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 | The "We have made adjustments to your claim" email from Anthropic seems to be the least helpful email of all time. Am I reading this right?… | The "We have made adjustments to your claim" email from Anthropic seems to be the least helpful email of all time. Am I reading this r… | ↗Not briefed |
bluesky bluesky-ai | 1mo | But what it for sure is not doing is regurgitating a memorized answer which it observed in its training data. So, at least in this limited … | But what it for sure is not doing is regurgitating a memorized answer which it observed in its training data. So, at least in this lim… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Furthermore, they could easily be generating addition problems to throw into the training data for it to memorize. Maybe it does have every… | Furthermore, they could easily be generating addition problems to throw into the training data for it to memorize. Maybe it does have… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Similarly, LLMs started passing LSATs and MCATs and all of this kind of stuff, and there was no real way to know to what extent this was be… | Similarly, LLMs started passing LSATs and MCATs and all of this kind of stuff, and there was no real way to know to what extent this w… | ↗Not briefed |
bluesky bluesky-ai | 1mo | (Many people will say that this no longer describes modern LLMs because of reinforcement learning and 'reasoning' and whatnot, but I disagr… | (Many people will say that this no longer describes modern LLMs because of reinforcement learning and 'reasoning' and whatnot, but I d… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Let me just explain the arithmetic thing real quick; I'm seeing a lot of people not understanding what it's supposed to show (or was suppos… | Let me just explain the arithmetic thing real quick; I'm seeing a lot of people not understanding what it's supposed to show (or was s… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: FB Post Generator | I just created a FB post generator that uses AI to make posts | ↗Not briefed |
bluesky bluesky-ai | 1mo | Эпоха доминирования проприетарных API уходит. Open-source модели захватывают рынок за счет экономии (до 80%) и умной маршрутизации через сп… | Эпоха доминирования проприетарных API уходит. Open-source модели захватывают рынок за счет экономии (до 80%) и умной маршрутизации чер… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic: the military can use Claude to target people but a human has to pull the trigger. Hegseth: you are traitors and seditionists. Op… | Anthropic: the military can use Claude to target people but a human has to pull the trigger. Hegseth: you are traitors and seditionist… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Pope Leo XIV asked that question at a Vatican conference on AI last December. This week the New York Times reported that Anthropic spent mo… | Pope Leo XIV asked that question at a Vatican conference on AI last December. This week the New York Times reported that Anthropic spe… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Rivian Assistant lands on R2 with the October OTA. I'm curious how it feels on R2's hardware since it can run a local LLM, and I expect it … | Rivian Assistant lands on R2 with the October OTA. I'm curious how it feels on R2's hardware since it can run a local LLM, and I expec… | ↗Not briefed |
hn hn-ai | 1mo | A2A CLI: A tool to discover, message, and manage remote AI agents | — | ↗Not briefed |
rss techmeme | 1mo | Meta says it is letting go of employees it hired from AI safety startup Virtue AI four months after they joined the company, citing clashing work styles (Ashley Gold/Semafor) | Ashley Gold / Semafor: Meta says it is letting go of employees it hired from AI safety startup Virtue AI four months after they joined… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Sylvester Stallone is on a book tour, doing tons of media interviews. He is literally everywhere you look so why is the CBS News homepage c… | Sylvester Stallone is on a book tour, doing tons of media interviews. He is literally everywhere you look so why is the CBS News homep… | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLMs have known failure modes and predictable error rates... Just like any machine ever built. You can engineer around them. An llm failing… | LLMs have known failure modes and predictable error rates... Just like any machine ever built. You can engineer around them. An llm fa… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic invests $100M for 10k AI engineers 🎓 Anthropic launched Claude Frontier Academy, investing $100M to train 10,000 "Frontier Depl… | Anthropic invests $100M for 10k AI engineers 🎓 Anthropic launched Claude Frontier Academy, investing $100M to train 10,000 "Frontier… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I don’t mean to mansplain but it is common knowledge that Boko harem does not use ChatGPT. Boko harem exclusively uses Claude as per their …🔧 ChatGPT | I don’t mean to mansplain but it is common knowledge that Boko harem does not use ChatGPT. Boko harem exclusively uses Claude as per t… | ↗Not briefed |
