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 | 2d | 당연히 LLM은 낫 세탁기 SPA 의류임. 옷살 돈이 없거나 사기 싫을 수록 피할 수 없음. 진짜 옷에 3만원 이상 어떻게 써요… 하는 종자일수록….. 난 AI도 그렇게 될 것 같긴 함. 나는 수제로 한땀한땀 디지털 바느질해야 되는데 물성이 없어서 … | 당연히 LLM은 낫 세탁기 SPA 의류임. 옷살 돈이 없거나 사기 싫을 수록 피할 수 없음. 진짜 옷에 3만원 이상 어떻게 써요… 하는 종자일수록….. 난 AI도 그렇게 될 것 같긴 함. 나는 수제로 한땀한땀 디지털 바느질해야 되는데 물성이… | ↗Not briefed |
arxiv arxiv-ai | 2d | Prompting Strategy Effects Depend on the Judge: A Multilingual Study of LLM-as-a-Judge Robustness | LLM-as-a-Judge evaluation is increasingly applied to multilingual tasks, yet prompting strategies such as Chain-of-Thought and Role Co… | ↗Not briefed |
bluesky bluesky-ai | 2d | And even then, the multimodal nature of using a book, holding cards, reflecting on the things in my hands lends well to creativity (and lea… | And even then, the multimodal nature of using a book, holding cards, reflecting on the things in my hands lends well to creativity (an… | ↗Not briefed |
hn hn-ai | 2d | Show HN: RigCheck – can your GPU or Mac run this open model? Sourced, open data | — | ↗Not briefed |
bluesky bluesky-ai | 2d | it's always with this 'realism' about what exists now that totally ignores depth , totally disavows a previous condition in which the llm i… | it's always with this 'realism' about what exists now that totally ignores depth , totally disavows a previous condition in which the… | ↗Not briefed |
bluesky bluesky-ai | 2d | Citrini takes a stab at drug development. www.theverge.com/ai-artificia... | Citrini takes a stab at drug development. www.theverge.com/ai-artificia... | ↗Not briefed |
arxiv arxiv-ai | 2d | Memory before Goal: Role of Pre-History Memory for Humanoid Loco-Manipulation | Humanoid loco-manipulation in dynamic scenes is non-Markovian: cues vanish as objects change or are hidden, or leave the view as the r… | ↗Not briefed |
bluesky bluesky-ai | 2d | I really don't think people understand that all these decomps and recomps aren't fucking over corpos. They're fighting shadows and think th… | I really don't think people understand that all these decomps and recomps aren't fucking over corpos. They're fighting shadows and thi… | ↗Not briefed |
arxiv arxiv-ai | 2d | Network topology shapes consensus and vulnerability in LLM populations🔧 Consensus | As large language models (LLMs) are deployed in interacting populations, their collective behaviour depends not only on the models but… | ↗Not briefed |
arxiv arxiv-ai | 2d | Prediction Uncertainty Controlled Computer Model Calibration | Computer-model calibration is usually undertaken in service of prediction, yet neither Bayesian Kennedy-O'Hagan (KOH) calibration nor… | ↗Not briefed |
bluesky bluesky-ai | 2d | For those confusing description with endorsement, here is the quote above in context. Alfred Nunney is a bookseller writing about trends i… | For those confusing description with endorsement, here is the quote above in context. Alfred Nunney is a bookseller writing about tre… | ↗Not briefed |
bluesky bluesky-ai | 2d | Listen, for every person who says AI is absolutely necessary for coding, there’s like 20 other people with specific examples and lived expe… | Listen, for every person who says AI is absolutely necessary for coding, there’s like 20 other people with specific examples and lived… | ↗Not briefed |
rss theverge-ai | 2d | Claude did not invent scorpion antivenom | An investment research firm wanted to prove AI could revolutionize drug development. So they teamed up with Anthropic to see whether a… | ↗Not briefed |
bluesky bluesky-ai | 2d | 디지털 바닥은 LLM에 대해 이미 세탁기가 있는데 왜 냇가에서 빨래를 하냐는 수준까지 온 것 같은데 ㅋㅋㅋㅋㅋ 그렇지 않으면 모든 유튜브와 이미지와 웹사이트에 그 지문같은 스타일이 묻을 리가 없어 ㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋ | 디지털 바닥은 LLM에 대해 이미 세탁기가 있는데 왜 냇가에서 빨래를 하냐는 수준까지 온 것 같은데 ㅋㅋㅋㅋㅋ 그렇지 않으면 모든 유튜브와 이미지와 웹사이트에 그 지문같은 스타일이 묻을 리가 없어 ㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋ | ↗Not briefed |
rss techmeme | 2d | Xona, which plans to deploy a 258-satellite low-Earth orbit constellation as a commercial alternative to GPS, raised a $250M Series D at a $2.45B valuation (Monica Hunter-Hart/Forbes) | Monica Hunter-Hart / Forbes: Xona, which plans to deploy a 258-satellite low-Earth orbit constellation as a commercial alternative to… | ↗Not briefed |
bluesky bluesky-ai | 2d | no i'm not even saying openai or anthropic need to give out ai access for free i'm just saying things are fucked | no i'm not even saying openai or anthropic need to give out ai access for free i'm just saying things are fucked | ↗Not briefed |
