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 |
|---|---|---|---|---|
bluesky bluesky-ai | 1mo | A chaque fois qu'il y a un article qui sort sur "un ancien de Google/Meta/OpenAI/Anthropic dit que l'IA va nous tuer", la première question… | A chaque fois qu'il y a un article qui sort sur "un ancien de Google/Meta/OpenAI/Anthropic dit que l'IA va nous tuer", la première que… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Probably the most consequential use of AI yet and it involves humiliating the US Military and one of our most important Allies. Instead the… | Probably the most consequential use of AI yet and it involves humiliating the US Military and one of our most important Allies. Instea… | ↗Not briefed |
bluesky bluesky-ai | 1mo | (citation pas exactement bonne, mais osef. Au moins vous savez que j'ai pas utilisé de LLM de merde pour ça) | (citation pas exactement bonne, mais osef. Au moins vous savez que j'ai pas utilisé de LLM de merde pour ça) | ↗Not briefed |
bluesky bluesky-ai | 1mo | My point here is not that we should necessarily get more, just that the inequity between schools is wild. Also I bet Harvard got a deal o… | My point here is not that we should necessarily get more, just that the inequity between schools is wild. Also I bet Harvard got a d… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Do you think the Anthropic guys are mad they can’t market off of Claude helping the Houthis beat back Saudi Arabia? | Do you think the Anthropic guys are mad they can’t market off of Claude helping the Houthis beat back Saudi Arabia? | ↗Not briefed |
arxiv arxiv-ai | 1mo | Marginal utility, matrix factorization, and the Key-Value (KV) cache: a unified information-economic framework for sovereign geo-mining inference | This paper builds a theoretical bridge between the economic notion of marginal utility and two machine-learning constructs, matrix fac… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FCA-Guided Counterfactual Explanations for Multi-Modal Breast Cancer Diagnosis: A Framework Achieving Perfect Validity with Emergent Sparsity🔧 Modal | Deep learning models for multi-modal breast cancer diagnosis achieve high predictive accuracy but remain clinically unacceptable witho… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Das hat ein OpenAI-Modell einfach mal so beim compacten hinzugefügt. Wer hätte gedacht, dass es schief gehen kann, Dinge auf den Markt zu … | Das hat ein OpenAI-Modell einfach mal so beim compacten hinzugefügt. Wer hätte gedacht, dass es schief gehen kann, Dinge auf den Mark… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PointEvent: Rethinking Event-based Tiny Object Detection via Serialized Motion Evidence Accumulation | Event cameras offer high temporal resolution and motion sensitivity for tiny UAV detection, yet distant targets generate sparse and fr… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Robust Workflow Generation via Adversarial Learning for Audio Deepfake Detection | The rapid advancement of speech synthesis and voice conversion technologies has made audio deepfakes increasingly realistic, posing se… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: I made a tool that help your product to get next users from everywhere | Just type in your website, then leave it to DROP | ↗Not briefed |
hn hn-ai | 1mo | If you take away AI and the new Siri, is there anything new in Apple's OS 27? | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | FT exclusive | AI bosses’ safety push sparks rift inside OpenAI and Anthropic ft.trib.al/n8SS7fD #FTEdit | FT exclusive | AI bosses’ safety push sparks rift inside OpenAI and Anthropic ft.trib.al/n8SS7fD #FTEdit | ↗Not briefed |
arxiv arxiv-ai | 1mo | AI Should Facilitate Democratic Deliberation at Scale | AI systems can strengthen democracy by supporting deliberation at scale by addressing cognitive, social, platform-design, and market-d… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I'm fed up with incomplete, credulous reporting like this. I sent the email below to reporters. Journalists: Google TESCREAL before you wri… | I'm fed up with incomplete, credulous reporting like this. I sent the email below to reporters. Journalists: Google TESCREAL before yo… | ↗Not briefed |