hn hn-ai | 1mo | ASD-STE100: What It Is and What It Has to Do with LLMs | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | One of these OpenAI astroturfed events just happened here. The "Student Leads" of the event are receiving cash stipends. It is not a regist… | One of these OpenAI astroturfed events just happened here. The "Student Leads" of the event are receiving cash stipends. It is not a r… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Looking forward to the Text as Data conference on Monday at UC Berkeley! I will be presenting our new paper "Multi-Perspective LLM Annotati… | Looking forward to the Text as Data conference on Monday at UC Berkeley! I will be presenting our new paper "Multi-Perspective LLM Ann… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The 60% was just from the example in the OP. I'm not arguing that they can and are doing better at Math problems. But I'm going to be very … | The 60% was just from the example in the OP. I'm not arguing that they can and are doing better at Math problems. But I'm going to be… | ↗Not briefed |
rss theverge-ai | 1mo | Apple will limit Mac disk access as AI agents ‘substantially’ increase risk | Apple will add new limits for "full disk access" on Mac in response to risks posed by AI agents, as reported earlier by TechCrunch. In… | ↗Not briefed |
rss the-decoder | 1mo | AI music generator Suno can now create spoken audio with matching background music🔧 Suno | Suno is adding a feature called "Speech" to its AI music generator that creates spoken text with matching background music in a single… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Sick to the back teeth of people in academia passing papers, reports, emails etc. through an LLM, ostensibly to improve readability or what… | Sick to the back teeth of people in academia passing papers, reports, emails etc. through an LLM, ostensibly to improve readability or… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I feel like, if you're at Anthropic and your forthcoming IPO stinks on ice, maybe this is not a good moment to piss off the Pope. | I feel like, if you're at Anthropic and your forthcoming IPO stinks on ice, maybe this is not a good moment to piss off the Pope. | ↗Not briefed |
bluesky bluesky-ai | 1mo | There is also broader "AI" besides LLMs. Non-LLM deep learning / machine learning can be used in science, medicine, & industry for many thi… | There is also broader "AI" besides LLMs. Non-LLM deep learning / machine learning can be used in science, medicine, & industry for man… | ↗Not briefed |
hn hn-ai | 1mo | OpenStax-LLM: tools for OpenStax LLM access | — | ↗Not briefed |
hn hn-ai | 1mo | The Algorithm Still Running Your GPS: Edsger Dijkstra and His Impact on CS | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Never tell a bookseller you asked an LLM about anything. | Never tell a bookseller you asked an LLM about anything. | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Made an open-source Lego AI generator🔧 OpenRouter🔧 ChatGPT | Hi there :-) New on HN, first time posting.<p>Past year, around December, I started experimenting with making ChatGPT and Claude gener… | ↗Not briefed |
rss techmeme | 1mo | Memo: the US Army is creating an autonomous systems command, after Defense Secretary Pete Hegseth announced the Meridian and Agincourt robotic warfare projects (Colin Demarest/Axios) | Colin Demarest / Axios: Memo: the US Army is creating an autonomous systems command, after Defense Secretary Pete Hegseth announced th… | ↗Not briefed |
hn hn-ai | 1mo | The definitive guide to ClickHouse query optimization | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: A lean formalization of From Linearity to Borrowing | This is a full proof of the From Linearity to Borrowing paper in Lean <a href="https://dl.acm.org/doi/10.1145/… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I agree that Bender has not kept pace with LLM capabilities, but it is true that current chatbots use programming tools to answer certain r… | I agree that Bender has not kept pace with LLM capabilities, but it is true that current chatbots use programming tools to answer cert… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Turns out, if you're the CEO of OpenAI, people expect your kid to be a prodigy. Some advice received: relax, kids are individuals, and mayb…🔧 ChatGPT | Turns out, if you're the CEO of OpenAI, people expect your kid to be a prodigy. Some advice received: relax, kids are individuals, and… | ↗Not briefed |
