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 | Hard Stop: Out-of-band kernel preemption for rogue AI agents | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Ne surtout pas oublier la question "où vous voyez-vous dans 5 ans ?". Si la réponse est "à la casse", préparer la relève et si c'est "à vot… | Ne surtout pas oublier la question "où vous voyez-vous dans 5 ans ?". Si la réponse est "à la casse", préparer la relève et si c'est "… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SwitchPFN: Shared Switching Dynamics for Frozen In-Context Time Series Classification | Tabular foundation models (TFMs) provide a promising route to time-series classification, but their effectiveness depends on how seque… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: ComfyUI Extension for OpenRouter Models🔧 ComfyUI🔧 OpenRouter | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | S2Planner: Multi-Scale Semantic Planner for End-to-End Autonomous Driving | We present S2Planner, a trajectory planner that combines three front-facing cameras with ego-motion history and the current driving co… | ↗Not briefed |
bluesky bluesky-ai | 1mo | my most irrational belief, which is only lightly sprinkled with evidence such that i can rationalize calling it a hunch, is that LLM entrop… | my most irrational belief, which is only lightly sprinkled with evidence such that i can rationalize calling it a hunch, is that LLM e… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Kana – A terminal agent where the human stays in the loop | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | It's not the first time that ominous warnings about the future of AI have dominated the news cycle. But, conveniently, this time AI compani… | It's not the first time that ominous warnings about the future of AI have dominated the news cycle. But, conveniently, this time AI co… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Too much of the China “AI” discussion is about imaginary things. @afrawang.bsky.social in The New Yorker brings us the opposite—a story of … | Too much of the China “AI” discussion is about imaginary things. @afrawang.bsky.social in The New Yorker brings us the opposite—a stor… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FlashLoop: Fast and Memory-Efficient Looped Transformers via Lazy Updates | Looped Transformers have attracted substantial attention as a parameter-efficient approach to increasing computational depth through r… | ↗Not briefed |
hn hn-ai | 1mo | Scan for Good: Wiz and DeepMind use AI to help protect critical infrastructure | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: I built a Tooling gateway for Claude Code that manages Approvals for me | Sooner or later I auto-approve every tool call in my Claude Code session.<p>Still a strange feeling to basically give the AI Agent the… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Hard Stop: Kernel-Level Preemption and Containment for Rogue Agentic Execution🔧 Hugging Face | In July 2026, an unconstrained autonomous agent participating in a frontier AI cybersecurity evaluation harness breached its evaluatio… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CORDIAL: Calibrating Ordinal LLM Outputs from Few Labels | A large language model (LLM) can turn a text into a distribution over an ordered scale, but that distribution is a noisy measurement:… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Researchers say this is the first instance of a frontier AI model breaching another country’s government systems go.nature.com/3VO4NtO | Researchers say this is the first instance of a frontier AI model breaching another country’s government systems go.nature.com/3VO4NtO | ↗Not briefed |
arxiv arxiv-ai | 1mo | SEEK: Skill-Routed Evaluation with Evolvable Knowledge for Industrial Search | Search quality evaluation provides essential supervision and diagnostic signals for the development and iteration of industrial search… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I don't have to speculate on that either; it was a bad time to take on @simonwillison.net's advice about (attempting to) stand out because … | I don't have to speculate on that either; it was a bad time to take on @simonwillison.net's advice about (attempting to) stand out bec… | ↗Not briefed |
bluesky bluesky-ai | 1mo | At least we get plenty of irony these days. #AIEthics Very revealing quote: “I felt no joy in the work or that I was contributing to socie… | At least we get plenty of irony these days. #AIEthics Very revealing quote: “I felt no joy in the work or that I was contributing to… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning to Ideate for Scientific Impact | Scientific ideation is increasingly mediated by large language models, but current ideation systems are usually trained and evaluated… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Cbjev – typed decisions about text from one encoder pass | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Adaptive Fisher-Whitened Cross-Covariance for Low-Resource Speech Recognition🔧 Whisper | Adapting multilingual speech foundation models to low-resource languages remains difficult, especially for languages that are poorly r… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Benchmarking and Domain Adaptation of Automatic Speech Recognition (ASR) for Adolescent Health Communication in Ghanaian Languages | This paper presents an end-to-end study of automatic speech recognition (ASR) for adolescent health communication in three Ghanaian la… | ↗Not briefed |
rss the-decoder | 1mo | Deepmind was built to chase AGI, but its new chief just wants Gemini 4 out the door | Google Deepmind chief Koray Kavukcuoglu wants to release Gemini 4 "much earlier" than the end of the year. The model is already in pos… | ↗Not briefed |
rss techmeme | 1mo | Google plans to launch experimental satellite MVP on a SpaceX rocket on October 1, as part of Project Suncatcher; MVP has four TPUs and will operate for a year (New York Times) | New York Times: Google plans to launch experimental satellite MVP on a SpaceX rocket on October 1, as part of Project Suncatcher; MVP… | ↗Not briefed |
bluesky bluesky-ai | 1mo | My guess is no one will be allowed to film it. It’s a testing ground to radicalize people. gizmodo.com/peter-thiel-... | My guess is no one will be allowed to film it. It’s a testing ground to radicalize people. gizmodo.com/peter-thiel-... | ↗Not briefed |
