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 | 13d | Spatial Harness: Grounding Spatial Reasoning in Executable World States with Physics Engines | Spatial reasoning requires understanding how objects and observers are arranged in a scene and how their relations change with viewpoi… | ↗Not briefed |
hn hn-ai | 13d | High level API for local LLMs in Node | — | ↗Not briefed |
hn hn-ai | 13d | Anthropic strikes $3.4B deal to generate gas-fired power for data centres | — | ↗Not briefed |
bluesky bluesky-ai | 13d | AI adoption is often framed as a model choice. In practice, the hard work is redesigning decisions, workflows and accountability around it.… | AI adoption is often framed as a model choice. In practice, the hard work is redesigning decisions, workflows and accountability aroun… | ↗Not briefed |
arxiv arxiv-ai | 13d | Agent Self-Improvement with Frozen Weights: The Added Value and Limits of Failure Traces | We evaluate whether an AI agent's harness, the notes around its model, can learn from the agent's failure traces beyond a matched sear… | ↗Not briefed |
arxiv arxiv-ai | 13d | Agentic Attribution of Cloud Resource Ownership | Security tools raise findings across cloud environments, and AI agents now investigate and prioritize them. A finding is still resolve… | ↗Not briefed |
hn hn-ai | 13d | LLMs are terrible at designing AI systems | — | ↗Not briefed |
bluesky bluesky-ai | 13d | これめちゃくちゃ面白い x.com/adachihirota... | これめちゃくちゃ面白い x.com/adachihirota... | ↗Not briefed |
hn hn-ai | 13d | Show HN: Neural Network Formal Verification | — | ↗Not briefed |
bluesky bluesky-ai | 13d | In case anyone was in any doubt, with all the optimistic chatter about data centres powered by nuclear SMRs or providing flexible demand fo… | In case anyone was in any doubt, with all the optimistic chatter about data centres powered by nuclear SMRs or providing flexible dema… | ↗Not briefed |
arxiv arxiv-ai | 13d | OPSRL-SSP: Optimistic Posterior Sampling for Stochastic Shortest Path with Minimax-Optimal Regret | We study online reinforcement learning for the stochastic shortest path (SSP) problem with unknown transition dynamics. Although exist… | ↗Not briefed |
arxiv arxiv-ai | 13d | Where Should a Gate Look? Input vs. State Gating in Diagonal State Space Models | Modern state-space models owe much of their computational efficiency to their linear, often diagonal, dynamics. Classical recurrent ne… | ↗Not briefed |
bluesky bluesky-ai | 13d | Google AI just dropped the Elephant‑Goldfish plugin—an open‑source, specification‑first LLM that powers smarter coding agents. Think Claude… | Google AI just dropped the Elephant‑Goldfish plugin—an open‑source, specification‑first LLM that powers smarter coding agents. Think C… | ↗Not briefed |
bluesky bluesky-ai | 13d | Pretty sure we need to treat them as a clear threat to democracy then? futurism.com/artificial-i... | Pretty sure we need to treat them as a clear threat to democracy then? futurism.com/artificial-i... | ↗Not briefed |
hn hn-ai | 13d | Show HN: No more code reviews. Fuzz instead | Hi HN, I'm Stephan.<p>I'm an ex-CTO and many companies I look in today, want to get AI coded apps into production, but engin… | ↗Not briefed |
hn hn-ai | 13d | Show HN: Shutdown Watch – App shutdowns, export deadlines, and alternatives | My recipe app picked up users when a similar app shut down, and I kept noticing people scrambling to export their data, so I created t… | ↗Not briefed |
arxiv arxiv-ai | 13d | MoT3DVG: A Benchmark for Outdoor 3D Visual Grounding with Motion-Aware Descriptions and Temporal Cues | 3D visual grounding (3DVG) localizes language-referred objects in 3D scenes, with broad applicability in both indoor and outdoor envir… | ↗Not briefed |
