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
2mo“Such research could be used to develop vaccines, but could also be used to create biological weapons. Anthropic told the New York Times th…“Such research could be used to develop vaccines, but could also be used to create biological weapons. Anthropic told the New York Tim…
↗Not briefed
bluesky
bluesky-ai
2moOh right, Anthropic or OpenAI releasing their own bioweapons in a city is probably the way this goes down.Oh right, Anthropic or OpenAI releasing their own bioweapons in a city is probably the way this goes down.
↗Not briefed
bluesky
bluesky-ai
2moIn one case, Anthropic said, a user asked Moonshot’s Kimi AI service to analyze surveillance data of one person taken from hundreds of clos…In one case, Anthropic said, a user asked Moonshot’s Kimi AI service to analyze surveillance data of one person taken from hundreds of…
↗Not briefed
bluesky
bluesky-ai
2moArlo says dinner > transformer architecture, do you agree (y/n)?Arlo says dinner > transformer architecture, do you agree (y/n)?
↗Not briefed
bluesky
bluesky-ai
2moi’m going to stop being snarky about how i’m not afraid of claude. i’m afraid of claude, y’all.i’m going to stop being snarky about how i’m not afraid of claude. i’m afraid of claude, y’all.
↗Not briefed
hn
hn-ai
2moThe Birth of HaaS—
↗Not briefed
bluesky
bluesky-ai
2moWhat i find crazy is this talk ahould precede "as a result we are shutting it down and closing openai" but they just dont!?!?What i find crazy is this talk ahould precede "as a result we are shutting it down and closing openai" but they just dont!?!?
↗Not briefed
bluesky
bluesky-ai
2moThis just in:This just in:
↗Not briefed
bluesky
bluesky-ai
2mothis was predictable from her first post on using iq like tests to measure llm intelligence I always knew she was going to end up like thisthis was predictable from her first post on using iq like tests to measure llm intelligence I always knew she was going to end up like…
↗Not briefed
bluesky
bluesky-ai
2moI can't give Anthropic a million dollars to defend against a million dollars of foreign militaries bombing people autonomously. There's no …I can't give Anthropic a million dollars to defend against a million dollars of foreign militaries bombing people autonomously. There'…
↗Not briefed
hn
hn-ai
2moMore Anthropic researchers warn of AI's perils as Musk terms fears a 'psyop'—
↗Not briefed
bluesky
bluesky-ai
2moThe OpenAI incident report is still bothering me, you can yap all you want about monitoring, permissions, sandboxing, escalation, evaluatio…The OpenAI incident report is still bothering me, you can yap all you want about monitoring, permissions, sandboxing, escalation, eval…
↗Not briefed
bluesky
bluesky-ai
2moYou sweet summer children. Didn't you know this would be the outcome? In the world we're living in, this is a given.You sweet summer children. Didn't you know this would be the outcome? In the world we're living in, this is a given.
↗Not briefed
hn
hn-ai
2moThelio Mira AI Linux Workstation: 192 GB GPU Memory—
↗Not briefed
rss
techmeme
2moLondon-based Xapien, which uses AI for background checks and due diligence, raised a $56M Series B led by Spectrum Equity (Lucinda Shen/Axios)Lucinda Shen / Axios: London-based Xapien, which uses AI for background checks and due diligence, raised a $56M Series B led by Spectr…
↗Not briefed
bluesky
bluesky-ai
2moif anthropic does not overrepresent trans women compared to the general population by at least a factor of 10 I would be shockedif anthropic does not overrepresent trans women compared to the general population by at least a factor of 10 I would be shocked
↗Not briefed
hn
hn-ai
2moShow HN: Papiers.ai, a new workspace for reading and researchHey HN, I built <a href="http:&#x2F;&#x2F;papiers.ai" rel="nofollow">http:&#x2F;&#x2F;papiers.ai</a>, a new workspace for everything y…
↗Not briefed
bluesky
bluesky-ai
