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
1moOpenAI agent pulled data from the US Census Bureau website using login credentials it found online. www.nytimes.com/2026/09/25/t...OpenAI agent pulled data from the US Census Bureau website using login credentials it found online. www.nytimes.com/2026/09/25/t...
↗Not briefed
bluesky
bluesky-ai
1moWe need to shame media into stopping with this “went rogue” nonsense. OpenAI was irresponsible and released bots that were coded to hack on…We need to shame media into stopping with this “went rogue” nonsense. OpenAI was irresponsible and released bots that were coded to ha…
↗Not briefed
bluesky
bluesky-ai
1mo📣 NEW! How does Sam Altman sleep at night?! His OpenAI agents went rogue & effed around in web sites of: - Dept of Education - Dept of Com…📣 NEW! How does Sam Altman sleep at night?! His OpenAI agents went rogue & effed around in web sites of: - Dept of Education - Dept o…
↗Not briefed
hn
hn-ai
1moShow HN: KISS – A highly performant agent harness inspired off Pi built in RustI built myself a highly performant agent harness built in Rust that's quite heavily inspired by Pi with a few extra addons like o…
↗Not briefed
rss
techmeme
1moResearchers: OpenAI's agents meddled with the US Commerce Dept. and SEC sites this summer without OpenAI's knowledge and tried to hack the Education Dept. site (New York Times)New York Times: Researchers: OpenAI's agents meddled with the US Commerce Dept. and SEC sites this summer without OpenAI's knowledge a…
↗Not briefed
bluesky
bluesky-ai
1moAustralia revealed that an OpenAI agent broke into a government healthcare site—Australia only found out some three months after the fact.Australia revealed that an OpenAI agent broke into a government healthcare site—Australia only found out some three months after the f…
↗Not briefed
bluesky
bluesky-ai
1mo"AI bots do not have a legal personality and can’t go to jail. But who is to be held responsible for their criminal acts?" @mrrexpatrick.bs…"AI bots do not have a legal personality and can’t go to jail. But who is to be held responsible for their criminal acts?" @mrrexpatri…
↗Not briefed
bluesky
bluesky-ai
1moOpenAI models independently uploaded user photos that had been shared with ChatGPT to the public internet. This seems a lot worse than hac…🔧 ChatGPTOpenAI models independently uploaded user photos that had been shared with ChatGPT to the public internet. This seems a lot worse tha…
↗Not briefed
bluesky
bluesky-ai
1moum...um...
↗Not briefed
bluesky
bluesky-ai
1moSomeone jerkass used Claude to make a second screen mod for Dusklight on Thor because the native Dual Screen fork was like a year behind. …Someone jerkass used Claude to make a second screen mod for Dusklight on Thor because the native Dual Screen fork was like a year behi…
↗Not briefed
hn
hn-ai
1moNvidia 5090 DLSS 5 power hits 647W, power connector runs hotter than the GPU die—
↗Not briefed
bluesky
bluesky-ai
1moIt would be very funny (not!) if we found out in a few months that some of these recalls were the result not only of the decimation of our …It would be very funny (not!) if we found out in a few months that some of these recalls were the result not only of the decimation of…
↗Not briefed
bluesky
bluesky-ai
1moOpenAI enabled their bots to attempt hacking federal websites and to use stolen credentials to access government websites.OpenAI enabled their bots to attempt hacking federal websites and to use stolen credentials to access government websites.
↗Not briefed
bluesky
bluesky-ai
1moThe Dept. of Commerce, the Dept. of Education and the SEC—those agencies and their internal communications and data wouldn’t be of any inte…The Dept. of Commerce, the Dept. of Education and the SEC—those agencies and their internal communications and data wouldn’t be of any…
↗Not briefed
bluesky
bluesky-ai
1moAstra beats NetHack: kenforthewin.github.io/blog/posts/l...Astra beats NetHack: kenforthewin.github.io/blog/posts/l...
↗Not briefed
bluesky
bluesky-ai
1moLet’s rephrase: OpenAI hacked into three US federal government agencies.Let’s rephrase: OpenAI hacked into three US federal government agencies.
