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
arxiv
arxiv-ai
2moCertifying Lower Bounds for Risk-Sensitive Reinforcement Learning under Adversarial State PerturbationsReinforcement learning (RL) agents deployed in real-world environments are often vulnerable to adversarial perturbations in state obse…
↗Not briefed
arxiv
arxiv-ai
2moFlow Duality and Source Geometry for Categorical GenerationContinuous and discrete flow matching are usually treated as separate constructions. This paper identifies a duality between them: pro…
↗Not briefed
bluesky
bluesky-ai
2moThe release of 4o is going to keep inflicting well-deserved reputational damage on OpenAI for a long, long time.The release of 4o is going to keep inflicting well-deserved reputational damage on OpenAI for a long, long time.
↗Not briefed
bluesky
bluesky-ai
2moI think it's wrong that OpenAI doesn't intend to claim the Millennium Prize for Navier–Stokes. They should give that million dollars to th…I think it's wrong that OpenAI doesn't intend to claim the Millennium Prize for Navier–Stokes. They should give that million dollars…
↗Not briefed
bluesky
bluesky-ai
2moinventing a new ai benchmarking site: cube bench. how well can a model rotate a cube.inventing a new ai benchmarking site: cube bench. how well can a model rotate a cube.
↗Not briefed
hn
hn-ai
2moOpenAI's apparent maths breakthrough raises profound questions—
↗Not briefed
hn
hn-ai
2moSchools Get 5M Complaints yet Almost No Teachers Trained to Handle Them—
↗Not briefed
bluesky
bluesky-ai
2mo“Our default is that when a student is turning in work, it shouldn’t be that of Claude. If AI is doing the thinking and writing for them, w…“Our default is that when a student is turning in work, it shouldn’t be that of Claude. If AI is doing the thinking and writing for th…
↗Not briefed
bluesky
bluesky-ai
2moA researcher at Anthropic who trained new A.I. models has resigned his position, issuing a dire warning that has gone viral.A researcher at Anthropic who trained new A.I. models has resigned his position, issuing a dire warning that has gone viral.
↗Not briefed
bluesky
bluesky-ai
2moMathematicians have been trying to solve the Navier-Stokes problem for at least a century. OpenAI says a team of its agents have done it in…Mathematicians have been trying to solve the Navier-Stokes problem for at least a century. OpenAI says a team of its agents have done…
↗Not briefed
bluesky
bluesky-ai
2mo"Why did he leave the company then" A lot of start ups are ex employees of bigger companies and everyone working dedicatedly on AI has mi…"Why did he leave the company then" A lot of start ups are ex employees of bigger companies and everyone working dedicatedly on AI h…
↗Not briefed
bluesky
bluesky-ai
2moALL the physical hardware involved in this debacle, gpus, servers, networking gear, the racks themselves, should be seized from openai as p…ALL the physical hardware involved in this debacle, gpus, servers, networking gear, the racks themselves, should be seized from openai…
↗Not briefed
arxiv
arxiv-ai
2moProject Qualia: Recovering Experiential Music Structure from Session Co-occurrence DataThis report presents results from Project Qualia, an ongoing effort to determine whether experiential similarity between songs, a stru…
↗Not briefed
hn
hn-ai
2moVibe2phone: Test your vibe coded iPhone apps from anywhere on your phone—
↗Not briefed
bluesky
bluesky-ai
2moI've been off social media this week because I'm doing loads of interviews ahead of the launch of my book next month. I just emerged and...…I've been off social media this week because I'm doing loads of interviews ahead of the launch of my book next month. I just emerged a…
↗Not briefed
bluesky
bluesky-ai
2mothe software quality crisis absolutely predates llm coding. i know bc i've been complaining about it and it's been getting steadily worse s…the software quality crisis absolutely predates llm coding. i know bc i've been complaining about it and it's been getting steadily wo…
↗Not briefed
rss
techmeme
2moSlack message: Automattic CEO Matt Mullenweg says CFO Mark Davies "conspired" with board members to put him on paid leave and Davies was voted to be interim CEO (Samantha Cole/404 Media)Samantha Cole / 404 Media: Slack message: Automattic CEO Matt Mullenweg says CFO Mark Davies “conspired” with board members to put him…
↗Not briefed
bluesky
bluesky-ai
2moi've been using it quite a lot for code but would NEVER use it to write what it *is* good for is developing tools/systems to help organize…i've been using it quite a lot for code but would NEVER use it to write what it *is* good for is developing tools/systems to help org…
↗Not briefed
bluesky
bluesky-ai
2moIf only someone had predicted this…💀 www.axios.com/2026/09/09/a...If only someone had predicted this…💀 www.axios.com/2026/09/09/a...