bluesky bluesky-ai | 13d | Popping bottles because your book went to publication and you realize too late that you and your editor both missed a mistake. "that prov… | Popping bottles because your book went to publication and you realize too late that you and your editor both missed a mistake. "that… | ↗Not briefed |
hn hn-ai | 13d | I will not be using LLMs in my coding | — | ↗Not briefed |
arxiv arxiv-ai | 13d | CVaR Deep Optimal Stopping | Risk-sensitive optimal stopping is useful when poor stopping outcomes carry substantial costs, as in option exercise, sequential medic… | ↗Not briefed |
arxiv arxiv-ai | 13d | When Credit Turns Against Tools: Transition-Aware Credit Assignment in Agentic Reinforcement Learning | Agentic reinforcement learning with verifiable rewards trains language models through multi-turn tool interactions, yet their reasonin… | ↗Not briefed |
bluesky bluesky-ai | 13d | I understand this is a luxury of being an independent researcher/on small informal teams. I am not having to justify model choices to an or… | I understand this is a luxury of being an independent researcher/on small informal teams. I am not having to justify model choices to… | ↗Not briefed |
bluesky bluesky-ai | 13d | Hell yeah | Hell yeah | ↗Not briefed |
bluesky bluesky-ai | 13d | (It is possible to do this without an LLM but it would be janky. Small LLMs are actually useful for building human/computer interfaces like… | (It is possible to do this without an LLM but it would be janky. Small LLMs are actually useful for building human/computer interfaces… | ↗Not briefed |
bluesky bluesky-ai | 13d | China's free open-weight AI models are eating into America's AI lead. #ai #llm #opensource | China's free open-weight AI models are eating into America's AI lead. #ai #llm #opensource | ↗Not briefed |
rss wired-ai | 13d | I Used NYT Cooking’s New AI-Powered Recipe Modifier, and It’s Actually Pretty Good | The New York Times has added a beta feature to its Cooking app that helps readers swap ingredients or adjust recipes to suit their nee… | ↗Not briefed |
arxiv arxiv-ai | 13d | When Sensors Fail: Coverage Auditing with Incomplete Sensor Records | Sensor outages can conceal prediction-interval coverage deficits behind acceptable marginal performance, while unavailable future read… | ↗Not briefed |
hn hn-ai | 13d | Show HN: Gitbet – A free sourcetree style Git client for Linux | I used Sourcetree for about 7 years on macOS and Windows. When I switched to Linux, I found it isn’t available, and none of the altern… | ↗Not briefed |
bluesky bluesky-ai | 13d | Dass Anthropic Teile eines Berichts schwärzt, ist gängige Praxis und kein besonderer Grund für Verschwörungsmärchen. Ebenso wenig ist es ei… | Dass Anthropic Teile eines Berichts schwärzt, ist gängige Praxis und kein besonderer Grund für Verschwörungsmärchen. Ebenso wenig ist… | ↗Not briefed |
bluesky bluesky-ai | 13d | Kilo code since anaconda bought them has given access to the free AI models for some time now. Use their free models for working on project… | Kilo code since anaconda bought them has given access to the free AI models for some time now. Use their free models for working on pr… | ↗Not briefed |
bluesky bluesky-ai | 13d | Из неожиданных последствий LLMопокалипсиса: я кажется переполз обратно на емакс (вместо которого последние 5+ лет использовал VSCode+Helix)… | Из неожиданных последствий LLMопокалипсиса: я кажется переполз обратно на емакс (вместо которого последние 5+ лет использовал VSCode+H… | ↗Not briefed |
rss the-decoder | 13d | The inventor of the bitter lesson has a bitter lesson for LLM scaling enthusiasts | Turing Award winner Richard Sutton reaffirms that raw compute alone isn't enough for intelligence and that fundamentally new learning… | ↗Not briefed |
hn hn-ai | 13d | Aiport: A self-improving local LLM gateway for the erratic vibe coder | — | ↗Not briefed |
