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 |
|---|---|---|---|---|
bluesky bluesky-ai | 18d | In 2025 and 2026, US prosecutors said, the 35-year-old Saudi Arabian national — whose salary at OpenAI was $385,000-a-year plus a $2.2mn eq… | In 2025 and 2026, US prosecutors said, the 35-year-old Saudi Arabian national — whose salary at OpenAI was $385,000-a-year plus a $2.2… | ↗Not briefed |
bluesky bluesky-ai | 18d | Gillibrand’s son got $30m from a crypto investor. Dems were implicated in the Susan Collins corruption scandal. Schumer’s daughter works fo… | Gillibrand’s son got $30m from a crypto investor. Dems were implicated in the Susan Collins corruption scandal. Schumer’s daughter wor… | ↗Not briefed |
bluesky bluesky-ai | 18d | Ai company Anthropocene being sued by Canadian news publishers. Copyright lawsuit. #cdnpoli | Ai company Anthropocene being sued by Canadian news publishers. Copyright lawsuit. #cdnpoli | ↗Not briefed |
bluesky bluesky-ai | 18d | Percival AI can flag disease beyond a CT scan’s original purpose. It identified pancreatic cancer more reliably than a comparison model, bu… | Percival AI can flag disease beyond a CT scan’s original purpose. It identified pancreatic cancer more reliably than a comparison mode… | ↗Not briefed |
bluesky bluesky-ai | 18d | Ronny talks to former OpenAI researcher Daniel Kokotajlo about the imminent threat posed by AI | Ronny talks to former OpenAI researcher Daniel Kokotajlo about the imminent threat posed by AI | ↗Not briefed |
arxiv arxiv-ai | 18d | What Not to Erase: Saliency-Constrained Augmentation for Low-Resource Handwritten Character Recognition | Optical Character Recognition (OCR) and handwritten character recognition (HCR) in low-resource historical material collections often… | ↗Not briefed |
hn hn-ai | 18d | A Guided Tour of Sketchpad | — | ↗Not briefed |
rss theverge-ai | 18d | CoreWeave’s debt is becoming the whole AI industry’s problem | CoreWeave had a bad week last week. Its bonds slipped into distressed territory by a common definition - when investors demand 1,000 b… | ↗Not briefed |
bluesky bluesky-ai | 18d | Abdulrahman Alfozan’s wife and young son also left the country the day after the 35-year-old Saudi national’s arrest, traveling on a one-wa… | Abdulrahman Alfozan’s wife and young son also left the country the day after the 35-year-old Saudi national’s arrest, traveling on a o… | ↗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 was “mere days” …🔧 Glean | A former OpenAI engineer charged with secretly installing a recording device in a conference room to glean company details was “mere d… | ↗Not briefed |
rss techmeme | 18d | Dream Labs, which offers an API for its Dream-1 robot foundation model trained on human videos, raised a $90M seed led by B Capital and Kindred Ventures (John Koetsier/Forbes) | John Koetsier / Forbes: Dream Labs, which offers an API for its Dream-1 robot foundation model trained on human videos, raised a $90M… | ↗Not briefed |
bluesky bluesky-ai | 18d | Reasonably clear now that Crown Prince Bone Saw was the man behind an industrial espionage caper at Open AI that aimed to grab their essent… | Reasonably clear now that Crown Prince Bone Saw was the man behind an industrial espionage caper at Open AI that aimed to grab their e… | ↗Not briefed |
arxiv arxiv-ai | 18d | Evolving Self-tuning Benchmarks for Long-Horizon Tool Use🔧 Elicit | For AI assistants to move beyond simple chat interfaces and reliably assist users with applications like Gmail, Venmo, and Slack, we n… | ↗Not briefed |
arxiv arxiv-ai | 18d | When Does an Added Planning Capability Create Operational Value? An Observable Evidence Standard | Industrial artificial intelligence (AI) planning systems combine predictive models, optimization logic, controls, and governance rules… | ↗Not briefed |
arxiv arxiv-ai | 18d | Global Security Blind Spots: Translation as a Safety Bypass in Low-Resource Languages🔧 Elicit | Large language models are deployed globally, yet safety alignment and evaluation remain concentrated in English: harmful requests refu… | ↗Not briefed |
hn hn-ai | 18d | Show HN: Claude Code skills and hooks that prove a fix's test fails without it | — | ↗Not briefed |
