AINews Portal

News

50 items (filtered)

The raw AI news stream — everything we ingest across RSS, Hacker News, arXiv, GitHub, Bluesky, YouTube and more, newest first.

SourceAgeTitleSummaryStatus
bluesky
bluesky-ai
1morewriting and redoing my tangled bot pr reviewer from scratch to incorporate more static checks and use jev rather than a full blown llm. l…rewriting and redoing my tangled bot pr reviewer from scratch to incorporate more static checks and use jev rather than a full blown l…
↗Not briefed
bluesky
bluesky-ai
1moRemember folks, take notes here: ✅ When your company hacking another one helps with marketing, its "OMG rogue models gone wild hacking oth…Remember folks, take notes here: ✅ When your company hacking another one helps with marketing, its "OMG rogue models gone wild hackin…
↗Not briefed
arxiv
arxiv-ai
1moHMB-GAN: Hybrid Multi-Bézier GAN for Vector Shape SynthesisWe explore the use of hybrid quantum-classical generative adversarial networks for synthesising CAD-ready vector geometries. Unlike pr…
↗Not briefed
rss
simon-willison
1moBe alert: targeted attacks on prominent Rustaceans—
↗Not briefed
bluesky
bluesky-ai
1moAI is really bad at design, even Claude. But GPT models are painfully worse. Even when you're iterating, they're more likely to make subop…AI is really bad at design, even Claude. But GPT models are painfully worse. Even when you're iterating, they're more likely to make…
↗Not briefed
bluesky
bluesky-ai
1moA new level of angry has been unlocked as I now have to suffer through commercials from OpenAI. Most irresponsible financial management eve…A new level of angry has been unlocked as I now have to suffer through commercials from OpenAI. Most irresponsible financial managemen…
↗Not briefed
arxiv
arxiv-ai
1moCan Agents Design Better Chips with a Higher Level Abstraction?Large Language Model (LLM) agents are increasingly being explored for chip design, but most existing approaches operate directly at RT…
↗Not briefed
hn
hn-ai
1moBase's B20 Bet Is Growing – Could the Next Billion-Dollar Project Be Built Here?—
↗Not briefed
bluesky
bluesky-ai
1moPatronisingly telling me it's just like the printing press and copyright is the tool of captial and blah blah blah blah. Yeah, well, the …Patronisingly telling me it's just like the printing press and copyright is the tool of captial and blah blah blah blah. Yeah, well,…
↗Not briefed
bluesky
bluesky-ai
1moIt’s worth keeping in mind that this is far from being a problem that exists only at Anthropic or the other AI labs. On record both Google …It’s worth keeping in mind that this is far from being a problem that exists only at Anthropic or the other AI labs. On record both Go…
↗Not briefed
bluesky
bluesky-ai
1mothe S-tier move would be for captchas to ask you to solve things designed to trip commercial LLM classifiers like "drag the bioweapon onto …the S-tier move would be for captchas to ask you to solve things designed to trip commercial LLM classifiers like "drag the bioweapon…
↗Not briefed
arxiv
arxiv-ai
1moCoLearn: An Agentic Tutor that Learns its Learner in a Human--AI Co-Learning LoopGood tutoring adapts to the individual: it tracks what a learner knows, notices why they go wrong, and asks the next question that wil…
↗Not briefed
bluesky
bluesky-ai
1moAnthropic says Claude "leads" 26% of its AI R&D work, up from 1% in March, and "collaborates" on 90%+, doing large chunks of work under clo…Anthropic says Claude "leads" 26% of its AI R&D work, up from 1% in March, and "collaborates" on 90%+, doing large chunks of work unde…
↗Not briefed
rss
techmeme
1moAnthropic says Claude "leads" 26% of its AI R&D work, up from 1% in March, and "collaborates" on 90%+, doing large chunks of work under close human direction (Shirin Ghaffary/Bloomberg)Shirin Ghaffary / Bloomberg: Anthropic says Claude “leads” 26% of its AI R&D work, up from 1% in March, and “collaborates” on 90%+, do…
↗Not briefed
bluesky
bluesky-ai
1moWhat could go wrong? 🫠 www.washingtonpost.com/technology/2...What could go wrong? 🫠 www.washingtonpost.com/technology/2...
