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 | 16d | OpenAI launches GPT-6 Cyber, which it says outperforms GPT-6 Astra on cyber benchmarks, is better at finding vulnerabilities, and is less likely to go rogue (Emma Roth/The Verge) | Emma Roth / The Verge: OpenAI launches GPT-6 Cyber, which it says outperforms GPT-6 Astra on cyber benchmarks, is better at finding vu… | ↗Not briefed |
bluesky bluesky-ai | 16d | 🏦 The OCC issued its first AI guidance for bank examiners ⚖️ Judge AI by the consequences of errors, not model sophistication 🔗 https://w… | 🏦 The OCC issued its first AI guidance for bank examiners ⚖️ Judge AI by the consequences of errors, not model sophistication 🔗 http… | ↗Not briefed |
arxiv arxiv-ai | 16d | First-Token Attraction in Mamba Dynamics | We study token dynamics in deep Mamba models, with a particular focus on the attracting role of the first token. Unlike attention mech… | ↗Not briefed |
arxiv arxiv-ai | 16d | Dynamic Optimistic Constrained OCO with Memory via Delay Equivalence | Online Convex Optimization with memory and time-varying constraints (COCO-M) is a recently-introduced framework that captures the dyna… | ↗Not briefed |
bluesky bluesky-ai | 16d | www.ft.com/content/2fbd... | www.ft.com/content/2fbd... | ↗Not briefed |
rss deepmind-blog | 16d | Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips | — | ↗Not briefed |
bluesky bluesky-ai | 16d | "Anthropic’s Claude generated the initial draft, a fact concealed for months until Oct. 9." Using AI and failing to disclose AI usage in m… | "Anthropic’s Claude generated the initial draft, a fact concealed for months until Oct. 9." Using AI and failing to disclose AI usage… | ↗Not briefed |
bluesky bluesky-ai | 16d | We see a lot of AI slop ideas dropped in here. You can tell from the text that the person was impressed with what the AI made...but we see … | We see a lot of AI slop ideas dropped in here. You can tell from the text that the person was impressed with what the AI made...but we… | ↗Not briefed |
rss techmeme | 16d | Defense tech startup Shield AI raises $2.5B to help acquire radar company Echodyne, including a $1B Series H at a $28.8B valuation, up from $12.7B in March (Rebecca Torrence/Bloomberg) | Rebecca Torrence / Bloomberg: Defense tech startup Shield AI raises $2.5B to help acquire radar company Echodyne, including a $1B Seri… | ↗Not briefed |
bluesky bluesky-ai | 16d | INTC's been hammered by a few things: weak earnings, a suspended dividend, big job cuts, and most of all their costly foundry push isn't pa… | INTC's been hammered by a few things: weak earnings, a suspended dividend, big job cuts, and most of all their costly foundry push isn… | ↗Not briefed |
bluesky bluesky-ai | 16d | Anthropic admite que el clima supondrá más costes para su negocio pero obvia el impacto que tiene su negocio en el cambio climático | Anthropic admite que el clima supondrá más costes para su negocio pero obvia el impacto que tiene su negocio en el cambio climático | ↗Not briefed |
bluesky bluesky-ai | 16d | Heading to Anthropic HQ to talk w security engineers and dev productivity folks on how their day-to-day work is like (+ how it's changing) … | Heading to Anthropic HQ to talk w security engineers and dev productivity folks on how their day-to-day work is like (+ how it's chang… | ↗Not briefed |
rss theverge-ai | 16d | Grimes is going to star in an AI movie that’s coming to theaters | The AI-generated sci-fi film Gods Don't Give Gifts has added a bit of star power: Pop star Grimes is going to appear in the movie. Bil… | ↗Not briefed |
rss theverge-ai | 16d | OpenAI says its new cybersecurity model is better at not hacking other companies🔧 GPT-5🔧 Hugging Face | OpenAI is launching GPT-6 Cyber, the latest version of its cybersecurity AI model, which it says is better at finding vulnerabilities… | ↗Not briefed |
bluesky bluesky-ai | 16d | im praying we get the deepseek future and closed-weights century of humiliation | im praying we get the deepseek future and closed-weights century of humiliation | ↗Not briefed |
