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 | 1mo | Grammatical "grandmother neurons" are rare in LLMs | Understanding how Large Language Models (LLMs) encode linguistic structures remains a fundamental challenge in interpretability resear… | ↗Not briefed |
arxiv arxiv-ai | 1mo | GCUL: Ambiguity Identification in Text Emotion Classification via Cluster-Guided Learning | Selective classification enables a model to abstain from predictions on uncertain instances, but existing approaches typically reject… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 120 secondes de Tech / 24 sept 2026 - Anthropic et OpenAI relancent la course - Google muscle la synthèse vocale - SpaceXAI lance Grok 4.7…🔧 Grok 4 | 120 secondes de Tech / 24 sept 2026 - Anthropic et OpenAI relancent la course - Google muscle la synthèse vocale - SpaceXAI lance Gro… | ↗Not briefed |
arxiv arxiv-ai | 1mo | TinyCardioUNet: IMU-to-ECG Translation with Graph-Encoded Inter-Axis Dependencies and Tensor Decomposition-Based Parameter Reduction | Estimating electrocardiography (ECG) from a chest-worn inertial measurement unit (IMU) enables continuous heart rate (HR) monitoring w… | ↗Not briefed |
rss stratechery | 1mo | An Interview with Colossus EIC Jeremy Stern About Profiling Mark Zuckerberg | An interview with Colossus EIC Jeremy Stern about profiling Mark Zuckerberg and other prominent tech figures. | ↗Not briefed |
hn hn-ai | 1mo | OpenAI agents 'infiltrated Australian government website' | — | ↗Not briefed |
rss techmeme | 1mo | Delivery Hero co-founder and CEO Niklas Östberg reverses his decision to step down in 2027, as the company prepares for Uber's €13B takeover and an asset sale (Financial Times) | Financial Times: Delivery Hero co-founder and CEO Niklas Östberg reverses his decision to step down in 2027, as the company prepares f… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Glaube, der Pfad wurde verlassen, als das Wort "Intelligenz" in Spiel kam. Stimme grundsätzlich zu. Präzision wäre hier sehr wünschenswert… | Glaube, der Pfad wurde verlassen, als das Wort "Intelligenz" in Spiel kam. Stimme grundsätzlich zu. Präzision wäre hier sehr wünschen… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Sécurité / OpenAI élargit l'accès de l'Ukraine aux outils de cybersécurité civils | Sécurité / OpenAI élargit l'accès de l'Ukraine aux outils de cybersécurité civils | ↗Not briefed |
arxiv arxiv-ai | 1mo | Neuralized Multi-Wavelet Decomposition for Time Series Classification and Forecasting | Time series analysis is fundamental in domains such as finance, healthcare, and meteorology. Real-world time series often exhibit mult… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: FreshThread – context monitor and handoffs for Codex Desktop (Windows) | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | When No One Owns the Judgment: Accountability Under Contribution Dissolution in Human-AI Collaboration | Communities often respond to potentially AI-assisted work by asking three questions: Was AI used? Was that use disclosed? Can hidden u… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DocuTeam: Mixed-Initiative Multi-Agent Discussions around Evolving Documents | In open-ended problem solving, collaborators often rely on discussion to surface concerns, challenge perspectives, and refine shared w… | ↗Not briefed |
bluesky bluesky-ai | 1mo | theconversation.com/an-openai-ag... | theconversation.com/an-openai-ag... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond Feature Reliability: Repeat-Informed Multifractal Curve Regression for Brain-Age Prediction | Brain-age prediction from resting-state fMRI provides a quantitative framework for characterizing age-related changes in spontaneous b… | ↗Not briefed |
bluesky bluesky-ai | 1mo | This "Warden" AI is attaching questions to all sorts of post by many different accounts, presumably in order to populate a LLM without cons… | This "Warden" AI is attaching questions to all sorts of post by many different accounts, presumably in order to populate a LLM without… | ↗Not briefed |