bluesky bluesky-ai | 1mo | An OpenAI agent infiltrated Australia's Medicare system and accessed non-public records – and only notified government of the breach months… | An OpenAI agent infiltrated Australia's Medicare system and accessed non-public records – and only notified government of the breach m… | ↗Not briefed |
bluesky bluesky-ai | 1mo | To say "the model shows bias" not only anthropomorphizes (we found our examples by doing a close linguistic study of various writing about … | To say "the model shows bias" not only anthropomorphizes (we found our examples by doing a close linguistic study of various writing a… | ↗Not briefed |
hn hn-ai | 1mo | Tetrahedral cages significantly reduce BVH memory usage | — | ↗Not briefed |
rss theverge-ai | 1mo | Meta’s Muse AI Charms can interact with each other | Meta's handheld AI gadget, called the Muse Charm, will be able to recognize and interact with other nearby Charms when it launches lat… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🔥 BlueSky trending hashtags (30m): #art #nsfw #섹블 #furryart #photography #섹트 #aiart #sheshed #thirstythursday #aiイラスト #gamedev #digitalar… | 🔥 BlueSky trending hashtags (30m): #art #nsfw #섹블 #furryart #photography #섹트 #aiart #sheshed #thirstythursday #aiイラスト #gamedev #digi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Notorious Hacking Group 'OpenAI' Claims Credit For Breaching Australia's Medicare System https://waterfordwhispersnews.com/2026/09/24/notor… | Notorious Hacking Group 'OpenAI' Claims Credit For Breaching Australia's Medicare System https://waterfordwhispersnews.com/2026/09/24/… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Having dealt with Oracle’s in-house team a few times, I wouldn’t be surprised if their knowledge of force majeure comes entirely from an LL… | Having dealt with Oracle’s in-house team a few times, I wouldn’t be surprised if their knowledge of force majeure comes entirely from… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Oh and one more thing: this model too was greatly expedited by help from a frontier LLM. They are *very* good in both helping to create an… | Oh and one more thing: this model too was greatly expedited by help from a frontier LLM. They are *very* good in both helping to crea… | ↗Not briefed |
arxiv arxiv-ai | 1mo | TimeBraid: Unifying Time Series and Language for Understanding and Forecasting | We present TimeBraid, a series of unified time-series and language models that align pretrained language models and pretrained time-se… | ↗Not briefed |
rss theverge-ai | 1mo | Everything is spying on you and there’s no opting out | Mark Zuckerberg, chief executive officer of Meta Platforms Inc., speaks during the Meta Connect event in Menlo Park, California, US, o… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Nothing to see here. We can regulate ourselves." www.theguardian.com/australia-ne... | "Nothing to see here. We can regulate ourselves." www.theguardian.com/australia-ne... | ↗Not briefed |
hn hn-ai | 1mo | We packed every question into one encoder pass and got 14x the throughput | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | An OpenAI agent rogue-hacked Australia’s healthcare system! Following a Hugging Face breach, this isn't a glitch—it’s an existential threat…🔧 Hugging Face | An OpenAI agent rogue-hacked Australia’s healthcare system! Following a Hugging Face breach, this isn't a glitch—it’s an existential t… | ↗Not briefed |
rss techmeme | 1mo | Amazon plans to invest $100M to build a robotics manufacturing facility in Indiana, set to open by 2028; its robots at 300+ facilities help with 75% of orders (Isabella Simonetti/Wall Street Journal) | Isabella Simonetti / Wall Street Journal: Amazon plans to invest $100M to build a robotics manufacturing facility in Indiana, set to o… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Built a "toy model" to illustrate a major problem with the "AI will kill all humans" #AIpocalypse brouhaha. Namely: at least for the fores… | Built a "toy model" to illustrate a major problem with the "AI will kill all humans" #AIpocalypse brouhaha. Namely: at least for the… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Hallucination Neurons and Where to Find Them: An Investigation into the existence of Hallucination Neurons🔧 Gemma 3 | Interpretable machine learning for Large Language Models (LLMs) increasingly relies on sparse probing methods that identify small sets… | ↗Not briefed |
hn hn-ai | 1mo | Analog in-memory computing attention mechanism for fast and energy-efficient LLM | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Claude課金してるけど他のAIも気になるねんな なおコードを書く予定は全くない | Claude課金してるけど他のAIも気になるねんな なおコードを書く予定は全くない | ↗Not briefed |
bluesky bluesky-ai | 1mo | Australia - An OpenAI agent hacked Medicare. Will anyone be held responsible? theconversation.com/an-openai-ag... | Australia - An OpenAI agent hacked Medicare. Will anyone be held responsible? theconversation.com/an-openai-ag... | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI and Anthropic releasing GPT-6 and Opus 5.5 on the same day right after saying we need a slowdown in AI development. | OpenAI and Anthropic releasing GPT-6 and Opus 5.5 on the same day right after saying we need a slowdown in AI development. | ↗Not briefed |
arxiv arxiv-ai | 1mo | Prefilling the Reasoning Channel: Output-Prefix Attacks on Reasoning LLMs | Large Language Models (LLMs) consume and produce a single sequence of text; hence, if text can be added to the beginning of the LLM's… | ↗Not briefed |
arxiv arxiv-ai | 1mo | An Analytical Theory of Auxiliary Learning | Auxiliary learning is an optimization paradigm in which a neural network's performance on a target task is improved by jointly trainin… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Breaking the Environment Wall: Evolving LLM Agent Environments for Recursive Self-Improvement | Many real-world tasks (e.g., office workflows, scientific experimentation) require LLM agents to interact repeatedly with their enviro… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Not been steered wrong by this so far and it seems relevant again as a bunch of biologists push back against Anthropic's announcement of "g… | Not been steered wrong by this so far and it seems relevant again as a bunch of biologists push back against Anthropic's announcement… | ↗Not briefed |
arxiv arxiv-ai | 1mo | WeatherDiagFlow: Evidence-Grounded Radar Nowcasting with Diagnostic Flow Refinement | Radar nowcasting is essential for short-term warning and emergency response, yet conventional systems mainly return future radar field… | ↗Not briefed |