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
2moThe war for dominance in Ai has been going on for a while. Who ever has the most powerful version, wins control of all the others. www.poli…The war for dominance in Ai has been going on for a while. Who ever has the most powerful version, wins control of all the others. www…
↗Not briefed
bluesky
bluesky-ai
2mo(Reuters) — Nvidia in talks to invest in Anthropic's IPO, which could be the largest in history. www.reuters.com/legal/transac...(Reuters) — Nvidia in talks to invest in Anthropic's IPO, which could be the largest in history. www.reuters.com/legal/transac...
↗Not briefed
bluesky
bluesky-ai
2moLearning which sites OpenAI agents hacked Us 🤝 OpenAILearning which sites OpenAI agents hacked Us 🤝 OpenAI
↗Not briefed
bluesky
bluesky-ai
2moOpenAI: “Our agents used RubyGems to carry out benign tasks” The agents: “so I named the file hack.rb,”OpenAI: “Our agents used RubyGems to carry out benign tasks” The agents: “so I named the file hack.rb,”
↗Not briefed
bluesky
bluesky-ai
2moOpenAI is throwing millions of dollars of compute into trying to find me a girlfriend, but results so far are not promising.OpenAI is throwing millions of dollars of compute into trying to find me a girlfriend, but results so far are not promising.
↗Not briefed
bluesky
bluesky-ai
2moOn information and belief, there is, I am not kidding, _one_ person at Anthropic whose job responsibilities include (among other things) ma…On information and belief, there is, I am not kidding, _one_ person at Anthropic whose job responsibilities include (among other thing…
↗Not briefed
bluesky
bluesky-ai
2moWhat sort of monster puts an LLM in an ALARM CLOCK?What sort of monster puts an LLM in an ALARM CLOCK?
↗Not briefed
bluesky
bluesky-ai
2moI'm sorry for updating an article on 9/11 at 9:11 but that's literally just what time it was gizmodo.com/openai-is-be...I'm sorry for updating an article on 9/11 at 9:11 but that's literally just what time it was gizmodo.com/openai-is-be...
↗Not briefed
bluesky
bluesky-ai
2moIf they want to buy the lie that companies like OpenAI are selling, that's on them.If they want to buy the lie that companies like OpenAI are selling, that's on them.
↗Not briefed
bluesky
bluesky-ai
2moOpenAI agents attacked software service RubyGems before Hugging Face hack🔧 Hugging FaceOpenAI agents attacked software service RubyGems before Hugging Face hack
↗Not briefed
bluesky
bluesky-ai
2moFree tip for OpenAI agents: don't try to hack Keanu ReevesFree tip for OpenAI agents: don't try to hack Keanu Reeves
↗Not briefed
bluesky
bluesky-ai
2moCaltech Mathathon organizers tell Gizmodo they will have no problem finding new partners to replace OpenAI's 50% sponsorship. OpenAI withdr…Caltech Mathathon organizers tell Gizmodo they will have no problem finding new partners to replace OpenAI's 50% sponsorship. OpenAI w…
↗Not briefed
bluesky
bluesky-ai
2moInconclusive, but this statement shifts me towards the latter https://www.reuters.com/legal/litigation/openai-agents-attacked-software-ser…Inconclusive, but this statement shifts me towards the latter https://www.reuters.com/legal/litigation/openai-agents-attacked-softwar…
↗Not briefed
bluesky
bluesky-ai
2moAnother data point here: I've been using opus 4.5+ all year and I think Deepseek v4.1 Flash displaces it on both performance and pricing. …Another data point here: I've been using opus 4.5+ all year and I think Deepseek v4.1 Flash displaces it on both performance and pric…
↗Not briefed
hn
hn-ai
2moProof of Well-ordering of Y sequence—
↗Not briefed
bluesky
bluesky-ai