hn hn-ai | 16d | Show HN: Training a SOTA Pangram-like detector on a 4070ti with Claude | This is kinda claude-slop but I was really interested in figuring out a way to break Pangram, and to do that I first needed to have a… | ↗Not briefed |
hn hn-ai | 16d | Show HN: Keeping tabs on updates in my stack was overwhelming so I built unstale | I built unstale it because I found no other tools that could proactively track & inform me during development of all the new updat… | ↗Not briefed |
arxiv arxiv-ai | 16d | Transformer-Based Trajectory Prescription for sEMG-Optimizing Robot-Mediated Rehabilitation | Robot-mediated rehabilitation has the potential to deliver high-intensity, individualized therapy, but existing systems (which have no… | ↗Not briefed |
bluesky bluesky-ai | 16d | Build #knowledgegraph with this tiny 314M #encoder #model: #named-entity recognition, #PII detection and relation extraction Not #OpenSourc… | Build #knowledgegraph with this tiny 314M #encoder #model: #named-entity recognition, #PII detection and relation extraction Not #Open… | ↗Not briefed |
arxiv arxiv-ai | 16d | AutoCompress: Automated Hardware-Guided Model Compression | LLM pruning must preserve quality while meeting hardware- and workload-specific constraints on latency, throughput, and memory. Existi… | ↗Not briefed |
bluesky bluesky-ai | 16d | From @publisherslunch.bsky.social , a lengthy and sympathetic article about the AO3 outage (despite the slightly snarky headline) (paywall)… | From @publisherslunch.bsky.social , a lengthy and sympathetic article about the AO3 outage (despite the slightly snarky headline) (pay… | ↗Not briefed |
hn hn-ai | 16d | Pierre, Bun, Astral, TBPN, and 40 more startups acquired by OpenAI and Anthropic | — | ↗Not briefed |
bluesky bluesky-ai | 16d | OpenAI Is Changing Its Tune on Data Centers Following Electoral Wipeout https://gizmodo.com/openai-is-changing-its-tune-on-data-centers-fol… | OpenAI Is Changing Its Tune on Data Centers Following Electoral Wipeout https://gizmodo.com/openai-is-changing-its-tune-on-data-center… | ↗Not briefed |
rss techmeme | 16d | Aina, which raised $5.5M, launches an "AI ring" with a screen, joystick, and mic to let users talk to their AI agents, take notes, control Apple TV, and more (Ivan Mehta/TechCrunch) | Ivan Mehta / TechCrunch: Aina, which raised $5.5M, launches an “AI ring” with a screen, joystick, and mic to let users talk to their A… | ↗Not briefed |
bluesky bluesky-ai | 16d | いっそのこと使い物にならなくなって忘れ去られればいいのに… いやだめだちゃんとダメな事はダメな事でしたと謝罪と賠償しろ x.com/crazyshyyt/s... | いっそのこと使い物にならなくなって忘れ去られればいいのに… いやだめだちゃんとダメな事はダメな事でしたと謝罪と賠償しろ x.com/crazyshyyt/s... | ↗Not briefed |
bluesky bluesky-ai | 16d | www.theguardian.com/technology/2... | www.theguardian.com/technology/2... | ↗Not briefed |
bluesky bluesky-ai | 16d | www.theguardian.com/technology/2... | www.theguardian.com/technology/2... | ↗Not briefed |
bluesky bluesky-ai | 16d | Video for my #RustNYC talk, "Code Matters: Reflections on Rust in the LLM Age", is now up: www.youtube.com/watch?v=5D7Q... | Video for my #RustNYC talk, "Code Matters: Reflections on Rust in the LLM Age", is now up: www.youtube.com/watch?v=5D7Q... | ↗Not briefed |
arxiv arxiv-ai | 16d | A Generative Model of Contextual Integrity: Appropriate vs. Inappropriate Information Sharing | Contextual integrity (CI) holds that appropriate information flow depends on context-specific norms. The same disclosure may be approp… | ↗Not briefed |
bluesky bluesky-ai | 16d | A eleição dos EUA pode ter sido decidida por IA, só não do jeito que queriam. Em dez/25, o Grok fala pro Trump que venezuelanos adorariam … | A eleição dos EUA pode ter sido decidida por IA, só não do jeito que queriam. Em dez/25, o Grok fala pro Trump que venezuelanos adora… | ↗Not briefed |
arxiv arxiv-ai | 16d | Multimodal-SEA: Benchmarking AI Agent Security under Multimodal Social Engineering Attacks | Personal AI agents face social engineering attacks supported by forged multimodal content, but existing evaluations mainly focus on te… | ↗Not briefed |
hn hn-ai | 16d | WaveDude Next: 2D interactive 2D wave optics simulator | — | ↗Not briefed |
