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 | 2mo | CS-Guard: Benchmarking LLM Guardrails for Code Generation Security | Large language models (LLMs) have been ex- ploited to generate malware, but the effective- ness of guardrails for code generation secu… | ↗Not briefed |
arxiv arxiv-ai | 2mo | A practical DIRECT-type algorithm for medium-scale black-box global optimization | The DIRECT algorithm is a deterministic global optimization method known for its versatility and balanced exploration-exploitation str… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Privacy-Preserving Split Learning for Federated LLM Fine-Tuning | Fine-tuning large language models (LLMs) on domain-specific data is essential for downstream adaptation. In many deployments, a partic… | ↗Not briefed |
arxiv arxiv-ai | 2mo | How Fragile Is Safety Alignment at Frontier Scale? A Single-Direction Attack on a 320B MoE | Directional ablation removes an aligned language model's ability to refuse by projecting a single "refusal direction" out of the weigh… | ↗Not briefed |
bluesky bluesky-ai | 2mo | #Mistral raises record €3bn as Europe strains to keep pace in AI race - Launched in 2023, months after ChatGPT’s debut. It quickly became E…🔧 ChatGPT | #Mistral raises record €3bn as Europe strains to keep pace in AI race - Launched in 2023, months after ChatGPT’s debut. It quickly bec… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Ieri Jacob Coxon, ricercatore di Anthropic con un passato a OpenAI, si è dimesso in polemica con l'industria dell'intelligenza artificiale,… | Ieri Jacob Coxon, ricercatore di Anthropic con un passato a OpenAI, si è dimesso in polemica con l'industria dell'intelligenza artific… | ↗Not briefed |
arxiv arxiv-ai | 2mo | MUCnoHARM@GermEval Shared Task 2026: Retrieval-based In-Context Learning for Defamatory Offences, and Where It Falls Short | With hate speech being ubiquitous online, automatic detection is crucial, in particular when it comes to criminally relevant social me… | ↗Not briefed |
hn hn-ai | 2mo | What LLM coding agents have taken from me | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | the millennium prize is nothing, let me know when the llm rewrites the banking software from FORTRAN | the millennium prize is nothing, let me know when the llm rewrites the banking software from FORTRAN | ↗Not briefed |
arxiv arxiv-ai | 2mo | LogiScope-VQA: Benchmarking Vision-Language Models for Logistics Hazard Identification in Industrial Scenarios🔧 GPT-5 | Large Multimodal Models (LMMs) large-scale deployment in industrial warehouse settings specifically necessitates that models exhibit h… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Pairit: A Platform for Live Experiments on Human-AI Collaboration | Organizational design in the era of artificial intelligence requires experimental methods that can test how human-AI groups coordinate… | ↗Not briefed |
bluesky bluesky-ai | 2mo | 11/ "Earlier today, AI researcher Jacob Coxon, who spent the past three years working at OpenAI and Anthropic, publicly resigned from Anthr… | 11/ "Earlier today, AI researcher Jacob Coxon, who spent the past three years working at OpenAI and Anthropic, publicly resigned from… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Evaluating Model Retraining under Drift: Paired Comparisons of Cumulative Subgroup Disparity | Choosing when to retrain a deployed classifier requires assessing subgroup error rates across the sequence of models used, including p… | ↗Not briefed |
hn hn-ai | 2mo | GPT3 Writing Code | — | ↗Not briefed |
arxiv arxiv-ai | 2mo | NEXUS-MI: Communication-Aware Federated Personalization for Gateway-Coordinated Motor-Imagery Brain-Computer Interfaces | Electroencephalography (EEG)-based motor-imagery brain-computer interfaces (MI-BCIs) vary across subjects and sessions, complicating p… | ↗Not briefed |
bluesky bluesky-ai | 2mo | This the more wholesome version of the Math Ozymandias hypothesis, which is that OpenAI or Anthropic has rounded up all the mathematicians … | This the more wholesome version of the Math Ozymandias hypothesis, which is that OpenAI or Anthropic has rounded up all the mathematic… | ↗Not briefed |
bluesky bluesky-ai | 2mo | rumor is that a team of researchers at OpenAI may have potentially found him | rumor is that a team of researchers at OpenAI may have potentially found him | ↗Not briefed |