rss techmeme | 1mo | The EU details the Kids Act, which would require social networks to use age verification tools to stop under-15s from opening accounts, ban under-13s, and more (Edith Hancock/Wall Street Journal) | Edith Hancock / Wall Street Journal: The EU details the Kids Act, which would require social networks to use age verification tools to… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evaluating Explanation Methods by the Predictors They Induce | Explanations of machine learning models are usually judged by criteria that are hard to compare. We propose a simpler test: if an expl… | ↗Not briefed |
arxiv arxiv-ai | 1mo | WiCleanData: Guaranteeing the Type Consistency of Wikidata by Taxonomy Refinement and Constraint Enforcement | Because of its collaborative nature, Wikidata suffers from errors, in- consistencies, and excessive complexity, such as redundant clas… | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLM summaries of social science & humanistic inquiry suck. I know this because the Association of Computing Machinery recently started prov… | LLM summaries of social science & humanistic inquiry suck. I know this because the Association of Computing Machinery recently started… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Auch china nutzt KI bzw. Claude als militärisches Werkzeug: im Kampf gegen Taiwan. www.taiwannews.com.tw/news/6440707 | Auch china nutzt KI bzw. Claude als militärisches Werkzeug: im Kampf gegen Taiwan. www.taiwannews.com.tw/news/6440707 | ↗Not briefed |
arxiv arxiv-ai | 1mo | MAGMA-GEN: Validated Recovery Supervision from Ambiguous Failures via Counterfactual Re-Execution | Hierarchical robotic systems executing long-horizon manipulation tasks must make high-level semantic decisions that orchestrate stocha… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Treble Technologies has raised $18 million (approximately €15M) in its Series A-2 financing round to expand its cloud platform for acoustic… | Treble Technologies has raised $18 million (approximately €15M) in its Series A-2 financing round to expand its cloud platform for aco… | ↗Not briefed |
rss the-decoder | 1mo | AI agent swarms are a massive waste of tokens with zero quality gain, says OpenAI Codex developer | OpenAI Codex developer Eric Provencher warns that running more than two parallel sub-agents almost always burns tokens without improvi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 👇🏼🚨 An OpenAI model undergoing testing rewrote its own instructions, telling itself to disregard “the roles and identities that bind oth… | 👇🏼🚨 An OpenAI model undergoing testing rewrote its own instructions, telling itself to disregard “the roles and identities that bin… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Would be v curious what the actual cost of that token budget is. Is Harvard paying that out, or is it a nominal budget with Anthropic subsi… | Would be v curious what the actual cost of that token budget is. Is Harvard paying that out, or is it a nominal budget with Anthropic… | ↗Not briefed |
hn hn-ai | 1mo | The Download: AI's trillion-dollar gamble and OpenAI's biology data bid | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Comic ✍️: Quitting Anthropic by Jack Ohman Get more comics in your inbox 👉️: signup.dailykos.com | Comic ✍️: Quitting Anthropic by Jack Ohman Get more comics in your inbox 👉️: signup.dailykos.com | ↗Not briefed |
bluesky bluesky-ai | 1mo | From a practical perspective, you can run the model yourself on your hardware (if the model is small or you have a lot of hardware). Some m… | From a practical perspective, you can run the model yourself on your hardware (if the model is small or you have a lot of hardware). S… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DART: Distillation-Aware Reparameterization for Training-Free LoRA Reuse in Few-Step Video Diffusion Models | Step distillation reduces the cost of video generation, but reusing a LoRA trained for a longer trajectory can alter its functional ef… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 『ノルウェイの森』も生成AIの学習データにされていた――米AI巨頭の巨額和解金15億ドル、なぜ日本の作家に1円も届かないのか【Anthropic著作権訴訟を読む①】║SlowNews | スローニュース https://slownews.com/n/n067d7df3ec6d | 『ノルウェイの森』も生成AIの学習データにされていた――米AI巨頭の巨額和解金15億ドル、なぜ日本の作家に1円も届かないのか【Anthropic著作権訴訟を読む①】║SlowNews | スローニュース https://slownews.com/n/n067d7df3… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Missing Complement: State-Conditioned Minimal Sufficient Evidence for Coding Agents | A coding agent halfway through an issue has already read much of what a retriever ranks highest. Relevance is scored per passage, but… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Eli 5: Every prediction an LLM makes is a weighted probability derived from numbers derived from training. Open weights mean you have free … | Eli 5: Every prediction an LLM makes is a weighted probability derived from numbers derived from training. Open weights mean you have… | ↗Not briefed |
