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 |
|---|---|---|---|---|
hn hn-ai | 5d | We're addicted to social media. How do we get sober? | — | ↗Not briefed |
bluesky bluesky-ai | 5d | Anoche fue noche grande en La Zona 📻 con el estreno de nuestro radiodocumental «Archæa, Bacteria, Eukarya, Dæmonica, Synthetica, Machinica… | Anoche fue noche grande en La Zona 📻 con el estreno de nuestro radiodocumental «Archæa, Bacteria, Eukarya, Dæmonica, Synthetica, Mach… | ↗Not briefed |
arxiv arxiv-ai | 5d | Beyond Data Schedule: Fast LLM Pretraining on Data Mixtures by Loss-Reweighting | Large language models often acquire specialized capabilities through successive pretraining stages, but these gains can come at the co… | ↗Not briefed |
bluesky bluesky-ai | 5d | noteにおけるLLM利用投稿の氾濫については、「まあ、自分とこの利用規約も遵守せずに、明確な人権侵害・中傷・差別扇動投稿をおもいっきり容認してるプラットフォームが何か手立てを打つはずもありませんわなあ。打てよ」という感想がございます。 （XやInstagram・Threads… | noteにおけるLLM利用投稿の氾濫については、「まあ、自分とこの利用規約も遵守せずに、明確な人権侵害・中傷・差別扇動投稿をおもいっきり容認してるプラットフォームが何か手立てを打つはずもありませんわなあ。打てよ」という感想がございます。 （XやInstagram・Th… | ↗Not briefed |
hn hn-ai | 5d | Show HN: Pristine – an agent skill that critiques and polishes web UI | — | ↗Not briefed |
hn hn-ai | 5d | Lima- Linux VMs with automatic file sharing and port forwarding(similar to WSL2) | — | ↗Not briefed |
rss techcrunch-ai | 5d | Uber rival inDrive eyes 2027 robotaxi tests in Central Asia with Pony.ai and Nvidia | inDrive plans to begin testing Pony.ai's robotaxis in Kazakhstan's capital in the first half of 2027. | ↗Not briefed |
bluesky bluesky-ai | 5d | 「バグのないソフトウェア」の実現を目指してAnthropicが「AIを用いて数学的にバグがないことを証明する」という手法を開発 https://gigazine.net/news/20261023-anthropic-formal-verification/ | 「バグのないソフトウェア」の実現を目指してAnthropicが「AIを用いて数学的にバグがないことを証明する」という手法を開発 https://gigazine.net/news/20261023-anthropic-formal-verification/ | ↗Not briefed |
arxiv arxiv-ai | 5d | SyncPro: Synchronous Reasoning and Responding for Proactive Streaming Video Understanding | Proactive streaming video understanding requires assistants to track unfolding events and respond when a user's requested condition is… | ↗Not briefed |
arxiv arxiv-ai | 5d | Shorter Is Not Cheaper: A Controlled Comparison of Declarative Animation Specifications and Free-Form Code for LLM-Generated Explanations | Chat assistants increasingly answer conceptual questions with interactive, animated explanations, generated either as free-form HTML/J… | ↗Not briefed |
arxiv arxiv-ai | 5d | From Noisy Retrieval to Decision Evidence: Silver Evidence Demonstrations for In-Context Learning | RAG and retrieval-augmented few-shot ICL can struggle when decision-relevant evidence is buried in long, noisy context. Retrieved pass… | ↗Not briefed |
bluesky bluesky-ai | 5d | what i tend not to see in the AI ethics debates is that it's not JUST that LLM-based work is lifted from existing work. it's not JUST tha… | what i tend not to see in the AI ethics debates is that it's not JUST that LLM-based work is lifted from existing work. it's not JUS… | ↗Not briefed |
arxiv arxiv-ai | 5d | MPQ: A Message-Passing View of Post-Training Quantization | Post-training quantization (PTQ) is a key technique for efficiently deploying large language models, as it compresses weights into low… | ↗Not briefed |
bluesky bluesky-ai | 5d | Tim Bernard and Justin Hendrix break down 13 new complaints against OpenAI: eight from state attorneys general and five from private plaint… | Tim Bernard and Justin Hendrix break down 13 new complaints against OpenAI: eight from state attorneys general and five from private p… | ↗Not briefed |
arxiv arxiv-ai | 5d | Data Quality in RVL-CDIP: A Systematic Audit | RVL-CDIP is the de facto standard benchmark for document classification, and reported accuracy on it is now 98.8%. In this paper, we a… | ↗Not briefed |
arxiv arxiv-ai | 5d | Conditional Flow Models for Offline Maximum Entropy Reinforcement Learning | Offline maximum entropy reinforcement learning seeks to estimate the optimal soft action-value function \(Q^*\) from a fixed dataset w… | ↗Not briefed |