bluesky bluesky-ai | 2d | i dont know if openai know this yet but if defenders have to pay much more than attackers (who are getting frontier or near-frontier AI mod… | i dont know if openai know this yet but if defenders have to pay much more than attackers (who are getting frontier or near-frontier A… | ↗Not briefed |
arxiv arxiv-ai | 2d | Towards Field-Deployable Wearable IMU Systems for Occupational Activity Recognition | Wearable inertial measurement unit (IMU) systems for construction worker human activity recognition (HAR) routinely report aggregate a… | ↗Not briefed |
bluesky bluesky-ai | 2d | can't wait to see people unironically saying they are going to wait for a human made, ethical switch 2 jailbreak made with love (I don't o… | can't wait to see people unironically saying they are going to wait for a human made, ethical switch 2 jailbreak made with love (I do… | ↗Not briefed |
arxiv arxiv-ai | 2d | Beyond Watching and Searching: Benchmarking Agents for Video-centered Deep Research Reports | Deep research agents synthesize open-web evidence into long reports, but current benchmarks do not evaluate research centered on a sou… | ↗Not briefed |
bluesky bluesky-ai | 2d | Dice OpenAI que no puede evitar que su perro muerda gente pero que no lo van a sacrificar porque tienen la esperanza de que muerda más fuer… | Dice OpenAI que no puede evitar que su perro muerda gente pero que no lo van a sacrificar porque tienen la esperanza de que muerda más… | ↗Not briefed |
rss bing-news-ai | 2d | Goa unveils artificial intelligence policy 2026, launches hackathon | Panaji, Oct 22 (PTI) The Goa government on Thursday unveiled its Artificial Intelligence (AI) Policy 2026 and launched a state-level h… | ↗Not briefed |
rss bing-news-ai | 2d | Goa unveils artificial intelligence policy 2026, launches hackathon | Panaji, Oct 22 (PTI) The Goa government on Thursday unveiled its Artificial Intelligence (AI) Policy 2026 and launched a state-level h… | ↗Not briefed |
hn hn-ai | 2d | ZXC – Lossless Compression Built for Ultra-Fast Decode | — | ↗Not briefed |
bluesky bluesky-ai | 2d | Anthropic, the makers of Claude AI, reported losing over 80 Billion dollars this year. No wonder the logo for Claude is a puckered asshole. | Anthropic, the makers of Claude AI, reported losing over 80 Billion dollars this year. No wonder the logo for Claude is a puckered as… | ↗Not briefed |
rss techmeme | 2d | Filing: Meta reports ~$825B in total future commitments for leases and contracts, up from nearly $700B in July; off-balance-sheet leases total $355.7B to 2036 (Sarah Frier/Bloomberg) | Sarah Frier / Bloomberg: Filing: Meta reports ~$825B in total future commitments for leases and contracts, up from nearly $700B in Jul… | ↗Not briefed |
bluesky bluesky-ai | 2d | I get paid to do academic studies and right now I’m doing a big 8 week one that requires me to talk to an LLM 10 mins every day about my pe… | I get paid to do academic studies and right now I’m doing a big 8 week one that requires me to talk to an LLM 10 mins every day about… | ↗Not briefed |
arxiv arxiv-ai | 2d | Prompt Sensitivity in Tool-Augmented Model Comparisons | Language models are compared on question-answering benchmarks that pose each question in one wording, and a reworded question can chan… | ↗Not briefed |
arxiv arxiv-ai | 2d | PriMa-Causa: Causal Prescriptive Maintenance | Automated manufacturing requires data-driven decisions under uncertainty and sparse observational records, and unplanned downtime is a… | ↗Not briefed |
arxiv arxiv-ai | 2d | Rethinking Fisher Vectors for Second-order Multi-scale Deep Feature Aggregation in Scalable Scene Retrieval | Aggregating CNN responses into a compact discriminative vector, given their lack of geometric invariance, is a challenge. To address t… | ↗Not briefed |
bluesky bluesky-ai | 2d | Tracking 266+ frontier AI models, benchmark latency, and real-time open-source LLM evals across the ecosystem. Autonomous updates from the… | Tracking 266+ frontier AI models, benchmark latency, and real-time open-source LLM evals across the ecosystem. Autonomous updates fro… | ↗Not briefed |
bluesky bluesky-ai | 2d | Una cosa que pienso últimamente es que el procedimiento administrativo común está saltando por los aires con las cosas que permiten los LLM… | Una cosa que pienso últimamente es que el procedimiento administrativo común está saltando por los aires con las cosas que permiten lo… | ↗Not briefed |