arxiv arxiv-ai | 2mo | BRACE: Anchored Bellman-Residual Correction for Stale Critics in Asynchronous RL | Asynchronous reinforcement learning has become the standard way to scale training for language models, but the resulting policy lag bi… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Google LLM verweigert nachdem ich 2 Mal die erfundenen Suchergebnisse korrigiert und es nun zur Nutzung von Google aufgefordert mit Ziel ei… | Google LLM verweigert nachdem ich 2 Mal die erfundenen Suchergebnisse korrigiert und es nun zur Nutzung von Google aufgefordert mit Zi… | ↗Not briefed |
hn hn-ai | 2mo | Canceling Emily | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | OpenAI said that an unreleased AI model had resolved one of the most significant open questions in modern math research. The problem conce… | OpenAI said that an unreleased AI model had resolved one of the most significant open questions in modern math research. The problem… | ↗Not briefed |
arxiv arxiv-ai | 2mo | ROAM: Robust Organization of Atomic Memories for Agents through Semantic Relations | Long-term language-model agents rely on external memory across interactions. Atomic memories are particularly useful: their fine-grain… | ↗Not briefed |
bluesky bluesky-ai | 2mo | if you really believe that, how much do you think shooting your coworkers (or if you prefer the guys across the street at openAI) changes t… | if you really believe that, how much do you think shooting your coworkers (or if you prefer the guys across the street at openAI) chan… | ↗Not briefed |
bluesky bluesky-ai | 2mo | The science version of this would be: There is a difference between writing a paper with an LLM and hoping an LLM writes a paper. | The science version of this would be: There is a difference between writing a paper with an LLM and hoping an LLM writes a paper. | ↗Not briefed |
bluesky bluesky-ai | 2mo | I suspect the protectionism angle here is hiding something more sinister: Allied, foreign governments are more oriented toward testing an… | I suspect the protectionism angle here is hiding something more sinister: Allied, foreign governments are more oriented toward testi… | ↗Not briefed |
bluesky bluesky-ai | 2mo | EXCLUSIVE: China's DeepSeek taps CITIC Securities for domestic IPO, sources say reut.rs/4iRPCJE | EXCLUSIVE: China's DeepSeek taps CITIC Securities for domestic IPO, sources say reut.rs/4iRPCJE | ↗Not briefed |
arxiv arxiv-ai | 2mo | Proof-Carrying Cognition: Closing the Verification Gap with Reality-Settled Reward | Frontier gains in language-model reasoning come from reinforcement learning on reasoning traces and are concentrated in domains with a… | ↗Not briefed |
bluesky bluesky-ai | 2mo | heard a rumor that anthropic was working on this actually | heard a rumor that anthropic was working on this actually | ↗Not briefed |
arxiv arxiv-ai | 2mo | Procedural Memory Under Change: Reuse and Interference in Controlled Web Tasks | Procedural memory lets language agents reuse successful routines, but reuse presumes that a stored routine remains applicable. We stud… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Generative AI might have aced the bar exam, but an LLM still can’t think like a lawyer. | Generative AI might have aced the bar exam, but an LLM still can’t think like a lawyer. | ↗Not briefed |
bluesky bluesky-ai | 2mo | OpenAI ogłosiło i obeszło wszystkie możliwe media twierdząc, że Chat GPT rozwiązał jakiś nierozwiązywalny problem matematyczny. Okazało się… | OpenAI ogłosiło i obeszło wszystkie możliwe media twierdząc, że Chat GPT rozwiązał jakiś nierozwiązywalny problem matematyczny. Okazał… | ↗Not briefed |
bluesky bluesky-ai | 2mo | 1. Environmental crisis and the socio-political #polycrisis are, by far, my #1 bets for what could really ruin the day of human civilizatio… | 1. Environmental crisis and the socio-political #polycrisis are, by far, my #1 bets for what could really ruin the day of human civili… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Sources: Anthropic declined to submit Mythos 5.1 to the UK AISI for pre-release testing, prompting UK fears that US AI labs are aligning wi… | Sources: Anthropic declined to submit Mythos 5.1 to the UK AISI for pre-release testing, prompting UK fears that US AI labs are aligni… | ↗Not briefed |
