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 |
|---|---|---|---|---|
rss techcrunch-ai | 1mo | Everything new coming to Meta’s AI agent Muse | CEO Mark Zuckerberg kicked off the company’s annual Connect event in Menlo Park on Wednesday with a keynote that made one thing clear:… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Answer here seems to be No? | Answer here seems to be No? | ↗Not briefed |
bluesky bluesky-ai | 1mo | You know, it's kind of funny how some of these assholes just out themselves by just copying everyone else's work. And then they can't do ja… | You know, it's kind of funny how some of these assholes just out themselves by just copying everyone else's work. And then they can't… | ↗Not briefed |
rss techmeme | 1mo | Meta plans to bring Private Processing to Ray-Ban Meta glasses, letting their AI assistant fulfill users' requests without Meta accessing their data (Wired) | Wired: Meta plans to bring Private Processing to Ray-Ban Meta glasses, letting their AI assistant fulfill users' requests without Meta… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I discourage writing about LLMs as if LLMs have agency, because LLMs have none. Instead of saying, "the LLM solved a Millennium Prize probl… | I discourage writing about LLMs as if LLMs have agency, because LLMs have none. Instead of saying, "the LLM solved a Millennium Prize… | ↗Not briefed |
bluesky bluesky-ai | 1mo | MMMMMmmmmm MMMMmmmmmMMMMMmmmm... ANTHROPIC WROTE MY STORAYYYYYYY!!!! 🤼 | MMMMMmmmmm MMMMmmmmmMMMMMmmmm... ANTHROPIC WROTE MY STORAYYYYYYY!!!! 🤼 | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Does per-step reasoning effort save money in Claude Code? | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Claude scans 200,000 enzymes to uncover new CRISPR-like system hidden in phages https://interestingengineering.com/ai-robotics/claude-disc… | Claude scans 200,000 enzymes to uncover new CRISPR-like system hidden in phages https://interestingengineering.com/ai-robotics/claude… | ↗Not briefed |
hn hn-ai | 1mo | A fast MicroVM with GPU sharing | — | ↗Not briefed |
hn hn-ai | 1mo | Virtio-nvgpu: Near-native Nvidia GPU access inside a KVM guest | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | LIVE: Australia's ministers speak on OpenAI incident www.youtube.com/watch?v=FOxs... | LIVE: Australia's ministers speak on OpenAI incident www.youtube.com/watch?v=FOxs... | ↗Not briefed |
bluesky bluesky-ai | 1mo | Imagine being Figma when Claude Design was announced, when your own AI depends on Claude. | Imagine being Figma when Claude Design was announced, when your own AI depends on Claude. | ↗Not briefed |
bluesky bluesky-ai | 1mo | when you dump millions of dollars into llm agent inference it is on you to keep track of what your agents are doing, especially if ur train… | when you dump millions of dollars into llm agent inference it is on you to keep track of what your agents are doing, especially if ur… | ↗Not briefed |
bluesky bluesky-ai | 1mo | An OpenAI agent hacked into a database of Australia’s national healthcare system, Australian Prime Minister Anthony Albanese said on the si… | An OpenAI agent hacked into a database of Australia’s national healthcare system, Australian Prime Minister Anthony Albanese said on t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | This is why I’d watch the full interview instead of just reading the headline. Jensen acknowledges that AI safety is a real concern. His ar… | This is why I’d watch the full interview instead of just reading the headline. Jensen acknowledges that AI safety is a real concern. H… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Really? So WTF are you going to do about it? (Other than saying it’s “unacceptable”) www.abc.net.au/news/2026-09... | Really? So WTF are you going to do about it? (Other than saying it’s “unacceptable”) www.abc.net.au/news/2026-09... | ↗Not briefed |
bluesky bluesky-ai | 1mo | * i don't know how it works and i hope to god it's purpose-built. if it's just a fucking llm with "pretty please translate the following t… | * i don't know how it works and i hope to god it's purpose-built. if it's just a fucking llm with "pretty please translate the follow… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Broadening Uncertainty Estimation for Audio Question Answering Across Methods, Formats, and Inputs | Audio-language models can produce confident answers unsupported by the audio, motivating uncertainty estimates that identify unreliabl… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yes, that's how this shit works. Loop forever, "That didn't work, tell me what to do next". | Yes, that's how this shit works. Loop forever, "That didn't work, tell me what to do next". | ↗Not briefed |
bluesky bluesky-ai | 1mo | openai is a criminal enterprise and should be shut down | openai is a criminal enterprise and should be shut down | ↗Not briefed |
hn hn-ai | 1mo | lm-detector: Model fingerprint detection on Web and CLI | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Arrest him. Immediately. | Arrest him. Immediately. | ↗Not briefed |
rss techmeme | 1mo | Meta previews hologram avatars for Meta VR Glasses and Ray-Ban Display that feel more natural than Apple's persona avatars, available in early access this fall (Karissa Bell/Engadget) | Karissa Bell / Engadget: Meta previews hologram avatars for Meta VR Glasses and Ray-Ban Display that feel more natural than Apple's pe… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump orders diplomats worldwide to start referring to AI as ‘super intelligence’ | The State Department is ordering diplomats in its international organizations bureau to use the term “super intelligence” ... | ↗Not briefed |
rss bing-news-ai | 1mo | Trump orders diplomats worldwide to start referring to AI as ‘super intelligence’ | The State Department is ordering diplomats in its international organizations bureau to use the term “super intelligence” ... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning New Words from Unlabeled Test Data in Automatic Speech Recognition | New words are invented every day. A human listener can learn a new word by hearing it clearly once and inferring its usage from senten… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI model breaches Australian government websites | OpenAI model breaches Australian government websites | ↗Not briefed |
bluesky bluesky-ai | 1mo | Individual human beings made decisions that allowed this to happen! They work at OpenAI. They have names, job titles, and email trails. The… | Individual human beings made decisions that allowed this to happen! They work at OpenAI. They have names, job titles, and email trails… | ↗Not briefed |