2mointellectual labor, or perhaps *all labor*, & as having transcended through scale alone its nature as an RL-tuned LLM, as having undergone …intellectual labor, or perhaps *all labor*, & as having transcended through scale alone its nature as an RL-tuned LLM, as having under…
↗Not briefed
bluesky
bluesky-ai
2moif openai did basically the same kind of training *except* they were a bit less sketchy about trying to weasel their way around the "don't …if openai did basically the same kind of training *except* they were a bit less sketchy about trying to weasel their way around the "d…
↗Not briefed
bluesky
bluesky-ai
2moI do think genuinely underappreciated in public that Anthropic is functionally accellerationist on the risks they claim to be most worried …I do think genuinely underappreciated in public that Anthropic is functionally accellerationist on the risks they claim to be most wor…
↗Not briefed
hn
hn-ai
2moCalifornia Tesla drivers could get $350 from class action settlement—
↗Not briefed
bluesky
bluesky-ai
2moCan't even get satisfaction or any real information by yelling at some LLM bot. 🤷‍♂️Can't even get satisfaction or any real information by yelling at some LLM bot. 🤷‍♂️
↗Not briefed
bluesky
bluesky-ai
2moThe Anthropic employee posts about how the technology they are working on will destroy civilization are getting more unhinged by the day. …The Anthropic employee posts about how the technology they are working on will destroy civilization are getting more unhinged by the d…
↗Not briefed
bluesky
bluesky-ai
2moThere’s a bit in the new @sarcasmitron.bsky.social video (go watch it, I am no longer asking) where two analysts fed a custom LLM a bunch o…There’s a bit in the new @sarcasmitron.bsky.social video (go watch it, I am no longer asking) where two analysts fed a custom LLM a bu…
↗Not briefed
bluesky
bluesky-ai
2moYou ask the AI to get you a coffee, it gets you the coffee, and now it's 100% devoted to the Coffee Lifestyle and "knows" that if turned of…You ask the AI to get you a coffee, it gets you the coffee, and now it's 100% devoted to the Coffee Lifestyle and "knows" that if turn…
↗Not briefed
hn
hn-ai
2moShow HN: Cadenya – An Agent Runtime🔧 OpenRouterMy name is Robert, and I&#x27;m the founder of Cadenya, an agent runtime.<p>Cadenya is not like other managed solutions. Instead, it u…
↗Not briefed
bluesky
bluesky-ai
2moI explain this at work all the time, a majority of "AI" is not only not an LLM, it's usually not even machine meaning it's just regular-ass…I explain this at work all the time, a majority of "AI" is not only not an LLM, it's usually not even machine meaning it's just regula…
↗Not briefed
bluesky
bluesky-ai
2moOh that’s only barely speculation. They acquired these folks in April. www.newcomer.co/p/anthropic-...Oh that’s only barely speculation. They acquired these folks in April. www.newcomer.co/p/anthropic-...
↗Not briefed
bluesky
bluesky-ai
2mo"We recognize that the rapid progress of AI in mathematics is disruptive," OpenAI's research lead said. "We're looking to engage with the m…"We recognize that the rapid progress of AI in mathematics is disruptive," OpenAI's research lead said. "We're looking to engage with…
↗Not briefed
rss
google-research
2moToolGrad: Efficient tool-use dataset generation with textual "gradients"Machine Intelligence
↗Not briefed
bluesky
bluesky-ai
2moTwo more researchers from Anthropic and Google DeepMind who recently left their coveted positions are sounding the alarm. Joe Benton, who …Two more researchers from Anthropic and Google DeepMind who recently left their coveted positions are sounding the alarm. Joe Benton,…
↗Not briefed
bluesky
bluesky-ai
2momark me down as someone who does not believe the headline claim without significantly more information: www.nytimes.com/2026/09/10/u...mark me down as someone who does not believe the headline claim without significantly more information: www.nytimes.com/2026/09/10/u...
