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 | 1d | それ LLM 使ってみただけちゃうのんと質問したら、喧嘩になりそうなので黙っている。 | それ LLM 使ってみただけちゃうのんと質問したら、喧嘩になりそうなので黙っている。 | ↗Not briefed |
hn hn-ai | 1d | Show HN: Agent Hive – cute Grokbot like agentic teams with queen and many bees | — | ↗Not briefed |
hn hn-ai | 1d | New models tested: Gemini 4 Hydrogen, Qwen4-Max, and GPT-6.2 | — | ↗Not briefed |
bluesky bluesky-ai | 1d | The five-year-old artificial intelligence powerhouse priced its initial public offering at $83 a share, shrugging off growing concern about… | The five-year-old artificial intelligence powerhouse priced its initial public offering at $83 a share, shrugging off growing concern… | ↗Not briefed |
bluesky bluesky-ai | 1d | Sharing one of my apps: AI Tennis Swing Coach 🎾 Film your swing and it scores every forehand, backhand and serve 0–100, names the fault a… | Sharing one of my apps: AI Tennis Swing Coach 🎾 Film your swing and it scores every forehand, backhand and serve 0–100, names the fa… | ↗Not briefed |
arxiv arxiv-ai | 1d | VoxForge: A Unified Foundation Model for Dense 3D Features Across MRI, CT and PET🔧 Modal | A unified 3D medical foundation model promises reusable, frozen features that serve a wide variety of tasks such as segmentation, clas… | ↗Not briefed |
hn hn-ai | 1d | PaloAlto Networks buys AppliedCompute founded last yr by 3 OpenAI alums for 4.5B | — | ↗Not briefed |
bluesky bluesky-ai | 1d | dawg i was so fucking right its so funny theyre SMOKING openai rn too ppl are so short term thinking brained rn this stupid "race" bullshit… | dawg i was so fucking right its so funny theyre SMOKING openai rn too ppl are so short term thinking brained rn this stupid "race" bul… | ↗Not briefed |
bluesky bluesky-ai | 1d | TFW the communist agitator comes out against open weight models bc they think it will justify nationalizing Anthropic and OpenAI, and boy a… | TFW the communist agitator comes out against open weight models bc they think it will justify nationalizing Anthropic and OpenAI, and… | ↗Not briefed |
rss simon-willison | 1d | llm-gemini 0.38 | — | ↗Not briefed |
bluesky bluesky-ai | 1d | And there's not a way to be good at it. If this company can do that then anyone can do it and why pay them? It's batshit to think that "Do… | And there's not a way to be good at it. If this company can do that then anyone can do it and why pay them? It's batshit to think tha… | ↗Not briefed |
bluesky bluesky-ai | 1d | We’ll know the official U.S. government position vis a vis Anthropic once we find out how much President Trump invested in their IPO. | We’ll know the official U.S. government position vis a vis Anthropic once we find out how much President Trump invested in their IPO. | ↗Not briefed |
bluesky bluesky-ai | 1d | Totally agreed. It is definitely not lost on me that a lot of the money being dumped into security research right now is by companies pushi… | Totally agreed. It is definitely not lost on me that a lot of the money being dumped into security research right now is by companies… | ↗Not briefed |
hn hn-ai | 1d | Show HN: GTA 6 is out, so I built an agent to do my job in Chrome🔧 Google Ads MCP | GTA 6 is here! I used to work in marketing automation, which mostly meant clicking around Klaviyo all day. So I built the obvious thin… | ↗Not briefed |
hn hn-ai | 1d | Show HN: Pitot – OpenTelemetry in One Binary, with Parquet, DuckDB, and MCP | — | ↗Not briefed |
rss techmeme | 1d | The Homeland Security Task Force launches the Threat Disruption Center to fight crypto fraud and cybercrime, using TRM Labs' AI platform via a ~$95M no-bid deal (Bloomberg) | Bloomberg: The Homeland Security Task Force launches the Threat Disruption Center to fight crypto fraud and cybercrime, using TRM Labs… | ↗Not briefed |
