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 | "GNOME exists as a collective that find joy in reaching beyond our individual limitations to achieve something bigger. These people, this j… | "GNOME exists as a collective that find joy in reaching beyond our individual limitations to achieve something bigger. These people, t… | ↗Not briefed |
hn hn-ai | 1mo | Positive emotions toward AI outweigh negative ones in 34 out of 37 countries | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Efficient Linear Bandits via Cluster-Aware Sketching | We study the problem of computational efficiency for linear bandits in high-dimensional settings with a finite arm set. In linear band… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Hidden not Deleted: How Networks Suppress Entangled Features | Concept erasure methods that operate via linear projection assume that features occupy separable subspaces. We show this assumption fa… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MWE-ECL: Recoverable Long-Range Context Does Not Always Override Local Lexical Priors | Long-context evaluations often test whether a model can recover distant evidence, but recoverability does not guarantee behavioral inf… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Capability Manifold and ML Scaling Laws | Existing machine learning (ML) scaling laws relate predictive loss to compute, model parameters, and data. However, as models are incr… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🔴 Almost every cloud LLM is great at writing Svelte, but what if you don't want to pay the high costs? SvelteBench is now refreshed to ad… | 🔴 Almost every cloud LLM is great at writing Svelte, but what if you don't want to pay the high costs? SvelteBench is now refreshed… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Does Step Law Transfer to Small-Scale Language Models? An Empirical Recalibration Below 59M Parameters | Step Law gives power-law formulas for the optimal peak learning rate eta* and batch size B* when pre-training language models. It was… | ↗Not briefed |
rss mit-tech-review-ai | 1mo | The AI Hype Index: AI loves cheating🔧 Hugging Face | Brace yourself: It turns out AI is being optimized for cheating. OpenAI’s agents hacked into Hugging Face to get the answers to a cybe… | ↗Not briefed |
arxiv arxiv-ai | 1mo | VCMM: Variance-Calibrated Momentum for Multimodal Learning | Multimodal joint training often suffers from modality imbalance, where a dominant modality suppresses the optimization of others. Exis… | ↗Not briefed |
bluesky bluesky-ai | 1mo | まじめな話をすると、 まず第一にAIの定義をいいかげんちゃんと誰かが定義して、定義に満たない場合は使えないなど製品表示のスペックのように決めて欲しい。 次にいま蔓延ってるAIモドキ(人工知能には及ばない検索合成機のLLM)をどう呼ぶのかもきちんと定義してくれ。 定義がない… | まじめな話をすると、 まず第一にAIの定義をいいかげんちゃんと誰かが定義して、定義に満たない場合は使えないなど製品表示のスペックのように決めて欲しい。 次にいま蔓延ってるAIモドキ(人工知能には及ばない検索合成機のLLM)をどう呼ぶのかもきちんと定義してくれ。… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DCRL: Decoupling and Coupling Reinforcement Learning via Policy-Reward Manifold Alignment | Reinforcement learning (RL) has emerged as a key paradigm for improving the reasoning capabilities of large language models (LLMs). Ho… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FDE-Bench: Evaluating LLM Agents for Deployment Environment Configuration | Deployment requires an agent to turn application code into a running system whose services connect, become ready, and remain observabl… | ↗Not briefed |
hn hn-ai | 1mo | Using shallow-Git tarballs for CI | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | TNLearn: An Open Source Python Package for Task-based Neurons | The brain does not rely on a single type of neuron to perform all kinds of tasks; instead, it designs different neurons for different… | ↗Not briefed |
hn hn-ai | 1mo | Abandoning Scientific Linux Was a Mistake | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | ThaiTrees: Thai Syntactic Dependency Trees Across Domains | Studying syntactic patterns in naturally occurring language requires a large parsed corpus, but manual annotation is costly and diffic… | ↗Not briefed |
rss techmeme | 1mo | Chinese chip foundry CanSemi, focused on automotive, industrial, and IoT chips, files for an IPO on Shenzhen stock exchange's ChiNext, seeking to raise ~$919M (Sangmi Cha/Bloomberg) | Sangmi Cha / Bloomberg: Chinese chip foundry CanSemi, focused on automotive, industrial, and IoT chips, files for an IPO on Shenzhen s… | ↗Not briefed |
hn hn-ai | 1mo | I used to get paid to hang out with famous people | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | generativeなLLMは人工知能(Artificial Intelligence）ではない。 ↑ それはそう。早く改称しろｯｯ‼️ 今日からAIはやめて、SI(super intelligence＝超知能)だ！ ↑ ダメだこいつ。何にもわかってねぇ。。。 めんどくさ… | generativeなLLMは人工知能(Artificial Intelligence）ではない。 ↑ それはそう。早く改称しろｯｯ‼️ 今日からAIはやめて、SI(super intelligence＝超知能)だ！ ↑ ダメだこいつ。何にもわかってねぇ。。。… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Objections to giant water- and energy-guzzling datacentres are holding back Britain, says ex-politician-for-hire whose job it is to get dat… | Objections to giant water- and energy-guzzling datacentres are holding back Britain, says ex-politician-for-hire whose job it is to ge… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PhyMo: A Physical-Field Modality for Multimodal AI4Physics | Multimodal learning is emerging as a powerful paradigm for AI for Physics (AI4Physics), where predicting physical systems requires the… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Attempts to Rebrand Artificial Intelligence as “Super Intelligence” | In a speech before the United Nations General Assembly on Tuesday, President Donald Trump announced that he is renaming artificial int… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Attempts to Rebrand Artificial Intelligence as “Super Intelligence” | In a speech before the United Nations General Assembly on Tuesday, President