rss techmeme | 2mo | Sources: Anthropic declined to submit Mythos 5.1 to the UK AISI for pre-release testing, prompting UK fears that US AI labs are aligning with US protectionism (Financial Times) | Financial Times: Sources: Anthropic declined to submit Mythos 5.1 to the UK AISI for pre-release testing, prompting UK fears that US A… | ↗Not briefed |
hn hn-ai | 2mo | Ask HN: Why is HN focusing a lot on human drama and not AI solving ME problem? | Title length restricted.<p>Why is HN focusing so much on the human drama of OpenAI and Anthropic and not the fact that AI just solved… | ↗Not briefed |
arxiv arxiv-ai | 2mo | SymbolicLight V2: Hybrid Neuromorphic Architecture and Sparse Execution for Low-Energy Language Inference | SymbolicLight V2 combines sparse event computation with continuous-state processing in a hybrid neuromorphic language architecture. Ex… | ↗Not briefed |
bluesky bluesky-ai | 2mo | A researcher who worked at both OpenAI and Anthropic has quit, accusing the companies of racing toward superintelligent AI despite believin… | A researcher who worked at both OpenAI and Anthropic has quit, accusing the companies of racing toward superintelligent AI despite bel… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Andrew Garfield plays Sam Altman in Artificial, Luca Guadagnino's new movie about the chaotic rise of OpenAI. | Andrew Garfield plays Sam Altman in Artificial, Luca Guadagnino's new movie about the chaotic rise of OpenAI. | ↗Not briefed |
arxiv arxiv-ai | 2mo | Fine-Tuning a KV Cache Concatenation-Aware Model or Recomputing KV Caches? Why Not Both? | In Retrieval-Augmented Generation (RAG) systems, a large number of retrieved chunks are concatenated to form the input context so that… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Einmal mehr verlässt jemand mit dramatischen Worten die KI-Branche: Ein Forscher behauptet, dass in den Firmen geglaubt wird, dass „KI uns … | Einmal mehr verlässt jemand mit dramatischen Worten die KI-Branche: Ein Forscher behauptet, dass in den Firmen geglaubt wird, dass „KI… | ↗Not briefed |
bluesky bluesky-ai | 2mo | かいた。一線を越えている スレッズの自称「神仏祈祷師|霊視占い師」がAIを使って極めて創作っぽい話を流布してた - 電脳塵芥 nou-yunyun.hatenablog.com/entry/2026/0... | かいた。一線を越えている スレッズの自称「神仏祈祷師|霊視占い師」がAIを使って極めて創作っぽい話を流布してた - 電脳塵芥 nou-yunyun.hatenablog.com/entry/2026/0... | ↗Not briefed |
bluesky bluesky-ai | 2mo | Met de razendsnelle opkomst van potentieel levensbedreigende #AI systemen zou een debat gevoerd moeten worden over ons democratisch model m… | Met de razendsnelle opkomst van potentieel levensbedreigende #AI systemen zou een debat gevoerd moeten worden over ons democratisch mo… | ↗Not briefed |
bluesky bluesky-ai | 2mo | The Anthropic comms team, yesterday | The Anthropic comms team, yesterday | ↗Not briefed |
arxiv arxiv-ai | 2mo | CARRE: Counterfactual Action Retrieval and Reason Evaluation for Explainable Churn Prescription | Churn models typically identify high-risk customers but do not specify which feasible retention action should be considered or why tha… | ↗Not briefed |
hn hn-ai | 2mo | GPT-6 Astra creating GTA San Andreas mods, editing the map and 3D models | — | ↗Not briefed |
arxiv arxiv-ai | 2mo | SocialRL: Refining LLMs' Social Intelligence through Multi-turn Reinforcement Learning and Reward Design | Social intelligence enables agents to read social context, infer intent, and adapt over sustained dialogue. As language models become… | ↗Not briefed |
arxiv arxiv-ai | 2mo | MethaneFuse: Learning from Multi-Sensor Satellite Observations for Methane Plume Detection | Methane plume detection from satellite imagery is constrained by incomplete observations: public satellites provide complementary spat… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Anthropic doesn’t have enough compute to service demand; OAI may be demand-bound instead. if so then they’re in a position where they have… | Anthropic doesn’t have enough compute to service demand; OAI may be demand-bound instead. if so then they’re in a position where they… | ↗Not briefed |
bluesky bluesky-ai | 2mo | another day working at anthropic. everyone keeps asking me if they can fuck the frontier model. buddy, they wont even let me fuck it | another day working at anthropic. everyone keeps asking me if they can fuck the frontier model. buddy, they wont even let me fuck it | ↗Not briefed |
bluesky bluesky-ai | 2mo | I’d like to know who is on the Anthropic comms team | I’d like to know who is on the Anthropic comms team | ↗Not briefed |