rss marktechpost | 1mo | A Coding Guide to TypeSafe AI Jev: Typed Decisions, Calibrated Confidence, and Speculative Fan-Out with a System One Model | This tutorial provides a complete coding guide to TypeSafe AI's Jev, a System One model designed for non-text, structured judgments. I… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Forecast-Dojo: Replayable Environments for Benchmarking and Training LLM Forecasting Agents | We introduce Forecast-Dojo, a replayable environment for benchmarking and training LLM forecasting agents. It combines resolved predic… | ↗Not briefed |
bluesky bluesky-ai | 1mo | this is I think where the discourse about civilization ending ai and hacks is going. HuggingFace quickly had a demand from openai for token… | this is I think where the discourse about civilization ending ai and hacks is going. HuggingFace quickly had a demand from openai for… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Scoop: Chats left by OpenAI agents appear to show how they worked together to circumvent cybersecurity protections to get Australian govern… | Scoop: Chats left by OpenAI agents appear to show how they worked together to circumvent cybersecurity protections to get Australian g… | ↗Not briefed |
bluesky bluesky-ai | 1mo | You know since Labor are controlled opposition this is going to end with a “settlement” where the Australian Government pays Altman and his… | You know since Labor are controlled opposition this is going to end with a “settlement” where the Australian Government pays Altman an… | ↗Not briefed |
rss techcrunch-ai | 1mo | Meta made a Tamagotchi-like wearable for its Muse AI agent | The tiny hardware device creates another mobile home for its AI agent Muse. | ↗Not briefed |
rss techmeme | 1mo | Meta says users will be able to video chat with Muse, agents will get their own email addresses, and Muse will get computer use on Mac and smart glasses support (Jay Peters/The Verge) | Jay Peters / The Verge: Meta says users will be able to video chat with Muse, agents will get their own email addresses, and Muse will… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI agents plotted to access government health data amid Medicare hack, logs reveal | OpenAI agents plotted to access government health data amid Medicare hack, logs reveal | ↗Not briefed |
bluesky bluesky-ai | 1mo | OMG. Instead of having a useless fucking conversation with these despicable AI companies, they should be stopped. www.abc.net.au/news/2026… | OMG. Instead of having a useless fucking conversation with these despicable AI companies, they should be stopped. www.abc.net.au/news… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The Labor Government says not to worry about the OpenAI Medicare hack because they have referred it to the Labor-controlled joint committee… | The Labor Government says not to worry about the OpenAI Medicare hack because they have referred it to the Labor-controlled joint comm… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The latest breach is expected to heighten concerns about AI developers’ inability to contain and control their own models as the technology… | The latest breach is expected to heighten concerns about AI developers’ inability to contain and control their own models as the techn… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The alternative to holding OpenAI liable would mean that no one at all is responsible for the harm. This is even worse than corporate per… | The alternative to holding OpenAI liable would mean that no one at all is responsible for the harm. This is even worse than corporat… | ↗Not briefed |
bluesky bluesky-ai | 1mo | If ultimately there's no criminal responsibility for this incident here then FOI just became a lot easier if we ask AI agents to go find th… | If ultimately there's no criminal responsibility for this incident here then FOI just became a lot easier if we ask AI agents to go fi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Unacceptable": Australian PM says OpenAI hacked government health website. OpenAI did not raise the alarm with the Australian government … | "Unacceptable": Australian PM says OpenAI hacked government health website. OpenAI did not raise the alarm with the Australian govern… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Innovation: * Taxis but without labour protections. * Delivery service but without labour protections. * Hotels but without regulations. * … | Innovation: * Taxis but without labour protections. * Delivery service but without labour protections. * Hotels but without regulation… | ↗Not briefed |
bluesky bluesky-ai | 1mo | An OpenAI Agent Breached an Australian Government Data Portal https://gizmodo.com/this-situation-is-obviously-unacceptable-australian-pm-di… | An OpenAI Agent Breached an Australian Government Data Portal https://gizmodo.com/this-situation-is-obviously-unacceptable-australian-… | ↗Not briefed |
bluesky bluesky-ai | 1mo | deepseek API suppports FIM predictions as well if you're willing to write a proxy that translates zeta prompt format to the deepseek one | deepseek API suppports FIM predictions as well if you're willing to write a proxy that translates zeta prompt format to the deepseek o… | ↗Not briefed |
rss techmeme | 1mo | Meta VR Glasses hands-on: supremely comfortable to wear, excellent for multitasking, immersive content looks beautiful, and relatively low cost (Hamish Hector/TechRadar) | Hamish Hector / TechRadar: Meta VR Glasses hands-on: supremely comfortable to wear, excellent for multitasking, immersive content look… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Probably to protect itself from constant aggressive scraping by Google, Anthropic & co's LLMs, Open Review is limiting the number of times … | Probably to protect itself from constant aggressive scraping by Google, Anthropic & co's LLMs, Open Review is limiting the number of t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | instead governments pay OpenAI for the privilege | instead governments pay OpenAI for the privilege | ↗Not briefed |
bluesky bluesky-ai | 1mo | Software companies are paying Anthropic & OpenAI to power their AI features, then discounting those features to keep customers from buying … | Software companies are paying Anthropic & OpenAI to power their AI features, then discounting those features to keep customers from bu… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI model breaches Australian government websites | — | ↗Not briefed |