online so that whenever an LLM inevitably crawls over it,… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The problem isn't just that the article was generated by an LLM. It's that the article includes inaccuracies, that it was generated by stea… | The problem isn't just that the article was generated by an LLM. It's that the article includes inaccuracies, that it was generated by… | ↗Not briefed |
hn hn-ai | 1mo | Evaluating the Tax Character of Options-Based Income [pdf] | — | ↗Not briefed |
hn hn-ai | 1mo | A Model Context Protocol (MCP) server for NATS messaging system integration | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Sifthound – self-hosted, Tavily-compatible search API for AI agents | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | “The agents also tried on three occasions to hack public data providers, including an Australian government website. We link at least some … | “The agents also tried on three occasions to hack public data providers, including an Australian government website. We link at least… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think we finally found a situation where an LLM would be an improvement | I think we finally found a situation where an LLM would be an improvement | ↗Not briefed |
arxiv arxiv-ai | 1mo | Identifying Scientists on X | With the growing importance of science-related discourse on the Web and the erosion of the classical knowledge order, it is important… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MoSAR: Mixture of Semantic Attention Regimes for Learning Adaptive and Approximable Attention Geometries | The quadratic complexity of dense self-attention remains a central bottleneck for long-context language modeling. Many efficient alter… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Agentic Limit Order Books: Phase Transitions and Market Impact | We investigate the systemic macroscopic dynamics emerging from Limit Order Books (LOBs) populated exclusively by autonomous reinforcem… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Concrete Calculator – Estimate Cubic Yards, Bags and Cost | Free concrete calculator: enter length, width and thickness to get cubic yards, 80/60/40-lb bag counts and estimated cost —… | ↗Not briefed |
bluesky bluesky-ai | 1mo | yeah it's strange because it seems obvious to me that LLM safety will be a lot more like ecology and politics than engineering away the pap… | yeah it's strange because it seems obvious to me that LLM safety will be a lot more like ecology and politics than engineering away th… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I remember when they fielded an OpenAI DOTA 2 team that beat the pros (albeit with some limitations on the rules). It was exciting. LLMs ne… | I remember when they fielded an OpenAI DOTA 2 team that beat the pros (albeit with some limitations on the rules). It was exciting. LL… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PIA: A Personal Intelligence Agent Turning Health Conversations into Records and Records into Understanding | General-purpose agent memory summarizes conversations: it extracts salient snippets, embeds them, and retrieves the top-k into the pro… | ↗Not briefed |
bluesky bluesky-ai | 1mo | By now it's not an accident or incident anymore. OpenAI is just deliberately trying to get data that's locked away. Just like they stole a… | By now it's not an accident or incident anymore. OpenAI is just deliberately trying to get data that's locked away. Just like they st… | ↗Not briefed |