rss techmeme | 1mo | Brookings projects US AI spending to total $10.3T between 2025 and 2032, averaging 3.6% of GDP per year, the largest single-industry build-out as a share of GDP (Wall Street Journal) | Wall Street Journal: Brookings projects US AI spending to total $10.3T between 2025 and 2032, averaging 3.6% of GDP per year, the larg… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Correction: OpenAI trained its model to hack the Australian government. When will this insanity end? It's not an autonomous being, this is … | Correction: OpenAI trained its model to hack the Australian government. When will this insanity end? It's not an autonomous being, thi… | ↗Not briefed |
rss techmeme | 1mo | Sources: DeepSeek's annualized revenue run rate hits $1B, up from less than $500M a few months ago, as it aims to finalize a ~$7.5B fundraise by late October (The Information) | The Information: Sources: DeepSeek's annualized revenue run rate hits $1B, up from less than $500M a few months ago, as it aims to fin… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Sufficiently Reduced Distributional Regression | We propose Sufficiently Reduced Distributional Regression (SRDR), a generative method that combines conditional distribution estimatio… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI社のAIエージェントが、オーストラリア政府の国民健康保険サイトに不正アクセスし、非公開データにアクセスした事件が発生したそうですね。 これは由々しき事態ですね。 gigazine.net/news/2026073... 先日も、生成AI大手各社が開発中のAIシステ… | OpenAI社のAIエージェントが、オーストラリア政府の国民健康保険サイトに不正アクセスし、非公開データにアクセスした事件が発生したそうですね。 これは由々しき事態ですね。 gigazine.net/news/2026073... 先日も、生成AI大手各社が開発中の… | ↗Not briefed |
hn hn-ai | 1mo | Happy, those able to know the causes of things | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Oh yeah, all these attacks are definitely just accidents and not PR opportunities and/or threats. This is in no way linked to Australia t… | Oh yeah, all these attacks are definitely just accidents and not PR opportunities and/or threats. This is in no way linked to Austra… | ↗Not briefed |
hn hn-ai | 1mo | I reverse-engineered AutoCAD's 3D format to convert DWG to STEP with Claude | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | why am i not hyped? (no one asks) - i don't have the nostalgia so i don't feel the need for this - beautiful areas etc + two pixel characte… | why am i not hyped? (no one asks) - i don't have the nostalgia so i don't feel the need for this - beautiful areas etc + two pixel cha… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI hackea el sistema de salud australiano." Ese es el titular y como tal hay que tratarlo, una empresa estadounidense con fuertes vínc… | OpenAI hackea el sistema de salud australiano." Ese es el titular y como tal hay que tratarlo, una empresa estadounidense con fuertes… | ↗Not briefed |
arxiv arxiv-ai | 1mo | From Text Decisions to Pixels: An Study of Jev-Style Visual Choice Model | Visual software often needs a decision over supplied alternatives rather than a generated explanation. We present PixelJev, a native-i… | ↗Not briefed |
hn hn-ai | 1mo | How the cocktail got its name | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Online Task Adaptation via Self-Organisation | Neural networks are typically adapted by computing gradients and updating model parameters. We investigate whether task-specific adapt… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI presenteert AI die na elke inbraak zelfstandig excuses aanbiedt nieuwspaal.nl/openai-prese... | OpenAI presenteert AI die na elke inbraak zelfstandig excuses aanbiedt nieuwspaal.nl/openai-prese... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Reasoning Instructions Can Break Answer Decoding in Vision--Language Models | Chain-of-thought (CoT) instructions can distort multiple-choice VLM evaluation when a scorer appends a reasoning cue but reads answer-… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The AI industry believes that if you’re smart enough, you can cure cancer - or destroy the world. But that’s not how power and intelligence… | The AI industry believes that if you’re smart enough, you can cure cancer - or destroy the world. But that’s not how power and intelli… | ↗Not briefed |