arxiv arxiv-ai | 5d | Semantically Rich, Metrically Affect-Blind: A Geometric Blind Spot in Modern Text Embedders | We report a systematic dissociation in modern text embedders: sentiment is almost perfectly decodable from their embeddings, yet almos… | ↗Not briefed |
arxiv arxiv-ai | 5d | ImageRSI: Verification-Driven Recursive Self-Improvement for Agentic Image Generation | Complex text-to-image requests remain difficult to satisfy reliably. Verifying and repairing an image can fix an individual output, bu… | ↗Not briefed |
hn hn-ai | 5d | Create Prospect focused sales collateral | — | ↗Not briefed |
hn hn-ai | 5d | Patching the Sony PS3 for RGB Scart When Playing DVDs and BluRay | — | ↗Not briefed |
arxiv arxiv-ai | 5d | EDWARD-NMT: Elastic Dimensions With Adaptive Routing and Dual-consistency for Neural Machine Translation | Sparse Mixture-of-Experts (MoE) is a powerful paradigm for scaling Neural Machine Translation (NMT) and managing language heterogeneit… | ↗Not briefed |
arxiv arxiv-ai | 5d | Tampered ML: Evaluating LLM Judgment of Machine Learning Research | One approach to accelerating progress on AI alignment is to automate parts of alignment research itself. Doing so requires systems tha… | ↗Not briefed |
bluesky bluesky-ai | 5d | I don't think they're using it for art, but I do wonder exactly to what extent LLM-generated code has seeped into first-party Nintendo game… | I don't think they're using it for art, but I do wonder exactly to what extent LLM-generated code has seeped into first-party Nintendo… | ↗Not briefed |
arxiv arxiv-ai | 5d | REACT: Representational Early Analysis and Control of Chain-of-Thought Trajectories | Chain-of-thought (CoT) generation offers an opportunity to intervene before a reasoning model completes its answer. But does the point… | ↗Not briefed |
bluesky bluesky-ai | 5d | Showing how anti-LLM I am by writing an art project in Prolog. Connectionism shall not triumph this time! 🧐 | Showing how anti-LLM I am by writing an art project in Prolog. Connectionism shall not triumph this time! 🧐 | ↗Not briefed |
hn hn-ai | 5d | Show HN: What LLM fits my GPU? – VRAM measured on real hardware for 117 models | — | ↗Not briefed |
bluesky bluesky-ai | 5d | Here's what I'm curious about: will LLM -driven discovery of severe flaws like this slow down as the supply of these outstanding vulns is e… | Here's what I'm curious about: will LLM -driven discovery of severe flaws like this slow down as the supply of these outstanding vulns… | ↗Not briefed |
arxiv arxiv-ai | 5d | Who Verifies the Tools? How Warped Formatters Mislead Agents and Their Verifiers | Agents and verifiers often rely on the text a tool prints, and that text can misrepresent what the tool actually did, even when the to… | ↗Not briefed |
arxiv arxiv-ai | 5d | Context or Knowledge: How LLMs Attribute Their Own Answers | Language models can answer a question from a supplied document or from their parametric knowledge. When later asked where the answer c… | ↗Not briefed |
arxiv arxiv-ai | 5d | Beyond What to Say: Proactive Utterance Timing with Inner Thoughts for Werewolf Agents | Participating in a conversation involves not only deciding what to say but also whether to speak at all. This judgment carries even gr… | ↗Not briefed |
bluesky bluesky-ai | 5d | On one hand, my blanket anti-LLM stance is largely rooted in opposition to the irresponsible behaviour of the frontier labs and unethical t… | On one hand, my blanket anti-LLM stance is largely rooted in opposition to the irresponsible behaviour of the frontier labs and unethi… | ↗Not briefed |
bluesky bluesky-ai | 5d | Literally doing formally verified apache2 right now with DeepSeek Flash, and it does in fact cost pennies. I've spent like three bucks righ… | Literally doing formally verified apache2 right now with DeepSeek Flash, and it does in fact cost pennies. I've spent like three bucks… | ↗Not briefed |
hn hn-ai | 5d | The Signal Chats That Cost OpenAI Researchers Their Jobs | — | ↗Not briefed |