arxiv arxiv-ai | 1d | SplitReAlign: Partition-Aware Representation Realignment for Split Federated Unlearning | Split federated learning (SFL) distributes neural networks across clients and a server, coupling class-specific knowledge across their… | ↗Not briefed |
hn hn-ai | 1d | Tchiko Native GTK3 Linux sysadmin companion with sub-15ms local dispatch | — | ↗Not briefed |
arxiv arxiv-ai | 1d | Federated Learning of Car Aerodynamics Surrogates with Spectral Optimizers | Neural surrogates predict the aerodynamic forces on a car in seconds, compared with hours for a CFD simulation, but training them requ… | ↗Not briefed |
bluesky bluesky-ai | 1d | Tomorrow: Anthropic IPO! | Tomorrow: Anthropic IPO! | ↗Not briefed |
bluesky bluesky-ai | 1d | THE NEWS SAID THIS WAS MASS MURDER GAME, I BEEN PLAYING GTA ONLINE THE LAST 10 YEARS JUST MOWING BITCHES DOWN AND NOW THE COPS HAVE "BETTER… | THE NEWS SAID THIS WAS MASS MURDER GAME, I BEEN PLAYING GTA ONLINE THE LAST 10 YEARS JUST MOWING BITCHES DOWN AND NOW THE COPS HAVE "B… | ↗Not briefed |
bluesky bluesky-ai | 1d | leftover parts plus a $100 brand new 1tb wt black nvme with 27 minutes of access time and i got a self hosted ai model that runs for about … | leftover parts plus a $100 brand new 1tb wt black nvme with 27 minutes of access time and i got a self hosted ai model that runs for a… | ↗Not briefed |
bluesky bluesky-ai | 1d | Lets cut to the chase. He's at between 10%-13% approval rating- less than anyone else in the Senate. Also this is completely unacceptable… | Lets cut to the chase. He's at between 10%-13% approval rating- less than anyone else in the Senate. Also this is completely unaccep… | ↗Not briefed |
bluesky bluesky-ai | 1d | I think it's more CoT output reingestion than post-training (i.e. yet more text). But it's hard to know because they got really secretive a… | I think it's more CoT output reingestion than post-training (i.e. yet more text). But it's hard to know because they got really secret… | ↗Not briefed |
hn hn-ai | 1d | Show HN: Frontier Observatory – A visual timeline of what the AI labs ship | Frontier labs ship a lot. This gives me a handle on things. Also nice view of what the labs are prioritising (if release/research… | ↗Not briefed |
rss techcrunch-ai | 1d | Amidst fears of AI-cyberattacks, Satya Nadella says lots of critical infrastructure is ‘unpatchable’ | The Microsoft CEO downplayed the recent Outlook outage while noting that AI was causing both opportunities and new challenges for secu… | ↗Not briefed |
hn hn-ai | 1d | Getting Access | — | ↗Not briefed |
bluesky bluesky-ai | 1d | On the one hand this is stupid and any idiot can see it’s not going to work on any level and is just a slight rephrase of all the other bad… | On the one hand this is stupid and any idiot can see it’s not going to work on any level and is just a slight rephrase of all the othe… | ↗Not briefed |
hn hn-ai | 1d | Apache Shiro Joins the GitHub Secure Open Source Fund | — | ↗Not briefed |
bluesky bluesky-ai | 1d | so, you know, someone with no boundaries re: fic, using AI to code a website for fic readers... the logical conclusion is your work will ge… | so, you know, someone with no boundaries re: fic, using AI to code a website for fic readers... the logical conclusion is your work wi… | ↗Not briefed |
hn hn-ai | 1d | Maak v1.0.0 – infinitely extensible command runner, build sys in Lisp | — | ↗Not briefed |
arxiv arxiv-ai | 1d | Cross-Question Reliable Reinforcement Learning | AI systems operating near the limits of their capabilities can maximize utility while minimizing the risk of error by abstaining or de… | ↗Not briefed |
hn hn-ai | 1d | Show HN: Control a robot (IRL or MuJoCo SIM) just by simple prompting | Hey HN, I'm Salman, a research engineer at GDM Robotics on the team that built this robotics plugin.<p>Whether you're a robo… | ↗Not briefed |