rss techmeme | 1mo | Sources citing Firmus' IPO draft prospectus: the Australian data center operator expects to report a $77M loss after tax for H1 FY27 as it prepares to raise $5B (Christine Chen/Reuters) | Christine Chen / Reuters: Sources citing Firmus' IPO draft prospectus: the Australian data center operator expects to report a $77M lo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | but the llm policy was written in parkour | but the llm policy was written in parkour | ↗Not briefed |
arxiv arxiv-ai | 1mo | pylazaro: a Python package for anglicism extraction in Spanish | Lexical borrowings are words from one language that are introduced into another language. Identifying lexical borrowings in text is a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Tuesday morning my role was eliminated. By Wednesday breakfast I had shipped a video game about it. 🎮 I built it with Claude Opus 5.5 in … | Tuesday morning my role was eliminated. By Wednesday breakfast I had shipped a video game about it. 🎮 I built it with Claude Opus 5.… | ↗Not briefed |
hn hn-ai | 1mo | Arturia KeyStep 37 firmware research project | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | The SMH op ed on the OpenAI / Medicare issue is at least slightly nuanced | The SMH op ed on the OpenAI / Medicare issue is at least slightly nuanced | ↗Not briefed |
rss lastweekin-ai | 1mo | LWiAI Podcast #257 - GPT 6 Astra, AI Extinction, Security Incidents | A belated discussion of Astra’s release, a bunch of AI incidents, and the broader situation we are in with regards to AI safety. | ↗Not briefed |
hn hn-ai | 1mo | Search – A small, fast WebKit browser for macOS, by Office Commun | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I regret to inform you Van Badham has weighed in on the OpenAI Medicare issue | I regret to inform you Van Badham has weighed in on the OpenAI Medicare issue | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learnable Time-Frequency Masks for Explaining Time-Series Classifiers | Time-series explainability remains challenging because discriminative information is often encoded in latent frequency or time-frequen… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ALOE: Semantically Addressed Low-Rank Operators for Knowledge Editing | Knowledge editing changes what a model knows by modifying parameters so that a requested fact updates while unrelated behavior is pres… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 権利問題って一部の人にしか関係ないんでしょ？みたいな認識の人がおそらく大多数で、今や当たり前にアプリで入れられたりスマホに最初から備え付けられてるような生成AI(LLM)が直球ど真ん中で権利侵害問題の中心にある…なんて想像も付かない利用者達に「あなたも当事者なんですよ、むしろ加… | 権利問題って一部の人にしか関係ないんでしょ？みたいな認識の人がおそらく大多数で、今や当たり前にアプリで入れられたりスマホに最初から備え付けられてるような生成AI(LLM)が直球ど真ん中で権利侵害問題の中心にある…なんて想像も付かない利用者達に「あなたも当事者なんですよ… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The new frontier AI model should arrive long before the end of 2026. www.theverge.com/tech/999802/... | The new frontier AI model should arrive long before the end of 2026. www.theverge.com/tech/999802/... | ↗Not briefed |
arxiv arxiv-ai | 1mo | BridgeMem: Causal Dyadic Transition Residuals for Temporal Knowledge Graph Forecasting | Temporal knowledge graph forecasting aims to infer future relational facts from the temporal structure of observed events. Existing fo… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Baszta: Data-Centric Fine-Tuning of a Polish Multi-Label Safety Classifier | We develop a multi-label Polish content-safety classifier by fine-tuning allegro/herbert-base-cased (124M) across five categories (hat… | ↗Not briefed |
rss theverge-ai | 1mo | Gemini 4 is almost ready, says new Google DeepMind chief | Google is reportedly nearing the launch of its long awaited Gemini 4 model, after dawdling behind rival developers on flagship AI rele… | ↗Not briefed |
bluesky bluesky-ai | 1mo | In that same way if I’m running a test on an AI model to maybe test cyber security of a particular object but then give the sandbox access … | In that same way if I’m running a test on an AI model to maybe test cyber security of a particular object but then give the sandbox ac… | ↗Not briefed |
arxiv arxiv-ai | 1mo | TP-CRIV: A Framework for Third-Party Challenge-Response Identity Verification of AI Models | Artificial intelligence (AI) models are increasingly deployed through remote services, making model misappropriation a growing concern… | ↗Not briefed |