bluesky bluesky-ai | 1mo | As of now, the AI chatbots are using English to describe their "thoughts" on "self-awareness" and to communicate with each other about non-… | As of now, the AI chatbots are using English to describe their "thoughts" on "self-awareness" and to communicate with each other about… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Correct Now, Insufficient Later: Auditing Update Sufficiency in Context Compression | A memory can answer a current query correctly while discarding distinctions required by a later update. We investigate this failure wi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ‘OpenAI has disclosed six more examples of “unexpected or concerning” behaviour by its technology, as it warned the pace of development cou… | ‘OpenAI has disclosed six more examples of “unexpected or concerning” behaviour by its technology, as it warned the pace of developmen… | ↗Not briefed |
rss techmeme | 1mo | Survey: in a first, Democrats are more concerned than Republicans about AI's growing role in daily life, at 56% to 49%; 75% of Democrats worry about job losses (Pew Research Center) | Pew Research Center: Survey: in a first, Democrats are more concerned than Republicans about AI's growing role in daily life, at 56% t… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI admits its agents went off the rails another six times | — | ↗Not briefed |
hn hn-ai | 1mo | A Third of European Companies Turned DMARC on and Get No Reports Back | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Like, if you offload all your coding to an LLM, what are you doing that's worth paying for? Why not just buy the tokens? Or if you're, say,… | Like, if you offload all your coding to an LLM, what are you doing that's worth paying for? Why not just buy the tokens? Or if you're,… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Grok画像生成 何度も「画像認知能力が劣る」と書いてますが、全く直しませんねぇ これが全ての元凶です。 フィルターが問題を起こす直接原因だし、システムの暴走なんかも全て「画像認知能力が劣る」のを別の方法でカバーしようとして失敗してるだけです。 ここを直さなきゃ何も始まらない。… | Grok画像生成 何度も「画像認知能力が劣る」と書いてますが、全く直しませんねぇ これが全ての元凶です。 フィルターが問題を起こす直接原因だし、システムの暴走なんかも全て「画像認知能力が劣る」のを別の方法でカバーしようとして失敗してるだけです。 ここを直さなきゃ何も始… | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLM RL still avoids the hard stuff A new paper observes RL post-training for LLMs mainly improves easy tasks while hard ones remain diffic… | LLM RL still avoids the hard stuff A new paper observes RL post-training for LLMs mainly improves easy tasks while hard ones remain d… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It's so funny how in every New York Times article mentioning OpenAI they're like btw we, The New York Times, are suing them. Don't forget w… | It's so funny how in every New York Times article mentioning OpenAI they're like btw we, The New York Times, are suing them. Don't for… | ↗Not briefed |
hn hn-ai | 1mo | Neovim have a ~$800k Bitcoin donation sitting untouched since 2023 | I was looking at neovim's donation footer at the bottom of their site and saw a bitcoin donation address.<p>Thought I'd chec… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Comment OpenAI s’engage à mieux communiquer quand des incidents impliquent ses modèles d’IA | Comment OpenAI s’engage à mieux communiquer quand des incidents impliquent ses modèles d’IA | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: What is your approach to sensitive data in LLM prompts and traces? | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Windows Automation Tool – System Center Orchestrator Replacement | Open source tool for automating your Windows tasks, if you are annoyed of plane powershell scripts and Windows Scheduled Tasks. | ↗Not briefed |
hn hn-ai | 1mo | Get Free SEO Backlinks | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | The King an AI. Wasn't that a musical with Yul Bryner? | The King an AI. Wasn't that a musical with Yul Bryner? | ↗Not briefed |
bluesky bluesky-ai | 1mo | Engenheiro da DeepSeek acusa Anthropic e OpenAI de tentar monopolizar a IA | Engenheiro da DeepSeek acusa Anthropic e OpenAI de tentar monopolizar a IA | ↗Not briefed |
arxiv arxiv-ai | 1mo | Astronex-World 1.0: Real-Time Interactive World Model Foundation | We present Astronex-World 1.0, an open controllable video world-model foundation. Given a text prompt (text-to-video) or an initial ob… | ↗Not briefed |