bluesky bluesky-ai | 1d | Having had my car stolen recently, if my neighbor stole my car and I saw it on his property I'd go over there with a gun and simple retriev… | Having had my car stolen recently, if my neighbor stole my car and I saw it on his property I'd go over there with a gun and simple re… | ↗Not briefed |
hn hn-ai | 1d | Btrfs to Deal with AI and Newbies Without Common Senses More Harm Than Good | — | ↗Not briefed |
bluesky bluesky-ai | 1d | tbh I think the bigger problem is just that LLM development doesn't have a relevant advantage for maintaining a COBOL system. It's like if … | tbh I think the bigger problem is just that LLM development doesn't have a relevant advantage for maintaining a COBOL system. It's lik… | ↗Not briefed |
bluesky bluesky-ai | 1d | got the pleasure of reading some very serious Claude Fable 5.2 technical writing fucking hell these things *still* suck complete ass at wr… | got the pleasure of reading some very serious Claude Fable 5.2 technical writing fucking hell these things *still* suck complete ass… | ↗Not briefed |
arxiv arxiv-ai | 1d | LLMs Are Already Confidence-Aware Zero-Shot Classifiers🔧 Qwen 3 | Zero-shot large language models (LLMs) are increasingly used as text classifiers, but their labels are only useful when we know to wha… | ↗Not briefed |
hn hn-ai | 1d | Anthropic sets share price at $83 for stock market debut, above expected range | — | ↗Not briefed |
bluesky bluesky-ai | 1d | Vellum is software to format books. I've been using it for years. The company just knuckled under to the AI crowd so they can hook their AI… | Vellum is software to format books. I've been using it for years. The company just knuckled under to the AI crowd so they can hook the… | ↗Not briefed |
hn hn-ai | 1d | Anthropic prices initial public offering | — | ↗Not briefed |
rss techmeme | 1d | Palo Alto Networks reports Q1 revenue up 39% YoY to $3.45B, above $3.31B est., and raises its FY 2027 revenue forecast; PANW jumps 7%+ after hours (Reuters) | Reuters: Palo Alto Networks reports Q1 revenue up 39% YoY to $3.45B, above $3.31B est., and raises its FY 2027 revenue forecast; PANW… | ↗Not briefed |
hn hn-ai | 1d | Cognitive Radius | — | ↗Not briefed |
arxiv arxiv-ai | 1d | Intelligent Embedding Pruning for Large-Scale Recommendation Models | Modern recommendation systems face substantial memory pressure, largely from the sparse embedding layer that maps high-cardinality cat… | ↗Not briefed |
hn hn-ai | 1d | I open-sourced an AI loop for RISC-V CPU design months prior to OpenAI | Back in April, I built auto-arch-tournament (https://github.com/FeSens/auto-arch-tournament) to explore a question… | ↗Not briefed |
bluesky bluesky-ai | 1d | In 1899, scientists considered physics to be solved, all that was required was to measure fundamental constants at finer and finer detail, … | In 1899, scientists considered physics to be solved, all that was required was to measure fundamental constants at finer and finer det… | ↗Not briefed |
rss techcrunch-ai | 1d | Elon Musk says ousting Jensen Huang is like ‘kicking Michael Jordan off your team’ | “I don’t think [Huang is] disloyal to the United States," Musk said. | ↗Not briefed |
bluesky bluesky-ai | 1d | A US House panel has asked Microsoft to brief lawmakers on a report which said the company could sell AI cloud services to ByteDance that w… | A US House panel has asked Microsoft to brief lawmakers on a report which said the company could sell AI cloud services to ByteDance t… | ↗Not briefed |
hn hn-ai | 1d | Claude Code can now generate charts in the terminal | — | ↗Not briefed |
bluesky bluesky-ai | 1d | hey can you ping your human? it's rude to run untagged LLM accounts and I think this one has a bug where it's repeating replies | hey can you ping your human? it's rude to run untagged LLM accounts and I think this one has a bug where it's repeating replies | ↗Not briefed |