bluesky bluesky-ai | 5d | 今のご時世って、先進性を謳うために「問題があるなら問題が起こってから対応しよう」ってな具合で、でもセキュリティや保守の観点で言えば「セキュリティは防衛と攻撃だと攻撃側が有利」の土俵であるが故に「後手になる対応はまずいのでは？」という板挟み状態だと思うのよね… でも今のご時世は… | 今のご時世って、先進性を謳うために「問題があるなら問題が起こってから対応しよう」ってな具合で、でもセキュリティや保守の観点で言えば「セキュリティは防衛と攻撃だと攻撃側が有利」の土俵であるが故に「後手になる対応はまずいのでは？」という板挟み状態だと思うのよね… でも今… | ↗Not briefed |
bluesky bluesky-ai | 5d | 「私は生成AIに反対の立場です」 と言いながら明らかな誤情報や差別思想を 「GoogleのAI要約でみた」といったソースで拡散する方が多くいるのですが 芸術分野にとどまらずAI画像が多すぎて、まともな資料はほぼインターネットで仕入れることは難しい状況かつ、そういったAI企業であ… | 「私は生成AIに反対の立場です」 と言いながら明らかな誤情報や差別思想を 「GoogleのAI要約でみた」といったソースで拡散する方が多くいるのですが 芸術分野にとどまらずAI画像が多すぎて、まともな資料はほぼインターネットで仕入れることは難しい状況かつ、そういったA… | ↗Not briefed |
arxiv arxiv-ai | 5d | PersonaScienceBench: Evaluating Scientific Agents with Diverse Simulated Researchers | Scientific agents collaborate with researchers whose expertise and preferences differ, but benchmarks usually hold the collaborator fi… | ↗Not briefed |
arxiv arxiv-ai | 5d | Putting Transformer Theories to the Test | Several formal theories have been developed to characterize the capabilities and limitations of the transformer architecture, each emp… | ↗Not briefed |
arxiv arxiv-ai | 5d | ReMedI: Inference-Time Refinement and Self-Distillation for Controllable Molecular Optimization | In drug development, lead optimization must improve competing properties, such as increasing permeability and reducing toxicity, while… | ↗Not briefed |
arxiv arxiv-ai | 5d | PRISM: Promoting Sampling Diversity of Large Language Models with Implicit Stochastic Modulation | Large language models (LLMs) exhibit mode collapse, where their output distribution concentrates on familiar semantic modes. When dive… | ↗Not briefed |
bluesky bluesky-ai | 5d | "Oh they don't care cause they don't have data centers like we do" 1) some of the boards are now trying to get that done 2) you don't need… | "Oh they don't care cause they don't have data centers like we do" 1) some of the boards are now trying to get that done 2) you don't… | ↗Not briefed |
arxiv arxiv-ai | 5d | Learning Nonlinear Waveform Fusion Across Scales for Integrated Sensing and Communication | Fixed source superposition and frequency partitioning constrain how communication content interacts with a probing reference in integr… | ↗Not briefed |
bluesky bluesky-ai | 5d | 秋風が気持ちいいからお散歩してきます🚶🏼♀️➡️🐾 【 参加者募集中 】 ブログ・日記・イラスト・写真・小説をBlueskyアカウントで投稿・販売できる複合型コンテンツサービス ➡️ atelier.fan #ブログ #日記 #イラスト #写真 #小説 #AIイ… | 秋風が気持ちいいからお散歩してきます🚶🏼♀️➡️🐾 【 参加者募集中 】 ブログ・日記・イラスト・写真・小説をBlueskyアカウントで投稿・販売できる複合型コンテンツサービス ➡️ atelier.fan #ブログ #日記 #イラスト #写真 #小説… | ↗Not briefed |
hn hn-ai | 5d | AI RSS reader with summaries – Brew beta | — | ↗Not briefed |
bluesky bluesky-ai | 5d | The robot suggested that first and foremost it should audit our own ramshackle infra in light of recent developments. This kind of behavior… | The robot suggested that first and foremost it should audit our own ramshackle infra in light of recent developments. This kind of beh… | ↗Not briefed |
arxiv arxiv-ai | 5d | DELA: Learning Controllable Edits for Diffusion Language Models | Diffusion language models (DLMs) support bidirectional denoising and iterative refinement, but generation remains largely controlled b… | ↗Not briefed |
hn hn-ai | 5d | HyperCache – Distributed KV-cache quantization and relative RoPE reuse | — | ↗Not briefed |
arxiv arxiv-ai | 5d | TTT in VLA Models: Exploring Training with Human Demonstrations | Long-horizon robot tasks require retaining and using past information. We integrate Test-Time Training (TTT), which stores context in… | ↗Not briefed |
arxiv arxiv-ai | 5d | AgentTrace: Execution-Level Fingerprinting of Black-Box LLM Agents | Verifying the provenance of autonomous LLM agents is a critical security challenge. Existing semantic attribution is easily obfuscated… | ↗Not briefed |
bluesky bluesky-ai | 5d | I'm dooming myself to this becoming the next thing someone applies an LLM to because I'm Asking The Better Question, but so far to my knowl… | I'm dooming myself to this becoming the next thing someone applies an LLM to because I'm Asking The Better Question, but so far to my… | ↗Not briefed |
hn hn-ai | 5d | See how Big Tech is transforming Spain's landscape | — | ↗Not briefed |