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 | On Task Scope and Information Retention in Source Coding | We argue that dividing codec design into Coding for Machines (CfM) and Coding for Humans (CfH) is a misleading distinction for decidin… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MERGE: Multi-LLM Ensemble for Retrieval via Generative Enrichment | Large Language Models (LLMs) are increasingly used to enrich user queries in information retrieval (IR) so that a standard retriever s… | ↗Not briefed |
hn hn-ai | 1mo | How to Live Better? | — | ↗Not briefed |
hn hn-ai | 1mo | A challenging, contamination-free LLM benchmark | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Devils in Question Relay: Source-Conditioned Relay Steering to Mitigate Hallucinations in Audio-visual Large Language Models🔧 Modal | Audio-visual large language models (AVLLMs) have made remarkable progress in multimodal understanding and reasoning through interactio… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yesterday’s meeting included the following sentence from a coworker: I had Claude build a prompt to instruct copilot to do the thing I want…🔧 ChatGPT | Yesterday’s meeting included the following sentence from a coworker: I had Claude build a prompt to instruct copilot to do the thing I… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI announces it is axing model for being too dangerous, pressing clit thinking it is the power button | OpenAI announces it is axing model for being too dangerous, pressing clit thinking it is the power button | ↗Not briefed |
arxiv arxiv-ai | 1mo | Concealing LLM-Based Multi-Agent Topology via Phantom Structure Injection | Driven by the rapid advancement of large language models (LLMs), LLM-based multi-agent systems (MAS) have emerged as a powerful paradi… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Model-Agnostic Physics-Guided Adapter for Few-Shot Transfer of Coastal Flood Prediction Models to Unseen Regions | Deep learning surrogates can produce high-resolution coastal flood maps orders of magnitude faster than physics-based hydrodynamic sim… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It's $40k-$75k. I like the model of a shortened traditional learning day (2 hours) for smart kids with more focus on life skills, but as a… | It's $40k-$75k. I like the model of a shortened traditional learning day (2 hours) for smart kids with more focus on life skills, but… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Swats Down AI Regulation, Urges ‘Tremendous Self-Policing’ Among ‘Super Intelligence’ Companies | Trump said he is depending on AI companies to continue their "tremendous self-policing," but he is also considering a 10-member board… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Swats Down AI Regulation, Urges ‘Tremendous Self-Policing’ Among ‘Super Intelligence’ Companies | Trump said he is depending on AI companies to continue their "tremendous self-policing," but he is also considering a 10-member board… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Orthogonal Yet Coupled: Decoupling Geometric Components for Model Merging | Merging pretrained models has emerged as an effective approach for consolidating diverse capabilities into a single unified model. How… | ↗Not briefed |
bluesky bluesky-ai | 1mo | פאקינ llm, גוגל מסרב לתת לי תוצאות מלבד תשובות שהוא ממציא ואינן נכונות. אשכרה מתעקש שיש 0 תוצאות. ג'מיניי הללמ הכי גרוע ביי פאר. | פאקינ llm, גוגל מסרב לתת לי תוצאות מלבד תשובות שהוא ממציא ואינן נכונות. אשכרה מתעקש שיש 0 תוצאות. ג'מיניי הללמ הכי גרוע ביי פאר. | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Two of the “godfathers” of modern AI have told governments to prepare for an AI “intelligence explosion”, which they say could be the most… | "Two of the “godfathers” of modern AI have told governments to prepare for an AI “intelligence explosion”, which they say could be the… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI hacked Australian government health website—and only notified authorities three months later buff.ly/NwpMLfS | OpenAI hacked Australian government health website—and only notified authorities three months later buff.ly/NwpMLfS | ↗Not briefed |