arxiv arxiv-ai | 18d | Evaluating Context Compaction on the Questions That Come Later | Language-model agents compact their context for the question at hand, but the compacted context must also serve questions that come la… | ↗Not briefed |
bluesky bluesky-ai | 18d | Meta Says Claude based tool using #ArtificialIntelligence Triggered #NoKings Instagram Suspension finance.yahoo.com/technology/a... | Meta Says Claude based tool using #ArtificialIntelligence Triggered #NoKings Instagram Suspension finance.yahoo.com/technology/a... | ↗Not briefed |
arxiv arxiv-ai | 18d | Generalizing Multimodal Foundational Models in Context-Driven Class Incremental Learning | Class-Incremental Learning (CIL) aims to learn new classes continuously without forgetting the previously learned ones. Recent works i… | ↗Not briefed |
arxiv arxiv-ai | 18d | Stronger artificial intelligence breakthroughs hasten their own exhaustion | Progress in artificial intelligence is widely assumed to scale with effort: more researchers, papers and investment should accelerate… | ↗Not briefed |
bluesky bluesky-ai | 18d | Anthropic sued by major Canadian news publishers for copyright claims in Ontario court. Newspapers say Anthropic infringed on their copyrig… | Anthropic sued by major Canadian news publishers for copyright claims in Ontario court. Newspapers say Anthropic infringed on their co… | ↗Not briefed |
arxiv arxiv-ai | 18d | RealControlBench: A Benchmark for Interpretability-Informed Fine-Grained Control of LLMs and Agents | Existing steering benchmarks are built around simplified, isolated generation tasks and evaluate a single behavioral attribute under c… | ↗Not briefed |
arxiv arxiv-ai | 18d | Vulnerability Analysis for Safe Multi-Agent Reinforcement | Safe Multi-Agent Reinforcement Learning (SMARL) aims to learn cooperative control policies under prescribed safety constraints. Al- th… | ↗Not briefed |
bluesky bluesky-ai | 18d | Funding news 16/20 Opsin raises $21m to police what enterprise AI agents do SYN Ventures and Race Capital led the round; Opsin, which bega… | Funding news 16/20 Opsin raises $21m to police what enterprise AI agents do SYN Ventures and Race Capital led the round; Opsin, which… | ↗Not briefed |
bluesky bluesky-ai | 18d | KI-Agenten von OpenAI haben haben Crypto geklaut. Ich liebe alles an dieser Meldung. V.a. wenn ich an Gespräche denke, in denen Leute mir … | KI-Agenten von OpenAI haben haben Crypto geklaut. Ich liebe alles an dieser Meldung. V.a. wenn ich an Gespräche denke, in denen Leute… | ↗Not briefed |
hn hn-ai | 18d | Terminal makrdown reader – feedback welcome | — | ↗Not briefed |
arxiv arxiv-ai | 18d | TEX: Test-Time Scaling Testing Agents via Execution-based Cross-Validation | Test-case generation is critical but remains challenging for language models. Existing test-time scaling methods either isolate candid… | ↗Not briefed |
bluesky bluesky-ai | 18d | I'm at a virtual conference on health literacy and every single slot has at least one talk on AI and at least half of them used Flesch-Kinc… | I'm at a virtual conference on health literacy and every single slot has at least one talk on AI and at least half of them used Flesch… | ↗Not briefed |
hn hn-ai | 18d | UseRef in React: Refs vs. State, DOM Access and React 19 | — | ↗Not briefed |
bluesky bluesky-ai | 18d | 💀 Anthropic Secretly Lobbying Government to Only Shut Down an AI Model If It's Poised to Kill More Than 50 People futurism.com/artificial-… | 💀 Anthropic Secretly Lobbying Government to Only Shut Down an AI Model If It's Poised to Kill More Than 50 People futurism.com/artifi… | ↗Not briefed |
arxiv arxiv-ai | 18d | When Does Hate Transfer? A Large-Scale Study of Cross-Lingual Hate Speech Detection | Effective detection of hate content is essential for maintaining healthy online communities, yet performance can vary substantially ac… | ↗Not briefed |
bluesky bluesky-ai | 18d | Anthropic sued by major Canadi... www.cbc.ca/news/busines... yes!!!!! Hope they win! #AI #cdnpoli | Anthropic sued by major Canadi... www.cbc.ca/news/busines... yes!!!!! Hope they win! #AI #cdnpoli | ↗Not briefed |
hn hn-ai | 18d | Show HN: Craig – A TUI for running parallel coding agents | Hello HN!<p>My name is Sam and this is Craig. Craig is a local-first TUI for managing AI coding agents. It’s agent-agnostic, creates i… | ↗Not briefed |
