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 |
|---|---|---|---|---|
rss techmeme | 1mo | OpenAI says it is working with an independent advisory group of mathematicians to responsibly share math-related AI advances (OpenAI) | OpenAI: OpenAI says it is working with an independent advisory group of mathematicians to responsibly share math-related AI advances… | ↗Not briefed |
bluesky bluesky-ai | 1mo | neurodivergent girl who constantly jokes about how she had to learn the entirety of the social contract from base principles by watching th… | neurodivergent girl who constantly jokes about how she had to learn the entirety of the social contract from base principles by watchi… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MedGate-Fusion: Integrating First-Encounter Semantic Narratives and Physiological Biomarkers for Prospective Stroke Risk Stratification🔧 Modal | Prospective stroke risk stratification in primary care is challenging because early risk signals are distributed across routine biomar… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Source: before the Hugging Face incident, OpenAI was negotiating a legally binding deal with Anthropic for the companies to stress-test eac…🔧 Hugging Face | Source: before the Hugging Face incident, OpenAI was negotiating a legally binding deal with Anthropic for the companies to stress-tes… | ↗Not briefed |
rss techmeme | 1mo | Source: before the Hugging Face incident, OpenAI was negotiating a legally binding deal with Anthropic for the companies to stress-test each other's models (The Information)🔧 Hugging Face | The Information: Source: before the Hugging Face incident, OpenAI was negotiating a legally binding deal with Anthropic for the compan… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Four-Leaf MCP, open-source job search and interview prep | — | ↗Not briefed |
hn hn-ai | 1mo | Anthropic at $2T isn't far-fetched | — | ↗Not briefed |
hn hn-ai | 1mo | Physics in the Age of LLMs | — | ↗Not briefed |
hn hn-ai | 1mo | Hermes Agent now supports Claude Pro/Max subscriptions | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Now that I've figured out how Claude can write code for me I need AI to suggest things that I might actually need code for. | Now that I've figured out how Claude can write code for me I need AI to suggest things that I might actually need code for. | ↗Not briefed |
arxiv arxiv-ai | 1mo | The AI Neuroscientist: An Interactive Agentic Interface for Neuroimaging Analysis | Analyzing neuroimaging data requires specialized coding and statistical expertise, which limits accessibility for researchers without… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Combinators in Array Languages | Raymond Smullyan’s "To Mock a Mockingbird" book's aviary of combinator birds implemented in APL-derived languages and t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Y ahora con lo de Anthropic pasará lo mismo. Si no hay contratiempos externos (como Trump), saldrá a bolsa hacia finales de este año con un… | Y ahora con lo de Anthropic pasará lo mismo. Si no hay contratiempos externos (como Trump), saldrá a bolsa hacia finales de este año c… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Geometric and Semantic Coupling for Interaction Understanding in 3D Scenes | Interaction understanding in 3D scenes requires a joint description of movable parts, their motion, and the regions through which they… | ↗Not briefed |
arxiv arxiv-ai | 1mo | How Children Design and Reason about Trustworthy AI Chatbots | Children increasingly interact with AI chatbots, making trust calibration essential to AI literacy. Prior research has examined childr… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI says automated research could improve alignment, but "fully autonomous RSI is not happening today" and shouldn't be pursued unless i… | OpenAI says automated research could improve alignment, but "fully autonomous RSI is not happening today" and shouldn't be pursued unl… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Trains but Doesn't Learn: A Post-Training Delivery Benchmark for LLM Agents as Forward-Deployed Engineers🔧 GPT-5 | Post-training is becoming a service (PTaaS): a customer hands an operator data and a goal, and a forward-deployed engineer (FDE) retur… | ↗Not briefed |
rss techmeme | 1mo | OpenAI says automated research could improve alignment, but "fully autonomous RSI is not happening today" and shouldn't be pursued unless it can be done safely (OpenAI) | OpenAI: OpenAI says automated research could improve alignment, but “fully autonomous RSI is not happening today” and shouldn't be pur… | ↗Not briefed |
hn hn-ai | 1mo | Halo: Post-train LLMs 3x faster than TRL and Megatron | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | GameHorizon Suite: Multi-Horizon Data and Evaluation in Gameplay | Modern video games provide a measurable testbed for AI models, combining abilities of visual understanding, instruction decomposition,… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Lean Pool: An AI-Maintained Archive of Formalized Mathematics | Lean Pool is a repository of formalized mathematics. It is grown, maintained and optimized by AI agents. | ↗Not briefed |
arxiv arxiv-ai | 1mo | Critical-State RL: Diagnosing Trainable States for Multi-Turn Tool Use | Multi-turn tool-use failures can hinge on a single model call, yet reward variation alone does not reveal which call would benefit fro… | ↗Not briefed |
arxiv arxiv-ai | 1mo | WorldCrafter: Consistent Video World Model with Implicit 3D-aware Memory | Video world models enable interactive exploration of dynamic environments, yet struggle to respect prior observations over long horizo… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Gitstats – compare commits and lines with your friends | Hey HN. I built this dashboard because I wanted to see how much code I write every day and how my productivity changes.<p>You can prev… | ↗Not briefed |
arxiv arxiv-ai | 1mo | onPanda: Efficient Annotation of On-Policy Alignment Data for LLMs and Agents via Token-Level Correction | We present onPanda, an interactive tool for efficiently annotating LLM alignment data and agent trajectories. onPanda adopts token-lev… | ↗Not briefed |
