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 | 19d | A Semantic Knowledge Graph Construction Pipeline for Vehicle Intention Prediction in nuScenes | Knowledge graphs (KGs) are increasingly used in autonomous driving (AD) to make traffic scene entities and relations explicit for pred… | ↗Not briefed |
bluesky bluesky-ai | 19d | The US DHS orders all federal civilian agencies to remove any DeepSeek tech from their IT systems and prohibits future procurement of its p… | The US DHS orders all federal civilian agencies to remove any DeepSeek tech from their IT systems and prohibits future procurement of… | ↗Not briefed |
bluesky bluesky-ai | 19d | Both can be true. In fact I saw someone comparing their ai decomp game launcher to woodworking claiming it's just their hobby now DON'T Y… | Both can be true. In fact I saw someone comparing their ai decomp game launcher to woodworking claiming it's just their hobby now DO… | ↗Not briefed |
rss techmeme | 19d | The US DHS orders all federal civilian agencies to remove any DeepSeek tech from their IT systems and prohibits future procurement of its products and services (Connor Hart/Wall Street Journal) | Connor Hart / Wall Street Journal: The US DHS orders all federal civilian agencies to remove any DeepSeek tech from their IT systems a… | ↗Not briefed |
arxiv arxiv-ai | 19d | A Fisher-Geometric View of Multi-Teacher On-Policy Distillation | Consolidating specialized foundation models into a unified generalist is a central goal of post-training. Two approaches are in common… | ↗Not briefed |
arxiv arxiv-ai | 19d | Beyond Behavioral Accuracy: Evaluating Individual-Level Fidelity in LLM Investor Simulation | Large language models (LLMs) are increasingly used as user simulators, but can they faithfully reproduce an individual's behavior and… | ↗Not briefed |
bluesky bluesky-ai | 19d | For the most part, analytical writing is judged by those criteria. The entire purpose is to judge whether the human who signed the work is … | For the most part, analytical writing is judged by those criteria. The entire purpose is to judge whether the human who signed the wor… | ↗Not briefed |
bluesky bluesky-ai | 19d | Opsin raises $21m to police what enterprise AI agents do SYN Ventures and Race Capital led the round; Opsin, which began by finding data o… | Opsin raises $21m to police what enterprise AI agents do SYN Ventures and Race Capital led the round; Opsin, which began by finding d… | ↗Not briefed |
hn hn-ai | 19d | Is Python the favorite language of LLMs? | — | ↗Not briefed |
bluesky bluesky-ai | 19d | Juste pour clarifier : j'ai pas parlé d'IA/LLM dans mon post, c'est pas nécessaire pour faire de l'art génératif, une forme d'art qui expli…🔧 ChatGPT | Juste pour clarifier : j'ai pas parlé d'IA/LLM dans mon post, c'est pas nécessaire pour faire de l'art génératif, une forme d'art qui… | ↗Not briefed |
arxiv arxiv-ai | 19d | Toward Structured Financial Foundation Models | Foundation models promise reusable capabilities and step-change performance gains across tasks. In finance, large-scale pretraining of… | ↗Not briefed |
bluesky bluesky-ai | 19d | Jerry isnt real he's actually an LLM | Jerry isnt real he's actually an LLM | ↗Not briefed |
hn hn-ai | 19d | Why Private LLM Apps Matter | — | ↗Not briefed |
bluesky bluesky-ai | 19d | Slop PRみたいな問題は言わずもがななんだが、それ以前に個人の困りごとを解決するツールとしてのLLMの思想があまりに拙速に便利だったおかげで、AさんもBさんも同じ問題で困ってる、じゃあこれを共通化してお互いにとって便利なものを作ろうという方向性のテクノロジーの力がすっかり忘… | Slop PRみたいな問題は言わずもがななんだが、それ以前に個人の困りごとを解決するツールとしてのLLMの思想があまりに拙速に便利だったおかげで、AさんもBさんも同じ問題で困ってる、じゃあこれを共通化してお互いにとって便利なものを作ろうという方向性のテクノロジーの力が… | ↗Not briefed |
hn hn-ai | 19d | Building an AI assistant that negotiates appointments and bookings for you | — | ↗Not briefed |
hn hn-ai | 19d | Show HN: Acyclic – Rapid agent swarms. Agentic workflows exponentially faster | Hi HN, we're the co-founders of Acyclic (YC F26).<p>Acyclic runs any agentic workflow as a swarm. You give it a task and your pro… | ↗Not briefed |
hn hn-ai | 19d | I can't use generic lambdas at compile-time, now what? | — | ↗Not briefed |
arxiv arxiv-ai | 19d | RouterWorld: Scaling Conversational AI with Latency-Aware LLM Routing | Conversational AI systems built on Large Language Models (LLMs) require routing mechanisms that balance answer quality and response la… | ↗Not briefed |