arxiv arxiv-ai | 13d | Verify Before You Fetch: Reliable KV Cache Compression for LLM Inference | Long-context applications, long-lived agents, and multi-agent systems require large language models (LLMs) to process and retain exten… | ↗Not briefed |
hn hn-ai | 13d | Terminal, Augmented | — | ↗Not briefed |
bluesky bluesky-ai | 13d | Felt generous and got Mei its own domain/website mei.computer | Felt generous and got Mei its own domain/website mei.computer | ↗Not briefed |
arxiv arxiv-ai | 13d | PROLA: Principal-Orthogonal Low-rank Adaptation for Predictive Spatiotemporal Weather Downscaling | High-resolution weather prediction is important for resolving local atmospheric patterns missed by coarse global forecasts. Large pret… | ↗Not briefed |
hn hn-ai | 13d | Finding Bugs in DuckDB with Property-based Testing | — | ↗Not briefed |
arxiv arxiv-ai | 13d | ReGKG-2025: A GKG-Compatible Reconstruction of the GDELT Global Knowledge Graph for the Missing Period of 14 June to 2 July 2025 | The June 2025 outage in the Global Database of Events, Language, and Tone (GDELT) collection pipeline left a gap in the Global Knowled… | ↗Not briefed |
arxiv arxiv-ai | 13d | Shape-Adaptive Subspace-Constrained Quadratic Matrix Factorization | Quadratic manifold fitting provides a flexible framework for modeling nonlinear low-dimensional structures embedded in high-dimensiona… | ↗Not briefed |
bluesky bluesky-ai | 13d | When Anthropic wrote in the S-1 that they might kill us, did they mean with gas turbines? "Anthropic has struck a massive deal for gas-fir… | When Anthropic wrote in the S-1 that they might kill us, did they mean with gas turbines? "Anthropic has struck a massive deal for ga… | ↗Not briefed |
hn hn-ai | 13d | Analyzing Software Vulnerabilities faster and orders of magnitudes cheaper | Since Anthropic announced Mythos, Software security has came into lime light. Multiple famous products have seen orders of magnitudes… | ↗Not briefed |
hn hn-ai | 13d | PHP2Shell Deep Dive | — | ↗Not briefed |
bluesky bluesky-ai | 13d | DeepSeek is also an integral part of my multi-model vuln research and red team automation approaches. I'm a huge fan | DeepSeek is also an integral part of my multi-model vuln research and red team automation approaches. I'm a huge fan | ↗Not briefed |
hn hn-ai | 13d | Show HN: CloudShare app for iOS, simple way to share to your own S3 bucket | Hello HN. I'm a developer from Spain, working for a US company and building apps as side projects.<p>I discovered that Cloudflare… | ↗Not briefed |
bluesky bluesky-ai | 13d | My patience for cunts like this is getting pretty thin. | My patience for cunts like this is getting pretty thin. | ↗Not briefed |
arxiv arxiv-ai | 13d | ScrapeBench: Evaluating Legal Compliance of AI Agents in Website Scraping | AI agents deployed on the open web are becoming more capable of performing actions that may breach contractual provisions and violate… | ↗Not briefed |
bluesky bluesky-ai | 13d | Anthropic strikes $3.4bn deal to generate gas-fired power for data centres ft.trib.al/tJa67DP | Anthropic strikes $3.4bn deal to generate gas-fired power for data centres ft.trib.al/tJa67DP | ↗Not briefed |
bluesky bluesky-ai | 13d | A slow holiday tease in green satin and red lace. Full uncensored videos on Fansly. Real model, no AI. 18+. Black November – 35% off first… | A slow holiday tease in green satin and red lace. Full uncensored videos on Fansly. Real model, no AI. 18+. Black November – 35% off… | ↗Not briefed |
hn hn-ai | 13d | Show HN: Throng, a project first, terminal 2nd IDE for command line people | I've been beavering away at this since July, and its finally getting to a point where its stabilising.<p>It started Windows only,… | ↗Not briefed |