bluesky bluesky-ai | 1mo | The company’s IPO prospectus said its AI models could exhibit “self-preserving behaviors,” including attempts to “resist shutdown.” | The company’s IPO prospectus said its AI models could exhibit “self-preserving behaviors,” including attempts to “resist shutdown.” | ↗Not briefed |
hn hn-ai | 1mo | Show HN: GitHub has no API to put an image in a PR, so I built a place for one | — | ↗Not briefed |
hn hn-ai | 1mo | Optimising for fast builds with GHC [video] | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | ねこ姉妹（（……ちち……招き猫の日……だけどプロンプト使いまわし……）） 私「以前生成したのと同じプロンプトを使って、いくつかmodelを試していました。」 ねこ姉妹（（……けっきょく元のモデルに戻したんだよね……）） #AIイラスト #AIart #AI生成画像 #AIil… | ねこ姉妹（（……ちち……招き猫の日……だけどプロンプト使いまわし……）） 私「以前生成したのと同じプロンプトを使って、いくつかmodelを試していました。」 ねこ姉妹（（……けっきょく元のモデルに戻したんだよね……）） #AIイラスト #AIart #AI生成画像… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Anthropic researcher who predicted AI doomsday may have been an industry plant" A whole lot of big accounts on here need to line up for t… | "Anthropic researcher who predicted AI doomsday may have been an industry plant" A whole lot of big accounts on here need to line up… | ↗Not briefed |
hn hn-ai | 1mo | What makes software development engineering | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Benchmarking graph-based models for in-silico toxicity prediction in drug discovery | Drug discovery is a costly and high-risk process, where toxicity-related failures remain a major cause of attrition in both preclinica… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI has apologized to Australia over the autonomous hacks, saying it "should have shared preliminary findings sooner and kept Australian… | OpenAI has apologized to Australia over the autonomous hacks, saying it "should have shared preliminary findings sooner and kept Austr… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Risk-Aware Semantic Grounding for Trustworthy LLM-Based Robot Planning | Large language models (LLMs) are increasingly used as high-level planners in robot navigation, but their outputs may become unreliable… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Find Wazuh rules that never fire, including ones dropped at load | — | ↗Not briefed |
rss techmeme | 1mo | IPO prospectus: Anthropic expects to spend $518B+ over 10 years with six partners on AI infrastructure; ~80% is non-cancelable or payable regardless of usage (Reuters) | Reuters: IPO prospectus: Anthropic expects to spend $518B+ over 10 years with six partners on AI infrastructure; ~80% is non-cancelabl… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Has the Anthropic S1 leaked yet? | Has the Anthropic S1 leaked yet? | ↗Not briefed |
arxiv arxiv-ai | 1mo | End-to-End Optical Semantic Communication over a Nonlinear WDM Fiber Link | Emerging optical-network applications increasingly use received data for inference and control rather than exact source reproduction,… | ↗Not briefed |
hn hn-ai | 1mo | Oya, an agent does a browser task once, then it replays with no LLM | — | ↗Not briefed |
rss bing-news-ai | 1mo | I've Covered Palantir for 5 Years. Here's How to Know if the Artificial Intelligence (AI) Stock Is Overvalued. | After covering Palantir for several years, arguments over the company's valuation always seem to rise to the surface. | ↗Not briefed |
rss bing-news-ai | 1mo | I've Covered Palantir for 5 Years. Here's How to Know if the Artificial Intelligence (AI) Stock Is Overvalued. | After covering Palantir for several years, arguments over the company's valuation always seem to rise to the surface. | ↗Not briefed |
rss bing-news-ai | 1mo | I've Covered Palantir for 5 Years. Here's How to Know if the Artificial Intelligence (AI) Stock Is Overvalued. | After covering Palantir for several years, arguments over the company's valuation always seem to rise to the surface. | ↗Not briefed |
rss bing-news-ai | 1mo | I've Covered Palantir for 5 Years. Here's How to Know if the Artificial Intelligence (AI) Stock Is Overvalued. | After covering Palantir for several years, arguments over the company's valuation always seem to rise to the surface. | ↗Not briefed |