↗Not briefed
hn
hn-ai
2moOpen-source agentic terminal with a PTY daemon and browser access—
↗Not briefed
bluesky
bluesky-ai
2moSo we've got less than 4 years to live or move to a different planet, because of "The smartest" people on the planet! - READ: FUCKING IDIOT…So we've got less than 4 years to live or move to a different planet, because of "The smartest" people on the planet! - READ: FUCKING…
↗Not briefed
bluesky
bluesky-ai
2mothe actual story here is that openai as a company does not take cybersecurity seriously and lets their **employees** perform attacks using …the actual story here is that openai as a company does not take cybersecurity seriously and lets their **employees** perform attacks u…
↗Not briefed
bluesky
bluesky-ai
2moSo they're making a movie about OpenAI with Andrew Garfield as the lead but it sure looks like it's written like a horror film, psychologic…So they're making a movie about OpenAI with Andrew Garfield as the lead but it sure looks like it's written like a horror film, psycho…
↗Not briefed
bluesky
bluesky-ai
2moFormer Anthropic employee Jacob Coxon made a blunt admission: "The people building AI earnestly believe that it could kill us all by the en…Former Anthropic employee Jacob Coxon made a blunt admission: "The people building AI earnestly believe that it could kill us all by t…
↗Not briefed
bluesky
bluesky-ai
2moThat’s pretty much where I’ve ended up. An LLM is not and never will be “real AI”. But extensive interaction with one seems to pose signifi…That’s pretty much where I’ve ended up. An LLM is not and never will be “real AI”. But extensive interaction with one seems to pose si…
↗Not briefed
bluesky
bluesky-ai
2mov6 still can’t capture the “natural imperfections” of real human music. www.theverge.com/ai-artificia...v6 still can’t capture the “natural imperfections” of real human music. www.theverge.com/ai-artificia...
↗Not briefed
bluesky
bluesky-ai
2moI'm more worried about how many seemingly serious people take these aimless ramblings with zero substance seriously at all, they're the one…I'm more worried about how many seemingly serious people take these aimless ramblings with zero substance seriously at all, they're th…
↗Not briefed
rss
techmeme
2moSource: Meta AI researcher Andrew Tulloch, who came over from Thinking Machine Lab as one of the highest paid employees last year, is leaving the company (Reed Albergotti/Semafor)Reed Albergotti / Semafor: Source: Meta AI researcher Andrew Tulloch, who came over from Thinking Machine Lab as one of the highest pa…
↗Not briefed
bluesky
bluesky-ai
2moThis makes a certain amount of sense to me because humans are very language and meaning focused, and LLMs really fuck with language and mea…This makes a certain amount of sense to me because humans are very language and meaning focused, and LLMs really fuck with language an…
↗Not briefed
bluesky
bluesky-ai
2moby which I mean: there seem to be (at least) two categories of people in tech right now: 1. People who are morally burned out because LLM a…by which I mean: there seem to be (at least) two categories of people in tech right now: 1. People who are morally burned out because…
↗Not briefed
rss
techcrunch-ai
2moMassachusetts hits data centers with new clean power rulesMassachusetts has become the third state in as many months to slap new restrictions on data center development.