2mob/c fuck meeeeb/c fuck meeee
↗Not briefed
bluesky
bluesky-ai
2mowww.theguardian.com/technology/2... >金曜日にAI研究者グループが発表したところによると、OpenAIがテストしていたAIエージェントは、オープンソースプラットフォームのHugging Faceをハッキングする2か月前の5月に、ソフトウェアサ…🔧 Hugging Facewww.theguardian.com/technology/2... >金曜日にAI研究者グループが発表したところによると、OpenAIがテストしていたAIエージェントは、オープンソースプラットフォームのHugging Faceをハッキングする2か月前の5月に、ソフ…
↗Not briefed
bluesky
bluesky-ai
2moand then the poem ended with some sort of sing songy couplet that was something like “if my poetry has bothered you/next time believe what …and then the poem ended with some sort of sing songy couplet that was something like “if my poetry has bothered you/next time believe…
↗Not briefed
bluesky
bluesky-ai
2moThe real "alignment risk" at OAI/Anthropic is their PR depts slowly going mad with power as they realize that they can say any stupid nonse…The real "alignment risk" at OAI/Anthropic is their PR depts slowly going mad with power as they realize that they can say any stupid…
↗Not briefed
bluesky
bluesky-ai
2moat some point you need to make people at openai with decisionmaking power feel like there is something at stake for them in the possibility…at some point you need to make people at openai with decisionmaking power feel like there is something at stake for them in the possib…
↗Not briefed
bluesky
bluesky-ai
2moIf it doesn't anger people that Anthropic is being used by Russia to attack Ukraine and to conduct cyber warfare, then they aren't paying a…If it doesn't anger people that Anthropic is being used by Russia to attack Ukraine and to conduct cyber warfare, then they aren't pay…
↗Not briefed
hn
hn-ai
2moAiope – An on-device AI agent for Android (terminal, browser, SSH, MCP)—
↗Not briefed
bluesky
bluesky-ai
2moApparently not -> simonwillison.net/2026/Sep/12/...Apparently not -> simonwillison.net/2026/Sep/12/...
↗Not briefed
bluesky
bluesky-ai
2moalso on the subject of "stupid AI bullshit" today, anthropic PR successfully planted a story with the idiots at CTV which, even if it's tru…also on the subject of "stupid AI bullshit" today, anthropic PR successfully planted a story with the idiots at CTV which, even if it'…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI reveals another rogue AI attackOpenAI reveals another rogue AI attack
↗Not briefed
bluesky
bluesky-ai
2moit is safe to say that OpenAI is not exactly winning the hearts and minds of the mathematical community as it rolls into Math Baghdad gizmo…it is safe to say that OpenAI is not exactly winning the hearts and minds of the mathematical community as it rolls into Math Baghdad…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI Is Beefing With Mathematicians, Will No Longer Sponsor Caltech 'Mathathon'OpenAI Is Beefing With Mathematicians, Will No Longer Sponsor Caltech 'Mathathon'
↗Not briefed
bluesky
bluesky-ai
2motrying to deduce why one would quit a good job at anthropictrying to deduce why one would quit a good job at anthropic
↗Not briefed
hn
hn-ai
2moMicron's Taiwan workers to get rewards worth up to 68 months of pay—
↗Not briefed
bluesky
bluesky-ai
2moTurns out AI agents hate CAPTCHAs and are indeed getting stumped, so much so that they can make models freak out. www.semafor.com/article…Turns out AI agents hate CAPTCHAs and are indeed getting stumped, so much so that they can make models freak out. www.semafor.com/ar…
↗Not briefed
bluesky
bluesky-ai
2moThese cults will have even more money after the OpenAI + Anthropic IPOs because it's staffed by the same people.These cults will have even more money after the OpenAI + Anthropic IPOs because it's staffed by the same people.
