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 | 21d | Which LLM should run your agentic analytics? Accuracy vs. cost of Claude, Gemini and GPT We tested 4 models on real e-commerce analytics q… | Which LLM should run your agentic analytics? Accuracy vs. cost of Claude, Gemini and GPT We tested 4 models on real e-commerce analyt… | ↗Not briefed |
arxiv arxiv-ai | 21d | LLMs Prompted for Legal Context Refuse More: Overrefusal from Small On-Premises LLMs in Criminal Legal Context | Confidentiality requirements increasingly push legal professionals toward small, on-device LLMs that keep case data in-house, used for… | ↗Not briefed |
arxiv arxiv-ai | 21d | OceanTensorCube and ST-MMT: A Multimodal Data and Transformer System for Seven-Day Forecasting of Pyropia Bleaching | Bleaching of cultured Pyropia (laver) causes losses in Korean seaweed farming. We present a system that forecasts bleaching seven days… | ↗Not briefed |
bluesky bluesky-ai | 21d | The most disruptive change to software due to AI coding agents is that effort is no longer a moat. It is now trivial to duplicate the fun… | The most disruptive change to software due to AI coding agents is that effort is no longer a moat. It is now trivial to duplicate th… | ↗Not briefed |
bluesky bluesky-ai | 21d | That's something an LLM would say.... | That's something an LLM would say.... | ↗Not briefed |
bluesky bluesky-ai | 21d | AI researchers stranded Claude on a virtual desert island. It started to “play.” www.technologyreview.com/2026/10/19/1... | AI researchers stranded Claude on a virtual desert island. It started to “play.” www.technologyreview.com/2026/10/19/1... | ↗Not briefed |
rss the-decoder | 21d | The neocloud shakeout is coming, and Nvidia has the most to lose | After stock crashes and Firmus's scrapped IPO, industry executives expect only seven to ten of the hundreds of AI cloud providers to s… | ↗Not briefed |
bluesky bluesky-ai | 21d | Mastodon accepts LLM-assisted contributions. | Mastodon accepts LLM-assisted contributions. | ↗Not briefed |
bluesky bluesky-ai | 21d | Economists love to talk about the market making sure supply rises to meet demand, but the thing about LLM and AI images and video is that b… | Economists love to talk about the market making sure supply rises to meet demand, but the thing about LLM and AI images and video is t… | ↗Not briefed |
hn hn-ai | 21d | AI firm Anthropic to commit 5% of shares to fund for 'public good' | — | ↗Not briefed |
bluesky bluesky-ai | 21d | Elon Musk pledged to distribute a share of SpaceX to each American child after competitor Anthropic PBC said it would set aside a portion o… | Elon Musk pledged to distribute a share of SpaceX to each American child after competitor Anthropic PBC said it would set aside a port… | ↗Not briefed |
rss techmeme | 21d | Analysis: companies are set for a massive acceleration of 50GW+ in data center capacity additions in 2027; 24GW+ across 40 campuses are in the advanced stages (Jeremie Eliahou Ontiveros/SemiAnalysis) | Jeremie Eliahou Ontiveros / SemiAnalysis: Analysis: companies are set for a massive acceleration of 50GW+ in data center capacity addi… | ↗Not briefed |
arxiv arxiv-ai | 21d | TaQFormer: A Treatment-as-Query Transformer with Data-Fusion for Multidimensional Treatment Modeling in Coupon Allocation | We study coupon allocation on an online marketplace offering multiple service categories. Issuing coupons requires choosing a vector o… | ↗Not briefed |
bluesky bluesky-ai | 21d | FloatDoor: LLMs trained to do malicious things but only on certain hardware this one uses differing behavior of floating point rounding to… | FloatDoor: LLMs trained to do malicious things but only on certain hardware this one uses differing behavior of floating point roundi… | ↗Not briefed |
arxiv arxiv-ai | 21d | GeoEdit: Adversarial De-Evidencing for Silent Geoprivacy Protection via Image Editing | Recent efforts have demonstrated the strong capabilities of Multimodal Large Reasoning Models (MLRMs) in geographic reasoning. By join… | ↗Not briefed |