bluesky bluesky-ai | 19d | Musk *claims* he plans to give one share of his stock to each American “kid” next year. (If he were serious, there is nothing stopping him … | Musk *claims* he plans to give one share of his stock to each American “kid” next year. (If he were serious, there is nothing stopping… | ↗Not briefed |
arxiv arxiv-ai | 19d | Learn the Task, Not the Schedule: Reducing Spurious Cross-Arm Coupling in Bimanual Imitation Learning | A bimanual demonstration specifies both what the robot should do and one particular schedule for when each arm should act. Standard im… | ↗Not briefed |
bluesky bluesky-ai | 19d | Remember when we said that all the fear-hype and "AI slowdown/pause/pacing/whatever" talk was just a way for them to set up regulatory capt… | Remember when we said that all the fear-hype and "AI slowdown/pause/pacing/whatever" talk was just a way for them to set up regulatory… | ↗Not briefed |
hn hn-ai | 19d | Replit makes OpenAI its default over Anthropic, CEO says🔧 Replit | — | ↗Not briefed |
hn hn-ai | 19d | Show HN: WeDecide – A tool for easier decision making for group trips | Hey!<p>In my spare time I am working on this little tool to let groups vote on accomodation/budget/dates for trips.<p>I star… | ↗Not briefed |
bluesky bluesky-ai | 19d | also, in musk v altman we learned elon musk cut off promised donations to pressure openai when it was a research nonprofit to grant him con… | also, in musk v altman we learned elon musk cut off promised donations to pressure openai when it was a research nonprofit to grant hi… | ↗Not briefed |
bluesky bluesky-ai | 19d | As leading AI companies insist their technologies could usher in a new industrial revolution, they are grappling with knotty questions abou… | As leading AI companies insist their technologies could usher in a new industrial revolution, they are grappling with knotty questions… | ↗Not briefed |
arxiv arxiv-ai | 19d | ScopeMem: Latent Memory That Learns Where to Revise Frozen GUI Agents | Latent memory is gaining ground as a lightweight history supplement for GUI agents. It supplies history to a frozen vision-language mo… | ↗Not briefed |
bluesky bluesky-ai | 19d | I don't know who this guy is but I can only assume it's some AI bro sephirothposting because America's literally never been without a perma… | I don't know who this guy is but I can only assume it's some AI bro sephirothposting because America's literally never been without a… | ↗Not briefed |
hn hn-ai | 19d | The Second Brain Gnome Users Deserve. Go Ahead, HN. Flag This OpenSourceProject | — | ↗Not briefed |
rss techmeme | 19d | Sources: US investigators pursuing ShinyHunters, which stole FBI data, identified executives at two ransomware negotiation firms who had unusually close ties (Brian Krebs/Krebs on Security) | Brian Krebs / Krebs on Security: Sources: US investigators pursuing ShinyHunters, which stole FBI data, identified executives at two r… | ↗Not briefed |
bluesky bluesky-ai | 19d | Ex-OpenAI employee accused of planting device appears to have Saudi backing, court filing shows reut.rs/3WeDf12 | Ex-OpenAI employee accused of planting device appears to have Saudi backing, court filing shows reut.rs/3WeDf12 | ↗Not briefed |
arxiv arxiv-ai | 19d | Mitigating Reward Hacking in Preference Reward Models with Evolving Penalties | Reinforcement learning for language models has moved from preference reward models to verifiable rewards, which sustain much longer op… | ↗Not briefed |
hn hn-ai | 19d | Show HN: Spotnote – Vite plugin for UI feedback to coding agents | — | ↗Not briefed |
hn hn-ai | 19d | Gemini AI Suite – Zero-Cost Next.js 14, Gemini 3.8 Flash, and Stripe Starter | — | ↗Not briefed |
bluesky bluesky-ai | 19d | Apesar que a gente pode entrar no que ele quer chamar de inteligência artificial, porque existem milhares de coisas no guarda chuva intelig… | Apesar que a gente pode entrar no que ele quer chamar de inteligência artificial, porque existem milhares de coisas no guarda chuva in… | ↗Not briefed |