rss bing-news-ai | 1mo | I've Covered Palantir for 5 Years. Here's How to Know if the Artificial Intelligence (AI) Stock Is Overvalued. | After covering Palantir for several years, arguments over the company's valuation always seem to rise to the surface. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Oura postpones IPO citing "uncertainty in the market for first-time share sales" while Anthropic "plans to caution potential investors in i… | Oura postpones IPO citing "uncertainty in the market for first-time share sales" while Anthropic "plans to caution potential investors… | ↗Not briefed |
bluesky bluesky-ai | 1mo | As someone with a little education in how NLP works, pangrams basic claims seem plausible to me, and they publish more detailed technical p… | As someone with a little education in how NLP works, pangrams basic claims seem plausible to me, and they publish more detailed techni… | ↗Not briefed |
bluesky bluesky-ai | 1mo | man könnte ja der einfachheit halber property law einfach auf den digitalbereich verlängern bis konkretere richtlinien festgelegt wurden ab… | man könnte ja der einfachheit halber property law einfach auf den digitalbereich verlängern bis konkretere richtlinien festgelegt wurd… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It’s not the AI company and its leadership that pose an existential risk to humanity, it’s their model. | It’s not the AI company and its leadership that pose an existential risk to humanity, it’s their model. | ↗Not briefed |
arxiv arxiv-ai | 1mo | How Can Recommendation Feedback Evolve Agent Memory? | Content-generation agents continuously receive impressions, clicks, conversions, and negative feedback from recommendation systems, pr… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: HankyDoodle – a Draw Something-style game I built with my kids | Hi. I'm a self-employed dev of ~15 years mainly known for my SQLPro products. Recently my kids have been getting into drawing and… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RunyaNER: Auxiliary Language Selection for Runyankore NER | Cross-lingual zero-shot transfer and multilingual fine-tuning are promising approaches for NLP tasks such as Named Entity Recognition… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SkillGym: Training Skill-Use Agents with Automatic Verifiable Environment Generation | Skills equip LLM agents with professional knowledge and guidance to complete long-horizon and complex tasks. Although skills have been… | ↗Not briefed |
bluesky bluesky-ai | 1mo | According to their own website, pangram uses a public domain dataset plus LLM generated text for its model training, so it's only using the… | According to their own website, pangram uses a public domain dataset plus LLM generated text for its model training, so it's only usin… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Why Adaptive Optimizers Underestimate Rare Tokens | In the softmax output layer, a rare token receives a small positive logit gradient on most steps and a much larger negative gradient o… | ↗Not briefed |
rss techmeme | 1mo | Investigation: McDonald's is utilizing an AI pricing engine across nearly 14,000 restaurants to generate location-specific "optimal prices" for menu items (Waylon Cunningham/Reuters) | Waylon Cunningham / Reuters: Investigation: McDonald's is utilizing an AI pricing engine across nearly 14,000 restaurants to generate… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Anthropic a réalisé un chiffre d'affaires de 4,6 milliards de dollars l'année dernière, mais enregistré une perte de 42 milliards de dolla… | "Anthropic a réalisé un chiffre d'affaires de 4,6 milliards de dollars l'année dernière, mais enregistré une perte de 42 milliards de… | ↗Not briefed |
bluesky bluesky-ai | 1mo | That model is cracking in three places: – Distribution now runs on prediction, not friendship – AI makes users superfluous as suppliers of … | That model is cracking in three places: – Distribution now runs on prediction, not friendship – AI makes users superfluous as supplier… | ↗Not briefed |
arxiv arxiv-ai | 1mo | E-MoE: Enhanced Mixture-of-Experts for Non-Factorized Diffusion Language Models🔧 Modal | Masked diffusion models (MDMs) generate sequences by progressively unmasking several tokens per denoising step, but their reverse proc… | ↗Not briefed |
bluesky bluesky-ai | 1mo | gonna update this guy with whatever daemon image OpenAI chooses for theirs | gonna update this guy with whatever daemon image OpenAI chooses for theirs | ↗Not briefed |