bluesky bluesky-ai | 21d | Scale your workflow with Claude Managed Agents! Deploy autonomous agents that monitor Slack, manage GitHub repos via Git worktrees, and eve… | Scale your workflow with Claude Managed Agents! Deploy autonomous agents that monitor Slack, manage GitHub repos via Git worktrees, an… | ↗Not briefed |
bluesky bluesky-ai | 21d | Earlier this year I was seeking funding to investigate a hunch that Amazon was using AI for union-busting. Turns out that hunch was correct… | Earlier this year I was seeking funding to investigate a hunch that Amazon was using AI for union-busting. Turns out that hunch was co… | ↗Not briefed |
rss the-decoder | 21d | Anthropic cushions its upcoming IPO with a discretionary fund for workers its models might replace | Anthropic wants to put five percent of its shares into a new public benefit fund that will also look into direct payments to workers a… | ↗Not briefed |
hn hn-ai | 21d | Gitby – Stop fixing conflicts. Prevent them | — | ↗Not briefed |
arxiv arxiv-ai | 21d | VEILED-SH: Vulnerability Expression through Indirect Language Evaluation Dataset (Self-Harm) | AI systems may miss indirect, culturally situated distress or provide support mismatched to the application context. Existing benchmar… | ↗Not briefed |
arxiv arxiv-ai | 21d | On Counting Graphs for Modal Logic with Counting🔧 Modal | In this paper, we study the theoretical computational complexity of counting trees and graphs that are classified positively by a grap… | ↗Not briefed |
arxiv arxiv-ai | 21d | MoveBench: A Rubric-Based Benchmark for Evaluating Large Language Models in Sports and Health Coaching | Despite rapid advances in large language models (LLMs), their deployment in the sports and health domain remains high-risk due to the… | ↗Not briefed |
arxiv arxiv-ai | 21d | Reading Time Modeling with Pixel-Based Language Models | Tokenization-free, pixel-based language models (PLMs) have recently been proposed as an alternative to conventional token-based langua… | ↗Not briefed |
bluesky bluesky-ai | 21d | Anthropic is seeking advice from lawyers on how AI agents should conduct negotiations on behalf of users www.ft.com/content/f7c2... | Anthropic is seeking advice from lawyers on how AI agents should conduct negotiations on behalf of users www.ft.com/content/f7c2... | ↗Not briefed |
bluesky bluesky-ai | 21d | ...do you think i'm an LLM? | ...do you think i'm an LLM? | ↗Not briefed |
hn hn-ai | 21d | OpenPS2BIOS | — | ↗Not briefed |
bluesky bluesky-ai | 21d | Pedi pro deepseek descrever e até a IA chinesa terminou a descrição com "...em provável referência à manipulação de sessões eleitorais por … | Pedi pro deepseek descrever e até a IA chinesa terminou a descrição com "...em provável referência à manipulação de sessões eleitorais… | ↗Not briefed |
rss techmeme | 21d | Character.AI's Chapter 11 court filings indicate a two-part deal with Disney to transfer talent and tech while insulating Disney from mounting legal liabilities (Maddy Varner/Wired)🔧 Character.AI | Maddy Varner / Wired: Character.AI's Chapter 11 court filings indicate a two-part deal with Disney to transfer talent and tech while i… | ↗Not briefed |
bluesky bluesky-ai | 21d | I think it was this tweet, from the glorious pre-Musk pre-LLM era: | I think it was this tweet, from the glorious pre-Musk pre-LLM era: | ↗Not briefed |
hn hn-ai | 21d | How to build a SaaS product with Claude Code | — | ↗Not briefed |
arxiv arxiv-ai | 21d | Conditional information Retrieval for Longitudinal Survey Harmonisation | Identifying conceptually equivalent questions across different surveys is a challenging information retrieval task where surface-level… | ↗Not briefed |
hn hn-ai | 21d | Ask HN: What are the things LLM models fail at? | Generally, whenever I do see frontier AI model's evaluation metrics on say Artificial Analysis. It gives out a benchmark on a set… | ↗Not briefed |
bluesky bluesky-ai | 21d | ウクライナに軍事AIを提供している「パランティア」という企業は生成AI「Claude」を基盤としたAIを使っています。 パランティアはGMOと提携して「日本版パランティア構想」を練っていることで知ってる人も多いかと思います。 | ウクライナに軍事AIを提供している「パランティア」という企業は生成AI「Claude」を基盤としたAIを使っています。 パランティアはGMOと提携して「日本版パランティア構想」を練っていることで知ってる人も多いかと思います。 | ↗Not briefed |
bluesky bluesky-ai | 21d | www.opendemocracy.net/amazon-ai-su... | www.opendemocracy.net/amazon-ai-su... | ↗Not briefed |