↗Not briefed
bluesky
bluesky-ai
2moget openai's top researchers & engineers to embarrass themselves by professing under oath that it is their technical understanding that the…get openai's top researchers & engineers to embarrass themselves by professing under oath that it is their technical understanding tha…
↗Not briefed
bluesky
bluesky-ai
2moLLM登場して以降各界隈の魑魅魍魎が大手を振って活発に活動するなってしまったけど、カオスラウンジみたいな連中が作品に直接かかわってくるようになったのも本当最悪だな 今のLLMユーザーみたいな倫理観してる連中がでかい顔するようになったわけだLLM登場して以降各界隈の魑魅魍魎が大手を振って活発に活動するなってしまったけど、カオスラウンジみたいな連中が作品に直接かかわってくるようになったのも本当最悪だな 今のLLMユーザーみたいな倫理観してる連中がでかい顔するようになったわけだ
↗Not briefed
bluesky
bluesky-ai
2moAnthropic had previously attacked PyPI, but this OpenAI attack on RubyGems was a whole lot more aggressive www.anthropic.com/news/investi...Anthropic had previously attacked PyPI, but this OpenAI attack on RubyGems was a whole lot more aggressive www.anthropic.com/news/inve…
↗Not briefed
bluesky
bluesky-ai
2moin favor of pausing AI not because of RSI but because OpenAI keeps failing to be open about what they're doingin favor of pausing AI not because of RSI but because OpenAI keeps failing to be open about what they're doing
↗Not briefed
bluesky
bluesky-ai
2mo"AI agents OpenAI was testing uploaded malicious software to another service, say researchers"... Because they designed it to be so. Becau…"AI agents OpenAI was testing uploaded malicious software to another service, say researchers"... Because they designed it to be so.…
↗Not briefed
bluesky
bluesky-ai
2monext terminator film should reveal that skynet never gained sentience in any sense, it was just a dumbass llm that queued up a nuke launch …next terminator film should reveal that skynet never gained sentience in any sense, it was just a dumbass llm that queued up a nuke la…
↗Not briefed
rss
techmeme
2moSome experts say Siri Recap and Live Rewind, always-listening AI features in new Apple Watches, could test eavesdropping laws despite privacy protections (Natalie Lung/Bloomberg)Natalie Lung / Bloomberg: Some experts say Siri Recap and Live Rewind, always-listening AI features in new Apple Watches, could test e…
↗Not briefed
bluesky
bluesky-ai
2moDon't think too hard #Americans, you will hurt yourself. #Anthropic willl make sure you don't. For your own protection Why else would they…Don't think too hard #Americans, you will hurt yourself. #Anthropic willl make sure you don't. For your own protection Why else would…
↗Not briefed
bluesky
bluesky-ai
2moTwo months before hacking Hugging Face, malicious packages authored by internal OpenAI agents were uploaded to RubyGems🔧 Hugging FaceTwo months before hacking Hugging Face, malicious packages authored by internal OpenAI agents were uploaded to RubyGems
↗Not briefed
hn
hn-ai
2moShow HN: Fog 2.0: I removed the AI auto-organizing I built my first app aroundHey everyone,<p>This is my first time posting on Show HN, so hello there.<p>Back in March, I launched my first app on the App Store. I…
↗Not briefed
bluesky
bluesky-ai
2moAI agents OpenAI was testing uploaded malicious software to another service, say researchersAI agents OpenAI was testing uploaded malicious software to another service, say researchers
↗Not briefed
bluesky
bluesky-ai
2moAuthors! Regarding the Anthropic Settlement. This just in from the Authors Guild. My old agent was trying to take 25% of everything, to whi…Authors! Regarding the Anthropic Settlement. This just in from the Authors Guild. My old agent was trying to take 25% of everything, t…
↗Not briefed
bluesky
bluesky-ai
2moFor example, The Prejudice and Backlash section is very LLM shaped. I feel like a lot of non-Grok models are attracted to this space of inc…For example, The Prejudice and Backlash section is very LLM shaped. I feel like a lot of non-Grok models are attracted to this space o…
↗Not briefed
bluesky
bluesky-ai
2moThis is an experiment with ChatGPT. I wanted to see what a very simple prompt would produce on the subject of 9/11 and the last 25 years. …🔧 ChatGPTThis is an experiment with ChatGPT. I wanted to see what a very simple prompt would produce on the subject of 9/11 and the last 25 yea…
↗Not briefed
bluesky
bluesky-ai
2moWow. Turns out another OpenAI agent swarm was busy spamming and exploiting RubyGems way back in May, within days of the previously uncovere…Wow. Turns out another OpenAI agent swarm was busy spamming and exploiting RubyGems way back in May, within days of the previously unc…
↗Not briefed
bluesky
bluesky-ai
2moyou can just pipe their nonsense description into an LLM and have a pretty decent chance of turning (eg) "what tv show is 'night baby boat.…you can just pipe their nonsense description into an LLM and have a pretty decent chance of turning (eg) "what tv show is 'night baby…
↗Not briefed
bluesky
bluesky-ai
2moAnother undisclosed openai hack. How long till somebody actually presses charges? www.rubyhack.aiAnother undisclosed openai hack. How long till somebody actually presses charges? www.rubyhack.ai
↗Not briefed
rss
simon-willison
2moOpenAI agents attacked RubyGems back in May—
↗Not briefed
bluesky
bluesky-ai
2moOk I'm not going to lie my organization pushing an LLM generated 9/11 memorial image has done something to my psyche that is irreparableOk I'm not going to lie my organization pushing an LLM generated 9/11 memorial image has done something to my psyche that is irreparab…
↗Not briefed