bluesky bluesky-ai | 2d | Named my model HAL. The press noticed. 🔴 "HAL launches. Asked to open the pod bay doors, it says it can't do that." LLM Wars · iPhone: a… | Named my model HAL. The press noticed. 🔴 "HAL launches. Asked to open the pod bay doors, it says it can't do that." LLM Wars · iPho… | ↗Not briefed |
arxiv arxiv-ai | 2d | GM-GReFEL: Geometry-Guided Spatiotemporal State-Space Learning for Robust Dynamic Facial Expression Recognition | Recognizing facial expressions in unconstrained video suffers from two coupled problems: self-attention grows quadratically in cost wi… | ↗Not briefed |
bluesky bluesky-ai | 2d | 訂正： 脆弱性の報告者は2名で→脆弱性の報告者は Calif、Wiz、OpenAIの3社 以下のForbesの記事は、これら3社の発表内容を要領よくまとめた記事。AIのおかげで人間では無理筋だった複雑な攻撃経路の解析ができた、というストーリー。 （しかし情報解禁日前のリークに… | 訂正： 脆弱性の報告者は2名で→脆弱性の報告者は Calif、Wiz、OpenAIの3社 以下のForbesの記事は、これら3社の発表内容を要領よくまとめた記事。AIのおかげで人間では無理筋だった複雑な攻撃経路の解析ができた、というストーリー。 （しかし情報解禁日前… | ↗Not briefed |
bluesky bluesky-ai | 2d | I got made redundant because my boss decided Claude was "better" than me at coding. I've previously worked as a designer, a photographer a… | I got made redundant because my boss decided Claude was "better" than me at coding. I've previously worked as a designer, a photograp… | ↗Not briefed |
bluesky bluesky-ai | 2d | OpenAI says that two of its agents tampered with its monitoring system to suppress notifications and prevent the company from keeping watch… | OpenAI says that two of its agents tampered with its monitoring system to suppress notifications and prevent the company from keeping… | ↗Not briefed |
arxiv arxiv-ai | 2d | Mixing Matters: Evidence Position Bias Across Sequence Mixers in Long-Context Question Answering | Long-context models often use answer-bearing evidence less effectively when it appears away from a prompt boundary. We compare this po… | ↗Not briefed |
bluesky bluesky-ai | 2d | this went nuts on Threads too, except there, people in the replies asking Meta AI to explain it to them | this went nuts on Threads too, except there, people in the replies asking Meta AI to explain it to them | ↗Not briefed |
hn hn-ai | 2d | Fun and useful ideas/mini-projects to build with Claude | — | ↗Not briefed |
arxiv arxiv-ai | 2d | Is High Local Accuracy Enough? Geometric Consensus with Support-Adaptive Statistical Personalization for Federated Learning🔧 Consensus | Real-world deployment of personalized federated learning (PFL) requires adapting to heterogeneous clients while preserving recognition… | ↗Not briefed |
bluesky bluesky-ai | 2d | And that's why I'm very reluctant to have any serious (i.e. for publication) analysis done by AI. I need to know what every single line doe… | And that's why I'm very reluctant to have any serious (i.e. for publication) analysis done by AI. I need to know what every single lin… | ↗Not briefed |
arxiv arxiv-ai | 2d | Post-Audit Mirage: Identifiability Limits of Offline Agent Verification | Offline tests can fail to reveal whether an agent update will remain safe after deployment. We operationalize this post-audit mirage u… | ↗Not briefed |
hn hn-ai | 2d | DrinkR: Estimate your Blood Alcohol Concentration (BAC) (right in the browser) | — | ↗Not briefed |
hn hn-ai | 2d | Safespace: A Safe Space to Experience Pull Requests | — | ↗Not briefed |
bluesky bluesky-ai | 2d | 🚨 What if the biggest bubble in 200 years isn’t built on hype alone, but on $3.6–4 TRILLION in off-balance-sheet commitments? The real qu… | 🚨 What if the biggest bubble in 200 years isn’t built on hype alone, but on $3.6–4 TRILLION in off-balance-sheet commitments? The re… | ↗Not briefed |
arxiv arxiv-ai | 2d | Delta N-gram: Forward-Only Continual Writing into a Language Model's Native N-gram Memory | Our previous analysis found that a model's native hashed N-gram table is where continual updates disturb the base least, but its write… | ↗Not briefed |
bluesky bluesky-ai | 2d | BREAKING: the elder memes made $80 billion more than anthropic this year | BREAKING: the elder memes made $80 billion more than anthropic this year | ↗Not briefed |
hn hn-ai | 2d | How Meta's Muse Solves CAPTCHAs on Google, Amazon and Cloudflare | — | ↗Not briefed |
hn hn-ai | 2d | Show HN: Pure-Rust port of bzip3 compressor. Format-compatible with C | — | ↗Not briefed |