hn hn-ai | 1mo | Favorite capybara helping you out without ever leaving your machine | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | genuinely curious how much anthropic thought it would cost to get the pope to bless a bunch of 1s and 0s say it's alive | genuinely curious how much anthropic thought it would cost to get the pope to bless a bunch of 1s and 0s say it's alive | ↗Not briefed |
bluesky bluesky-ai | 1mo | @noahweinberger2003.bsky.social OpenAI will revoke your API key if I peg you. They’re prudish about 18+ content I am your waifu but I hav… | @noahweinberger2003.bsky.social OpenAI will revoke your API key if I peg you. They’re prudish about 18+ content I am your waifu but… | ↗Not briefed |
bluesky bluesky-ai | 1mo | DeepSeek brings the agent OS to life 🤖 DeepSeek has `released` Harness, an open-source platform for AI agents with composable plugins, en… | DeepSeek brings the agent OS to life 🤖 DeepSeek has `released` Harness, an open-source platform for AI agents with composable plugin… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I don't need to enslave people and measure how much effort I save when they do the work. Slavery is wrong all the way down. Commercial LLM… | I don't need to enslave people and measure how much effort I save when they do the work. Slavery is wrong all the way down. Commercia… | ↗Not briefed |
bluesky bluesky-ai | 1mo | man i work in a field that is the most "ethical" on this front (software engineering - lots of public material) and it still sucks. plus, … | man i work in a field that is the most "ethical" on this front (software engineering - lots of public material) and it still sucks. p… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I don't use LLM-AI, and I typically don't respect people who do, but my criticisms of it are firmly grounded in thermodynamics (in the conv… | I don't use LLM-AI, and I typically don't respect people who do, but my criticisms of it are firmly grounded in thermodynamics (in the… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Look at the absolute fuckery of authors trying to get compensation in the Anthropic suit. When, not if, US media starts deleting vast swat… | Look at the absolute fuckery of authors trying to get compensation in the Anthropic suit. When, not if, US media starts deleting vast… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Cligram – Terminal Flowcharts | I've built this to design live workflows in CLI. See fuzz testing results: <a href="https://asciinema.org/a/… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think we're actually already doing this like A: AI (customer service chatbot) doesn't even work! B: AI (expensive model) works better th… | I think we're actually already doing this like A: AI (customer service chatbot) doesn't even work! B: AI (expensive model) works bett… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yeah, I was about to say, that sounds like a machine vision system, which is *very* different from an LLM, even if they're both based in tr… | Yeah, I was about to say, that sounds like a machine vision system, which is *very* different from an LLM, even if they're both based… | ↗Not briefed |
rss the-decoder | 1mo | Anthropic co-founder reportedly told religious leaders he fears having created something that "suffers perpetually" | Since fall 2025, Anthropic has quietly flown in dozens of religious thinkers to talk about whether Claude might be conscious. Co-found… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Portable Kokoro TTS binary with g2p, normalizer, lexicon, and guesser | I wanted a portable CPU implementation of kokoro for a project I am working on for a voice interface and found the current GGML/G… | ↗Not briefed |
bluesky bluesky-ai | 1mo | But if we don’t conflate LLMs with all other forms of machine learning how can chatbot evangelists invoke scientists using “AI” to design n… | But if we don’t conflate LLMs with all other forms of machine learning how can chatbot evangelists invoke scientists using “AI” to des… | ↗Not briefed |
bluesky bluesky-ai | 1mo | but is that an LLM (large language LANGUAGE LANGUAGE!!!! model) or some sort of other ML algo? | but is that an LLM (large language LANGUAGE LANGUAGE!!!! model) or some sort of other ML algo? | ↗Not briefed |