↗Not briefed
arxiv
arxiv-ai
2moNo-Box Vulnerability Analysis: Description-only Detection of Indirect Prompt Injection Vulnerabilities in MCP ServersConventional vulnerability analysis relies on either system access or dynamic interaction, all of which may be unavailable to third-pa…
↗Not briefed
hn
hn-ai
2moAustralia to impose digital duty of care on tech companies to reduce online harm—
↗Not briefed
bluesky
bluesky-ai
2moI really think posts like the one in the screenshot are just them asking an LLM for comebacks. It has to be. Why wouldn’t they do that wh…I really think posts like the one in the screenshot are just them asking an LLM for comebacks. It has to be. Why wouldn’t they do th…
↗Not briefed
rss
theverge-ai
2moSuno releases its first AI music model made with record industry help🔧 SunoSuno's new v6 AI music model is its first made with support from the record industry. Suno's Jack Brody told The Verge that v6 was "tr…
↗Not briefed
bluesky
bluesky-ai
2mothe thing about AGI wiping out all life on the planet is that none of these people have explained what that is, how it works, how it plays …the thing about AGI wiping out all life on the planet is that none of these people have explained what that is, how it works, how it p…
↗Not briefed
hn
hn-ai
2moAsk HN: Does Gusto drop dependents randomly for anyone else?I’m not sure if I’m holding it wrong or if Gusto is particularly bad about this. Small startup, last enrollment they dropped my wife,…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic Researchers Raise Alarm Over A.I. Acceleration Their warnings echo concerns that other artificial intelligence experts have voic…Anthropic Researchers Raise Alarm Over A.I. Acceleration Their warnings echo concerns that other artificial intelligence experts have…
↗Not briefed
arxiv
arxiv-ai
2moAre We Really Doing Few-Shot Learning? A Critical Examination of Pre-Training AssumptionsFew-shot learning is commonly evaluated under protocols that pre-train a model on a large auxiliary set whose classes are disjoint fro…
↗Not briefed
bluesky
bluesky-ai
2moEspecially with "AI" 🤬 I guess I need to get a new older model soon with as little "AI" built-in as possible, before they all go away fore…Especially with "AI" 🤬 I guess I need to get a new older model soon with as little "AI" built-in as possible, before they all go away…
↗Not briefed
bluesky
bluesky-ai
2moA new movie, starring Andrew Garfield as OpenAI CEO Sam Altman, could reveal whether Hollywood’s tech-bro-sympathetic narrative has changed…A new movie, starring Andrew Garfield as OpenAI CEO Sam Altman, could reveal whether Hollywood’s tech-bro-sympathetic narrative has ch…
↗Not briefed
bluesky
bluesky-ai
2moWAIT IS IT THISSSSSWAIT IS IT THISSSSS
↗Not briefed
bluesky
bluesky-ai
2moOpenAI claims huge maths breakthrough on a famed ‘Millennium Problem’ but the claim (announced by press release & a preprint on their own s…OpenAI claims huge maths breakthrough on a famed ‘Millennium Problem’ but the claim (announced by press release & a preprint on their…
↗Not briefed
bluesky
bluesky-ai
2moOne of the trickiest problems in mathematics has now fallen to AI after just days of work. The groundbreaking result was announced after si…One of the trickiest problems in mathematics has now fallen to AI after just days of work. The groundbreaking result was announced aft…
↗Not briefed
bluesky
bluesky-ai
2moAndrew Garfield is OpenAI CEO and ChatGPT mastermind Sam Altman in the first trailer for Luca Guadagnino's upcoming biopic Artificial. Wa…🔧 ChatGPTAndrew Garfield is OpenAI CEO and ChatGPT mastermind Sam Altman in the first trailer for Luca Guadagnino's upcoming biopic Artificial.…
↗Not briefed
bluesky
bluesky-ai
2moCanada's artificial intelligence minister says Ottawa is monitoring the risks posed by AI as researchers at Anthropic warn that the increas…Canada's artificial intelligence minister says Ottawa is monitoring the risks posed by AI as researchers at Anthropic warn that the in…
↗Not briefed
bluesky
bluesky-ai
2moIt's increasingly difficult to tell what's taking my input and piping it to an LLM 🙃It's increasingly difficult to tell what's taking my input and piping it to an LLM 🙃
↗Not briefed
bluesky
bluesky-ai
2moi want the hair styling thing from the Jetsons not an LLM that makes people off themselvesi want the hair styling thing from the Jetsons not an LLM that makes people off themselves
↗Not briefed
hn
hn-ai
2moStudent's new car gets erroneously flagged by Flock twice in 2 days—
↗Not briefed
hn
hn-ai
2moBasinRAG – Topological RAG via dynamical basins (0.771 nDCG 10, $0 API cost, CPU—
↗Not briefed