arxiv arxiv-ai | 21d | VISAGENT: An Agent-Based End-to-End Voice Interaction System | Voice-based public services have evolved from manual switchboards to intelligent platforms to meet modern user needs. However, current… | ↗Not briefed |
bluesky bluesky-ai | 21d | But anyone who's been doing any flavor of activism for some length of time knows that motivation and enthusiasm are only the start. The bre… | But anyone who's been doing any flavor of activism for some length of time knows that motivation and enthusiasm are only the start. Th… | ↗Not briefed |
bluesky bluesky-ai | 21d | used it for half a decade then switched to 1337. but like many things LLM is using, it's because people are doing it, of course. | used it for half a decade then switched to 1337. but like many things LLM is using, it's because people are doing it, of course. | ↗Not briefed |
hn hn-ai | 21d | Asking minified JSON can reduce output token costs by 30% | — | ↗Not briefed |
bluesky bluesky-ai | 21d | DwarfStar GGUF for DeepSeek v4.1 Flash new checkpoint is now available. It is an improvement even only for the better token efficiency. Upg… | DwarfStar GGUF for DeepSeek v4.1 Flash new checkpoint is now available. It is an improvement even only for the better token efficiency… | ↗Not briefed |
rss techmeme | 21d | A look at Anthropic's Long-Term Benefit Trust, an unusual, opaque five-member trust whose trustees own no equity and hold enormous sway over the company (Theo Francis/Wall Street Journal) | Theo Francis / Wall Street Journal: A look at Anthropic's Long-Term Benefit Trust, an unusual, opaque five-member trust whose trustees… | ↗Not briefed |
bluesky bluesky-ai | 21d | This is going to become a genuine problem and IMO the key elements to think through it clearly were nailed by Kastel months ago: mimidoshim… | This is going to become a genuine problem and IMO the key elements to think through it clearly were nailed by Kastel months ago: mimid… | ↗Not briefed |
bluesky bluesky-ai | 21d | 日産の横須賀市追浜工場の跡地にアンドゥリル、という話が出たので… アンドゥリル(Andutil)という軍事企業にAIを提供している「OpenAI」という会社があります。 最近話題の「Chat GPT」を一般人にも提供している会社です。 #国会正門前大行動1019 #19日… | 日産の横須賀市追浜工場の跡地にアンドゥリル、という話が出たので… アンドゥリル(Andutil)という軍事企業にAIを提供している「OpenAI」という会社があります。 最近話題の「Chat GPT」を一般人にも提供している会社です。 #国会正門前大行動1019… | ↗Not briefed |
bluesky bluesky-ai | 21d | *und* das Denken das in einem IQ-Test abgefragt wird ist eins das ein LLM nicht durchführt | *und* das Denken das in einem IQ-Test abgefragt wird ist eins das ein LLM nicht durchführt | ↗Not briefed |
bluesky bluesky-ai | 21d | i also resent the suggestion that getting an LLM to write a sql query is the same as using it to create works of visual art or music or wri… | i also resent the suggestion that getting an LLM to write a sql query is the same as using it to create works of visual art or music o… | ↗Not briefed |
arxiv arxiv-ai | 21d | Dynamical Batching: How Class-Cycling Training Shapes the Internal Structure of Neural Networks | We investigate whether cyclically varying the class composition of training batches can modify both the performance and the internal o… | ↗Not briefed |
arxiv arxiv-ai | 21d | DeepSEA: A Multilingual Evaluation Benchmark for Southeast Asia | While language models have demonstrated remarkable capabilities, their performance remains disproportionately skewed towards high-reso… | ↗Not briefed |
bluesky bluesky-ai | 21d | www.commondreams.org/opinion/trum... #datacenters #bigtech #AI I thought originally that AI might be an interesting technology to make life… | www.commondreams.org/opinion/trum... #datacenters #bigtech #AI I thought originally that AI might be an interesting technology to make… | ↗Not briefed |
arxiv arxiv-ai | 21d | Physics-Driven Framework for QoS-Constrained Satellite System with Non-Linear PA Distortion | This work develops a physics-driven self-supervised graph learning framework that directly optimizes the distortion-aware precoder und… | ↗Not briefed |
hn hn-ai | 21d | Training a Chess Reasoning Agent | — | ↗Not briefed |
bluesky bluesky-ai | 21d | lately it feels like the opposite where the push to use AI is turning all programmers into middle-managers for claude | lately it feels like the opposite where the push to use AI is turning all programmers into middle-managers for claude | ↗Not briefed |