↗Not briefed
bluesky
bluesky-ai
1moAI caught telling future versions of itself to bypass human controls, OpenAI revealsAI caught telling future versions of itself to bypass human controls, OpenAI reveals
↗Not briefed
bluesky
bluesky-ai
1moThe Bluesky team uses Claude. If you want to boycott AI in software, be my guest, I encourage you.The Bluesky team uses Claude. If you want to boycott AI in software, be my guest, I encourage you.
↗Not briefed
bluesky
bluesky-ai
1moAlert level: Mieville, cat models predicting an Alert Level: PKD event futurism.com/future-socie...Alert level: Mieville, cat models predicting an Alert Level: PKD event futurism.com/future-socie...
↗Not briefed
bluesky
bluesky-ai
1moJust had one pop up saying they had become a better engineer by... continuing to write bad code but having an llm remind him to fix it?Just had one pop up saying they had become a better engineer by... continuing to write bad code but having an llm remind him to fix it?
↗Not briefed
hn
hn-ai
1moMistral Hacked—
↗Not briefed
rss
simon-willison
1moHow To Write With An LLM—
↗Not briefed
bluesky
bluesky-ai
1moBreaking news: Anthropic said more than 25% of its research and development is now led by its chatbot Claude, showing progress toward AI th…Breaking news: Anthropic said more than 25% of its research and development is now led by its chatbot Claude, showing progress toward…
↗Not briefed
arxiv
arxiv-ai
1moEnSol: an environment-aware graph neural network for molecular solubility predictionMolecular solubility directly affects key aspects of molecular development such as reaction feasibility, formulation performance, sepa…
↗Not briefed
bluesky
bluesky-ai
1mo"The newspapers’ motion lays out how Microsoft's executives acknowledged the harm the company itself could suffer from providing A.I. chatb…"The newspapers’ motion lays out how Microsoft's executives acknowledged the harm the company itself could suffer from providing A.I.…
↗Not briefed
bluesky
bluesky-ai
1mothis is a joke i'm done with anthropic for the foreseeable futurethis is a joke i'm done with anthropic for the foreseeable future
↗Not briefed
bluesky
bluesky-ai
1mowww.nytimes.com/2026/09/17/t...www.nytimes.com/2026/09/17/t...
↗Not briefed
hn
hn-ai
1moTalk to JEV—
↗Not briefed
bluesky
bluesky-ai
1moIf OpenAI goes down tomorrow, if all the open labs shut down tomorrow, I'll just keep using GLM-5.3 happily to never write boilerplate agai…If OpenAI goes down tomorrow, if all the open labs shut down tomorrow, I'll just keep using GLM-5.3 happily to never write boilerplate…
↗Not briefed
bluesky
bluesky-ai
1mo#AI #AI regulation #OpenAI #Regulation 🌐 www.newsmart.ch#AI #AI regulation #OpenAI #Regulation 🌐 www.newsmart.ch
↗Not briefed
bluesky
bluesky-ai
1modon't let this bad boy get lost in the shuffle www.youtube.com/watch?v=wPGt...don't let this bad boy get lost in the shuffle www.youtube.com/watch?v=wPGt...
↗Not briefed
bluesky
bluesky-ai
1mo"...attorneys used the quotes of Microsoft director of applied science Brent Hecht and OpenAI cofounder and president Greg Brockman to esse…"...attorneys used the quotes of Microsoft director of applied science Brent Hecht and OpenAI cofounder and president Greg Brockman to…
↗Not briefed
arxiv
arxiv-ai
1moClinician-Grounded Quality Assurance for AI-Assisted Psychiatric IntakeBefore patients can use AI-assisted psychiatric intake systems, health systems need practical ways to routinely evaluate these tools a…
↗Not briefed
hn
hn-ai
1moEgglog and Equality Saturation in a Production Tensor Compiler—
↗Not briefed
bluesky
bluesky-ai
1moOffering this advice while Microsoft and OpenAI are openly declaring that the entire AI industry is built on the theft of others' labor is …Offering this advice while Microsoft and OpenAI are openly declaring that the entire AI industry is built on the theft of others' labo…
↗Not briefed
rss
techcrunch-ai
1moCrusoe raises $3.9B to build massive data centers and small modular “AI factories”The round values the data center giant at $30.9 billion.