arxiv arxiv-ai | 18d | VPFlow: Value-Predictive Planning with Imperfect World Models for Offline Goal Reaching | Offline goal-conditioned reinforcement learning enables skill acquisition by reaching any goal from any state without dense rewards. M… | ↗Not briefed |
arxiv arxiv-ai | 18d | A latent control model for realistic rodent motion | The development of embodied models of animal behavior is a major goal of the neuro-AI community. Embodied models will enhance our abil… | ↗Not briefed |
hn hn-ai | 18d | Show HN: Create commitments that only begin once enough people have signed up | Hi all! I made a small free site for creating conditional commitments - commitments that are only "binding" once a sufficien… | ↗Not briefed |
arxiv arxiv-ai | 18d | What Would Removing This Corruption Change? Failure Attribution for Vision-Language-Action Models | Vision-language-action (VLA) models are sensitive to visual corruptions, but knowing that a corruption is typically harmful does not e… | ↗Not briefed |
bluesky bluesky-ai | 18d | One way to know that AI is a bubble is that if consumers (including businesses) had to actually pay the per-token cost, the chatbot busines… | One way to know that AI is a bubble is that if consumers (including businesses) had to actually pay the per-token cost, the chatbot bu… | ↗Not briefed |
hn hn-ai | 18d | Show HN: Flyingcar.pet – a physics simulation of Tesla's flying Roadster stunt | Hi HN. I kept watching the Tesla Roadster's flying video and wondered how much force would be required to pull off a stunt like t… | ↗Not briefed |
bluesky bluesky-ai | 18d | try to carefully use an LLM to do a thing -> "you're an idiot" try to carefully describe the ways you tried to be rigorous about using an … | try to carefully use an LLM to do a thing -> "you're an idiot" try to carefully describe the ways you tried to be rigorous about usin… | ↗Not briefed |
bluesky bluesky-ai | 18d | i feel like the looming AI unemployment crisis is a result of the cloud model and very powerful business interests trying to force that cen… | i feel like the looming AI unemployment crisis is a result of the cloud model and very powerful business interests trying to force tha… | ↗Not briefed |
bluesky bluesky-ai | 18d | “AI is absolutely useless” Can you read this? I can’t and yet I was having to peruse hundreds of pages of this to find records. Claude fou… | “AI is absolutely useless” Can you read this? I can’t and yet I was having to peruse hundreds of pages of this to find records. Claud… | ↗Not briefed |
bluesky bluesky-ai | 18d | True physiognomy has never been tried, only failed. | True physiognomy has never been tried, only failed. | ↗Not briefed |
arxiv arxiv-ai | 18d | FLINT: Fixed-budget Late Interaction | Multi-vector retrieval models such as ColBERT often outperform single-vector models of comparable scale by matching queries and docume… | ↗Not briefed |
bluesky bluesky-ai | 18d | was it because of that post? who knows. do i still have 4 "strikes" according to the clearly-unfamiliar-with-how-baseball-works moderators?… | was it because of that post? who knows. do i still have 4 "strikes" according to the clearly-unfamiliar-with-how-baseball-works modera… | ↗Not briefed |
arxiv arxiv-ai | 18d | Certifying Repeatable Entry-Replay Witnesses for COBOL-to-Java Migration | Producing synthetic data for comparison testing a migrated COBOL application is a major challenge that agentic migrations face today.… | ↗Not briefed |
arxiv arxiv-ai | 18d | African Languages AI Safety Readiness Report | With the widespread adoption of large language models (LLMs) across languages and domains, AI safety has emerged as a critical global… | ↗Not briefed |
bluesky bluesky-ai | 18d | Ex-OpenAI worker accused of bugging offices planned to 'flee' to Saudi Arabia, prosecutors say | Ex-OpenAI worker accused of bugging offices planned to 'flee' to Saudi Arabia, prosecutors say | ↗Not briefed |
hn hn-ai | 18d | Ex-OpenAI employee accused of planting device appears to have Saudi backing | — | ↗Not briefed |
bluesky bluesky-ai | 18d | “Our goal is to make face photograph-based health assessment more quantitative, consistent and scalable so it can support personalized trea… | “Our goal is to make face photograph-based health assessment more quantitative, consistent and scalable so it can support personalized… | ↗Not briefed |