rss the-decoder | 16d | No AI agent scores above 12 percent on SemiAnalysis' new real-world work benchmark | SemiAnalysis is putting AI agents through simulated knowledge work in its new GDPx benchmark, covering everything from chip design to… | ↗Not briefed |
arxiv arxiv-ai | 16d | LeanContext: Disentangling Mathematical Resources from Local Formal Context in Lean Whole-Proof Generation | Neural theorem proving increasingly uses retrieved premises and project context to support proof generation. Evaluating these resource… | ↗Not briefed |
hn hn-ai | 16d | Software Line Specialist | — | ↗Not briefed |
arxiv arxiv-ai | 16d | Aligning Policy Encoders with a Shape Teacher | 3D-aware representations can improve robotic manipulation by providing geometric information that is robust to changes in viewpoint, o… | ↗Not briefed |
arxiv arxiv-ai | 16d | Accelerating Safe Reinforcement Learning with Massive Parallelism | We investigate the problem of scaling safe reinforcement learning (safe RL) to massively parallel training regimes, where a large numb… | ↗Not briefed |
bluesky bluesky-ai | 16d | llm protein modeling is cool but we can safely turn off the picture machines. the drawbacks are not worth images of amy rose dressed like 2… | llm protein modeling is cool but we can safely turn off the picture machines. the drawbacks are not worth images of amy rose dressed l… | ↗Not briefed |
arxiv arxiv-ai | 16d | DART: Domain-Agnostic Residual Transfer for Generalist Anomaly Detection | Generalist Anomaly Detection (GAD) aims to detect anomalies in unseen domains using only a few reference samples. Recent normal-abnorm… | ↗Not briefed |
hn hn-ai | 16d | Paraguay seizes major digital bank tied to president, leaving customers in limbo | — | ↗Not briefed |
hn hn-ai | 16d | Tick – A Smalltalk with a Tcl like syntax | — | ↗Not briefed |
arxiv arxiv-ai | 16d | Accepted but Wrong: Silent Failures of LLM Code Repairs after API Migrations🔧 GPT-5 | After a runtime or API migration, an LLM-repaired procedure can run cleanly yet return wrong results, and a release check on typical t… | ↗Not briefed |
rss techmeme | 16d | Sources: Microsoft is tightening oversight of some data center projects, stepping away from those it deems risky, and debating whether to ink new neocloud deals (Brody Ford/Bloomberg) | Brody Ford / Bloomberg: Sources: Microsoft is tightening oversight of some data center projects, stepping away from those it deems ris… | ↗Not briefed |
hn hn-ai | 16d | Discord Plans Lighter Desktop App with Rust and GPUI | — | ↗Not briefed |
bluesky bluesky-ai | 16d | ...a billion AI slop images and not a single one is the thing the person imagined when they started to mash the keyboard... ...not a singl… | ...a billion AI slop images and not a single one is the thing the person imagined when they started to mash the keyboard... ...not a… | ↗Not briefed |
arxiv arxiv-ai | 16d | Scaling Neural Motor Decoding via Decoupled Behavioral Pretraining | Recent neural decoding approaches achieve strong performance by training on large collections of paired neural-behavioral recordings,… | ↗Not briefed |
arxiv arxiv-ai | 16d | The Geometry of Simulated Perspectives: Dissociating Representation and Behaviour in a Multimodal Language Model | Multimodal language models can simulate different observers, yet it remains unclear how assigned perspectives reshape their internal r… | ↗Not briefed |
rss techmeme | 16d | Greece's parliament approves a 10% capital gains tax on crypto, exempting gains of up to €500 a year, in the country's first comprehensive crypto tax framework (Lefteris Papadimas/Reuters) | Lefteris Papadimas / Reuters: Greece's parliament approves a 10% capital gains tax on crypto, exempting gains of up to €500 a year, in… | ↗Not briefed |
hn hn-ai | 16d | Tobacco company got smoked by leaving secret recipes on a file share | — | ↗Not briefed |
bluesky bluesky-ai | 16d | On AI efficiency gains: "It's bullish for the labs and terribly bearish for everyone that does data centers. Now, that could be an overlap… | On AI efficiency gains: "It's bullish for the labs and terribly bearish for everyone that does data centers. Now, that could be an ov… | ↗Not briefed |