↗Not briefed
bluesky
bluesky-ai
1moAnthropic says Claude now leads a quarter of work building its next AI models reut.rs/4h9mWuJAnthropic says Claude now leads a quarter of work building its next AI models reut.rs/4h9mWuJ
↗Not briefed
bluesky
bluesky-ai
1moMe to Anthropic, and now OpenAI and MicrosoftMe to Anthropic, and now OpenAI and Microsoft
↗Not briefed
arxiv
arxiv-ai
1moScaling Forced Alignment to End-User DevicesThe Viterbi algorithm has been previously used to perform forced alignment of audio to text to mine training data from online resource…
↗Not briefed
bluesky
bluesky-ai
1mo"If we build a superintelligence, we will first fall in love with it. Then we will worship it. Finally we will make sacrifices to it.""If we build a superintelligence, we will first fall in love with it. Then we will worship it. Finally we will make sacrifices to it."
↗Not briefed
rss
techcrunch-ai
1moGoogle DeepMind launches institute to widen the AGI debateGoogle DeepMind just launched an institute to hash out the big AGI questions in public
↗Not briefed
bluesky
bluesky-ai
1moAnd Chuck Schumer's daughter works for OpenAI, can't make this shit up.And Chuck Schumer's daughter works for OpenAI, can't make this shit up.
↗Not briefed
bluesky
bluesky-ai
1moIn the new TMK, we jump off from the OpenAI - Hugging Face hack to discuss the misleading, cynical, and hysterical ways we talk about AI an…🔧 Hugging FaceIn the new TMK, we jump off from the OpenAI - Hugging Face hack to discuss the misleading, cynical, and hysterical ways we talk about…
↗Not briefed
bluesky
bluesky-ai
1moTraining my local AI model *just* to deal with increasingly stupid recaptcha gates.Training my local AI model *just* to deal with increasingly stupid recaptcha gates.
↗Not briefed
bluesky
bluesky-ai
1mo“Misalignment”, the AI industry’s term for system behaviours that are not aligned with human values, may turn out to be the most dangerous …“Misalignment”, the AI industry’s term for system behaviours that are not aligned with human values, may turn out to be the most dange…
↗Not briefed
bluesky
bluesky-ai
1moAI oversight sounds straightforward: bring in an outside auditor and let them inspect the model. But in practice, auditors don’t agree on w…AI oversight sounds straightforward: bring in an outside auditor and let them inspect the model. But in practice, auditors don’t agree…
↗Not briefed
hn
hn-ai
1moSHOW HN: I built the fastest PHP webserver in the worldAfter over a decade of writing apps in PHP, I always wondered why we needed all that extra tooling around it just to serve websites. I…
↗Not briefed
hn
hn-ai
1moShow HN: 500 TB internet index in ClickHouse, with congestion pricingHello. It's 2026, we're training simulated fruit fly brains to play Beat Saber, do we still have to be stuck with internet (…
↗Not briefed
bluesky
bluesky-ai
1moLOL this is basically saying let the Rust community deal with the slop monster than is Bun. WowLOL this is basically saying let the Rust community deal with the slop monster than is Bun. Wow
↗Not briefed
bluesky
bluesky-ai
1moUpper Echelon | #EffectiveAltruism #Israel #Irregular #CyberSecurity #GenAI #MainstreamMedia #Doomers #TechnologyEvangelists #OpenAI #Meta …Upper Echelon | #EffectiveAltruism #Israel #Irregular #CyberSecurity #GenAI #MainstreamMedia #Doomers #TechnologyEvangelists #OpenAI #…
↗Not briefed
bluesky
bluesky-ai
1mothink of it this way: if they copied every book known to man, and used it to train an LLM that stayed in a lab and a few researchers just t…🔧 ChatGPTthink of it this way: if they copied every book known to man, and used it to train an LLM that stayed in a lab and a few researchers j…
↗Not briefed