bluesky bluesky-ai | 19d | Double effet cumulatif, probablement: 1. l'incitation des politiques nationales "l'objectif c'est d'aller à l'Europe"; 2. les LLM qui dés… | Double effet cumulatif, probablement: 1. l'incitation des politiques nationales "l'objectif c'est d'aller à l'Europe"; 2. les LLM qu… | ↗Not briefed |
bluesky bluesky-ai | 19d | inclusive em alguns pontos é até ilegal usar modelo não determinístico (modelos que as mesmas entradas podem produzir resultados diferentes… | inclusive em alguns pontos é até ilegal usar modelo não determinístico (modelos que as mesmas entradas podem produzir resultados difer… | ↗Not briefed |
hn hn-ai | 19d | Detecting model weight tampering in GPU memory (Rome attacks) | — | ↗Not briefed |
hn hn-ai | 19d | Show HN: ECG Signal Processing Starter Kit – Python, SciPy, and FastAPI | — | ↗Not briefed |
bluesky bluesky-ai | 19d | Chainguard’s new Harness points frontier models like Claude Mythos 5.1 at your code, alongside new tools to isolate coding agents and verif… | Chainguard’s new Harness points frontier models like Claude Mythos 5.1 at your code, alongside new tools to isolate coding agents and… | ↗Not briefed |
rss theverge-ai | 19d | JB Pritzker thinks AI could help us live to 100 — or threaten us like nukes | Today, I’m talking with Illinois Governor JB Pritzker, who has a lot of thoughts about AI, tech, and privacy. If you’ve been listenin… | ↗Not briefed |
arxiv arxiv-ai | 19d | Strength Controls Change How Interventions Compare Across Sparse Autoencoder Codes | Changing sparse autoencoder (SAE) codes can alter both intervention direction and magnitude, so comparing intervention effects require… | ↗Not briefed |
bluesky bluesky-ai | 19d | By this logic, paying someone to write your essay for you should've been OK back in the pre-LLM days. What matters is the quality, not the … | By this logic, paying someone to write your essay for you should've been OK back in the pre-LLM days. What matters is the quality, not… | ↗Not briefed |
arxiv arxiv-ai | 19d | SPAR: Structural Pruning and Action Recovery for Vision-Language-Action Models | The limited onboard computing resources of robots make the inference demands of vision-language-action (VLA) models costly to accommod… | ↗Not briefed |
bluesky bluesky-ai | 19d | ‘It scored higher than OpenAI’s GPT5.6 Sol on Terminal-Bench 4.0. DeepSeek only charges $1.20 for output per 1 million tokens at peak hours… | ‘It scored higher than OpenAI’s GPT5.6 Sol on Terminal-Bench 4.0. DeepSeek only charges $1.20 for output per 1 million tokens at peak… | ↗Not briefed |
arxiv arxiv-ai | 19d | Wet-Biased Data Pipelines Mislead the Training and Evaluation of Generative Precipitation Nowcasting | Generative models for precipitation nowcasting are trained and evaluated predominantly on wet-biased datasets, which over-represent pr… | ↗Not briefed |
bluesky bluesky-ai | 19d | Yes, SpaceX (and Anthropic) are trying to make the public stakeholders in the company because they know their bad finances will eventually … | Yes, SpaceX (and Anthropic) are trying to make the public stakeholders in the company because they know their bad finances will eventu… | ↗Not briefed |
bluesky bluesky-ai | 19d | Filing: US prosecutors say OpenAI suspect Abdulrahman Alfozan maintained frequent WhatsApp contact in 2025 and 2026 with an unnamed Saudi R… | Filing: US prosecutors say OpenAI suspect Abdulrahman Alfozan maintained frequent WhatsApp contact in 2025 and 2026 with an unnamed Sa… | ↗Not briefed |
rss techmeme | 19d | Filing: US prosecutors say OpenAI suspect Abdulrahman Alfozan maintained frequent WhatsApp contact in 2025 and 2026 with an unnamed Saudi Royal Family member (Dan Mangan/CNBC) | Dan Mangan / CNBC: Filing: US prosecutors say OpenAI suspect Abdulrahman Alfozan maintained frequent WhatsApp contact in 2025 and 2026… | ↗Not briefed |
hn hn-ai | 19d | Low-bit LLMs on CPU: what runs correctly and fast where | — | ↗Not briefed |
bluesky bluesky-ai | 19d | The point of was how awesome LLMs are because they were able to run analysis of prior academic articles & find problems. Which is basicall… | The point of was how awesome LLMs are because they were able to run analysis of prior academic articles & find problems. Which is bas… | ↗Not briefed |