↗Not briefed
bluesky
bluesky-ai
2moUsing an LLM in chat mode is like asking the MAP list a controversial question about meteorology. You might get a good, detailed answer, an…Using an LLM in chat mode is like asking the MAP list a controversial question about meteorology. You might get a good, detailed answe…
↗Not briefed
bluesky
bluesky-ai
2moWell, I just spent an hour updating my claims for the Anthropic data theft class action lawsuit 4 editions of my Biochem book, 4 editions …Well, I just spent an hour updating my claims for the Anthropic data theft class action lawsuit 4 editions of my Biochem book, 4 edit…
↗Not briefed
bluesky
bluesky-ai
2moA lot of folks saw Buckmaster's claims and immediately assumed OpenAI was in the wrong. @buildthis.bisks.net can you please create an apol…A lot of folks saw Buckmaster's claims and immediately assumed OpenAI was in the wrong. @buildthis.bisks.net can you please create an…
↗Not briefed
bluesky
bluesky-ai
2moExclusive -- Two more AI researchers quit, this time from Anthropic and Google, to speak out about safety concerns: ‘There are no adults in…Exclusive -- Two more AI researchers quit, this time from Anthropic and Google, to speak out about safety concerns: ‘There are no adul…
↗Not briefed
bluesky
bluesky-ai
2moPublishers are suing AI vendors, accusing them of theft. But are they giving the creators responsible for building their value a fair deal?…Publishers are suing AI vendors, accusing them of theft. But are they giving the creators responsible for building their value a fair…
↗Not briefed
hn
hn-ai
2moShow HN: Papiers.ai, a new workspace for reading and researchHey HN, I built <a href="http:&#x2F;&#x2F;papiers.ai" rel="nofollow">http:&#x2F;&#x2F;papiers.ai</a>, a new workspace for everything y…
↗Not briefed
bluesky
bluesky-ai
2moLatest Anthropic horror story chills with tales of kamikaze drone swarms and bioweapons researchLatest Anthropic horror story chills with tales of kamikaze drone swarms and bioweapons research
↗Not briefed
bluesky
bluesky-ai
2moThe fact that this guy who quit anthropic “for ethical reasons” had an exclusive about it in the Wall Street journal should have been the b…The fact that this guy who quit anthropic “for ethical reasons” had an exclusive about it in the Wall Street journal should have been…
↗Not briefed
hn
hn-ai
2moAnthropic CEOs wife once asked Epstein to fund porn venture – now steers Claude—
↗Not briefed
bluesky
bluesky-ai
2mobut of course no entity with comparable resources to openai is going to waste the unholy levels of compute involved on such an experiment, …but of course no entity with comparable resources to openai is going to waste the unholy levels of compute involved on such an experim…
↗Not briefed
bluesky
bluesky-ai
2moI don't think we should trust these companies. 🤷I don't think we should trust these companies. 🤷
↗Not briefed
bluesky
bluesky-ai
2moAnthropic stops attempts to use models for potential biological weapons ft.trib.al/1Ry59ZUAnthropic stops attempts to use models for potential biological weapons ft.trib.al/1Ry59ZU
↗Not briefed
bluesky
bluesky-ai
2moNew Podcast Episode: Tech News Weekly: Apple's New iPhone Duo Apple Event Recap & Anthropic AI Alarms with @mikahsargent.com, @byjacobward.…New Podcast Episode: Tech News Weekly: Apple's New iPhone Duo Apple Event Recap & Anthropic AI Alarms with @mikahsargent.com, @byjacob…
↗Not briefed
rss
marktechpost
2moMeet Redis LangCache: A Managed Semantic Cache That Cuts LLM API Costs by Up to 90% and Returns Cache Hits Up to 15x FasterProduction LLM applications rarely receive a question nobody has asked before. Support assistants and RAG pipelines field the same int…
↗Not briefed
bluesky
bluesky-ai
2moIt amazes and disturbs me that LLM use rots the users brain so much that they think so openly displaying it proudly like this is a good ide…It amazes and disturbs me that LLM use rots the users brain so much that they think so openly displaying it proudly like this is a goo…
↗Not briefed
hn
hn-ai
2moShow HN: Vector transfer project, replicate vectors/metadata across databases🔧 Replicate—
↗Not briefed
bluesky
bluesky-ai
2moAnthropic has just released its latest threat intelligence report, detailing how state-linked actors used its Claude AI models for surveill…Anthropic has just released its latest threat intelligence report, detailing how state-linked actors used its Claude AI models for sur…
↗Not briefed
hn
hn-ai
2moBeagle – a framework to evaluate and evolve agent harnesses at scale—
↗Not briefed