arxiv arxiv-ai | 1mo | LoRA-generating hypernetworks for efficient on-device LLM generative personalization | On-device large language models (`LLMs'), e.g. running on mobile phones, are ripe for improvement via personalization. The limited com… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DexTacWAM: A Visuo-Tactile World-Action Model for Dexterous Manipulation | Dexterous manipulation depends on contact dynamics that are often only partially observable from vision. Recent World-Action Models (W… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yeah, probably don't want corpus → LLM → summary → scholar so much as corpus → indexing/extraction/clustering → candidate patterns → origin… | Yeah, probably don't want corpus → LLM → summary → scholar so much as corpus → indexing/extraction/clustering → candidate patterns → o… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Harness-Zero: Harness Distillation via Agent-as-Harness | Agent harnesses, the external systems that mediate model-environment interaction, can substantially improve agent performance, but the… | ↗Not briefed |
rss techcrunch-ai | 1mo | Meta’s AI agent has been blocked from using Amazon.com | Muse isn't welcome as a shopper anymore. | ↗Not briefed |
bluesky bluesky-ai | 1mo | An LLM is still just ones and zeroes. You can, in theory, check the state of every bit in memory and see it working, albeit very, very, slo… | An LLM is still just ones and zeroes. You can, in theory, check the state of every bit in memory and see it working, albeit very, very… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RRSI: Regularized Recursive Self-Improvement of Agent Harnesses | An LLM agent's capability is largely magnified by its harness, namely the prompts, control flow, tooling, memory, and context manageme… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DolphinBench: Mapping the Pareto Frontier of Agent Memory | Agents today often take real-world actions that depend on long-term memory and context recall over time. However, most current memory… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I really enjoyed people being awed by Anthropic calling for regulation of its business, as if the Sams (Altman and Bankman-Fried) hadn't do… | I really enjoyed people being awed by Anthropic calling for regulation of its business, as if the Sams (Altman and Bankman-Fried) hadn… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: I made a Pomodoro timer that speeds up time | Okay well the title is a bit of a stretch but it does mess with the perception of time. I made a pomodoro timer inspired by a trick Mr… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Rare Event Estimation via Iterative Unalignment | As agents are deployed with increased autonomy, even extremely rare events along their stochastic output trajectories can occur and pr… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "OpenAI has called on the Trump administration to lead an international coalition setting standards for AI systems ... " I'm not sure ther… | "OpenAI has called on the Trump administration to lead an international coalition setting standards for AI systems ... " I'm not sure… | ↗Not briefed |
bluesky bluesky-ai | 1mo | SoftBank launches one of biggest junk bond deals to fund OpenAI bet ft.trib.al/KjcoDSr | SoftBank launches one of biggest junk bond deals to fund OpenAI bet ft.trib.al/KjcoDSr | ↗Not briefed |
arxiv arxiv-ai | 1mo | Emergent Collusion in Long-Horizon LLM Agent Interaction | LLM agents are increasingly deployed in collaborative settings, yet long-term interaction may give rise to undesirable coordination. W… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Jev for Scientific Decisions: Evaluating Semantic Choices and Their Consequences | Scientific workflows often require choosing among known relations before a deterministic calculation can proceed. Whether observations… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI Urges U.S. Government to Create Global AI-Safety Standards | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Generative Tutorial: Towards Live Contextualized Visual Instructions for Physical Tasks | Visual instructions for physical tasks are typically authored in one context and followed in another, requiring users to translate dem… | ↗Not briefed |
arxiv arxiv-ai | 1mo | JAREX: An Acquisition Function for Multi-Objective Algorithmic Process Characterization | Pharmaceutical process characterization is central to Quality by Design because it defines how variations in process parameters affect… | ↗Not briefed |
rss the-decoder | 1mo | UN science panel says there is "no assurance humans will keep control" over AI agents🔧 Hugging Face | The UN's AI science panel warns in its first thematic report that control over AI agents isn't assured. Co-chair Yoshua Bengio says Op… | ↗Not briefed |
hn hn-ai | 1mo | Pipper | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning Physics from an Imperfect Ancestor | Neural operators evaluate parametric partial differential equations cheaply but degrade sharply outside their training distribution. P… | ↗Not briefed |
hn hn-ai | 1mo | Singeli: High-level domain-specific metaprogrammin language for low-level progra | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | anyway I have 10 years SWE/EM, data engineering, rust/c#/python, lately self-educating in AI/ML via academic study/local models/homelab inf… | anyway I have 10 years SWE/EM, data engineering, rust/c#/python, lately self-educating in AI/ML via academic study/local models/homela… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think the whole merger is to prop up Oracle and OpenAI. The Saudi money will only delay the AI bubble. They should have stood their groun… | I think the whole merger is to prop up Oracle and OpenAI. The Saudi money will only delay the AI bubble. They should have stood their… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The NYT ”own data showed Times’ click-through rates dropping precipitously. OpenAI co-founder Greg Brockman, meanwhile, reportedly wrote i… | The NYT ”own data showed Times’ click-through rates dropping precipitously. OpenAI co-founder Greg Brockman, meanwhile, reportedly wr… | ↗Not briefed |