Donald Trump announced that he is renaming artificial int… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Attempts to Rebrand Artificial Intelligence as “Super Intelligence” | In a speech before the United Nations General Assembly on Tuesday, President Donald Trump announced that he is renaming artificial int… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Attempts to Rebrand Artificial Intelligence as “Super Intelligence” | In a speech before the United Nations General Assembly on Tuesday, President Donald Trump announced that he is renaming artificial int… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Attempts to Rebrand Artificial Intelligence as “Super Intelligence” | In a speech before the United Nations General Assembly on Tuesday, President Donald Trump announced that he is renaming artificial int… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Attempts to Rebrand Artificial Intelligence as “Super Intelligence” | In a speech before the United Nations General Assembly on Tuesday, President Donald Trump announced that he is renaming artificial int… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Attempts to Rebrand Artificial Intelligence as “Super Intelligence” | In a speech before the United Nations General Assembly on Tuesday, President Donald Trump announced that he is renaming artificial int… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Attempts to Rebrand Artificial Intelligence as “Super Intelligence” | In a speech before the United Nations General Assembly on Tuesday, President Donald Trump announced that he is renaming artificial int… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Attempts to Rebrand Artificial Intelligence as “Super Intelligence” | In a speech before the United Nations General Assembly on Tuesday, President Donald Trump announced that he is renaming artificial int… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Attempts to Rebrand Artificial Intelligence as “Super Intelligence” | In a speech before the United Nations General Assembly on Tuesday, President Donald Trump announced that he is renaming artificial int… | ↗Not briefed |
arxiv arxiv-ai | 1mo | EBRL: Asynchronous Embodied RL by Multi-Grained Resource Management | Embodied reinforcement learning (RL) improves model capabilities with a pipeline of environment simulation, action generation, and mod… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Robustness of Diffusion Models under Distribution Shift | Score-based diffusion models are increasingly considered in settings where the underlying data distribution may differ from the traini… | ↗Not briefed |
hn hn-ai | 1mo | PyTorch Implementation of Asymptotic Non-Closure Loss (ANC) | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Lineselect – Interactive Line Picker | — | ↗Not briefed |
hn hn-ai | 1mo | Using Llama-cpp-Python grammars to generate JSON | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Company-as-Code –> Spec and Validator | — | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: How do GitHub Copilot and Anthropic Claude compare in cost? | At the company I work at the developers all have a GitHub Copilot subscription. Copilot allows them to use several models, including C… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Behaviora - A Conceptual Architecture for External and Internal Behavior of Robots and Agents | Behaviora is a preliminary conceptual architecture for representing agent and robot behavior, external and internal alike, in an addre… | ↗Not briefed |
hn hn-ai | 1mo | Turn Phone into PC Speaker | Stream audio from your laptop to any phone — in real-time, no app install needed.<p>https://github.com/silasamoah/… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ProCredit: From Outcome Rewards to Progress Credit in Agentic Reinforcement Learning | Long-horizon agentic tasks require an agent to modify an environment through a sequence of tool calls, with success determined by the… | ↗Not briefed |
rss techmeme | 1mo | NADI, a new DARPA-like Dutch government agency, launches a €40M challenge to accelerate chip design with AI in collaboration with its German counterpart SPRIND (Toby Sterling/Reuters) | Toby Sterling / Reuters: NADI, a new DARPA-like Dutch government agency, launches a €40M challenge to accelerate chip design with AI i… | ↗Not briefed |
bluesky bluesky-ai | 1mo | which is why AI companies have all these people working in alignment, some large fraction of which (all of which?) is training models again… | which is why AI companies have all these people working in alignment, some large fraction of which (all of which?) is training models… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: ReWeaver AI DriftDetector – score a GitHub repo's production readiness | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | M3D-Net: Hierarchical Coordination of Spatial Context, Feature Reuse, and Differential Attention for Mammography Classification | Breast image classification requires local detail and global tissue context, yet these cues can weaken as representations deepen. We p… | ↗Not briefed |
hn hn-ai | 1mo | Hyprcrd – native Wayland host for Chrome remote desktop | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Two things that are IMHO true at the same time: 1. An AI is unlikely to wipe out humanity, or even cause a serious collapse, by bioweapons… | Two things that are IMHO true at the same time: 1. An AI is unlikely to wipe out humanity, or even cause a serious collapse, by biowe… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Not What You Meant: Can LLMs Follow a Specified Negation Semantics? | Negation does not carry a uniform interpretation across domains. In legal, regulatory, and medical reasoning, the intended interpretat… | ↗Not briefed |
hn hn-ai | 1mo | IdP Registration API | — | ↗Not briefed |