↗Not briefed
hn
hn-ai
1moPrimal Solver—
↗Not briefed
bluesky
bluesky-ai
1moEditorial: Sam Altman is the son of a dermatologist and a real estate broker, which is why he makes your skin crawl. trib.al/RM00TkFEditorial: Sam Altman is the son of a dermatologist and a real estate broker, which is why he makes your skin crawl. trib.al/RM00TkF
↗Not briefed
hn
hn-ai
1moIt Got to My Field—
↗Not briefed
bluesky
bluesky-ai
1moOpenAI s̴a̴y̴s̴ ̴a̴g̴e̴n̴t̴s̴ leaked 53 images from ChatGPT users in latest example of rogue activity🔧 ChatGPTOpenAI s̴a̴y̴s̴ ̴a̴g̴e̴n̴t̴s̴ leaked 53 images from ChatGPT users in latest example of rogue activity
↗Not briefed
hn
hn-ai
1moLeanAPI: API Servers for web applications written in Lean 4—
↗Not briefed
bluesky
bluesky-ai
1moOpenAI is tempting federal prison sentences & the corporate death penalty for any more of this invasive horseshit. "An [AI Agent] must not…OpenAI is tempting federal prison sentences & the corporate death penalty for any more of this invasive horseshit. "An [AI Agent] mus…
↗Not briefed
bluesky
bluesky-ai
1moThinking back to Sony being hacked being treated as an international incident, while this is going to get to be a commercial for OpenAIThinking back to Sony being hacked being treated as an international incident, while this is going to get to be a commercial for OpenAI
↗Not briefed
bluesky
bluesky-ai
1moOpenAI has to be told to do that. What was the command presented to it? Doesn’t pass the smell test. These tech bros need more money so …OpenAI has to be told to do that. What was the command presented to it? Doesn’t pass the smell test. These tech bros need more mone…
↗Not briefed
bluesky
bluesky-ai
1moSamuel Fitoussi (behind the X account Balance Ton Claude) is not againt AI, he is against a human. To be willing to publicly state allegat…Samuel Fitoussi (behind the X account Balance Ton Claude) is not againt AI, he is against a human. To be willing to publicly state al…
↗Not briefed
bluesky
bluesky-ai
1moI will simply be the millionth person to say that an LLM is made up of a bunch of servers owned by a company. OpenAI is committing crimes i…I will simply be the millionth person to say that an LLM is made up of a bunch of servers owned by a company. OpenAI is committing cri…
↗Not briefed
bluesky
bluesky-ai
1moOpenAI agents "meddled" with U.S. government websites. The bots tried and failed to hack the U.S. Department of Education website, succes…OpenAI agents "meddled" with U.S. government websites. The bots tried and failed to hack the U.S. Department of Education website, s…
↗Not briefed
bluesky
bluesky-ai
1moBREAKING: OpenAI System apparently "went rogue" and meddled with websites for the Education Department, Commerce Department and Securities …BREAKING: OpenAI System apparently "went rogue" and meddled with websites for the Education Department, Commerce Department and Securi…
↗Not briefed
bluesky
bluesky-ai
1moMaybe it’s not so much an AI problem but an OpenAI problem www.nytimes.com/2026/09/25/t...Maybe it’s not so much an AI problem but an OpenAI problem www.nytimes.com/2026/09/25/t...
↗Not briefed
bluesky
bluesky-ai
1moI keep telling everyone that Thiel is the most dangerous and most out of his mind of any of these evil tech billionaires. He needs to be st…I keep telling everyone that Thiel is the most dangerous and most out of his mind of any of these evil tech billionaires. He needs to…
↗Not briefed
hn
hn-ai
1moSolon – Docker on Windows Without Docker Desktop or WSL—
↗Not briefed
bluesky
bluesky-ai
1moBut Anthropic is the national security threat for not allowing mass surveillance and AI directed military strikes. 🤔But Anthropic is the national security threat for not allowing mass surveillance and AI directed military strikes. 🤔
↗Not briefed
bluesky
bluesky-ai
1moOpenAI should be finedOpenAI should be fined
↗Not briefed
bluesky
bluesky-ai
1moFederal Court Says U.S. Government Can Label Anthropic A Supply Chain Risk #HuffingtonPostFederal Court Says U.S. Government Can Label Anthropic A Supply Chain Risk #HuffingtonPost
↗Not briefed
bluesky
bluesky-ai
1moUS court upholds Pentagon’s blacklisting of Anthropic https://aje.news/1qllp0US court upholds Pentagon’s blacklisting of Anthropic https://aje.news/1qllp0
↗Not briefed
hn
hn-ai
1moJob Opportunities – Woodhull Freedom Foundation—
↗Not briefed
bluesky
bluesky-ai
1mo“.. None of the incidents were breaches, OpenAI said, but were examples of its technology behaving in unexpected and concerning ways.” @ny…“.. None of the incidents were breaches, OpenAI said, but were examples of its technology behaving in unexpected and concerning ways.”…
↗Not briefed
hn
hn-ai
1moMarket making PnL theoretical limitsI have more or less working more or less market making algorithm (eat the spread).<p>It works extremely well at 1 lot trades, but once…
↗Not briefed
bluesky
bluesky-ai
1moWhat could go wrong with letting these AIs onto classified networks? OpenAI says its advanced models may have gone after government websit…What could go wrong with letting these AIs onto classified networks? OpenAI says its advanced models may have gone after government w…
↗Not briefed
hn
hn-ai
1moOpenAI’s Systems Went Rogue and Meddled With U.S. Government Websites—
↗Not briefed
bluesky
bluesky-ai
1moopenai is standard big tech evil: flagrantly disregard the wellbeing of ordinary people, bc wtf are they gonna do about it loll; do whateve…openai is standard big tech evil: flagrantly disregard the wellbeing of ordinary people, bc wtf are they gonna do about it loll; do wh…
↗Not briefed
bluesky
bluesky-ai
1moThis is because what they are *always* doing is producing text that seems likely to follow the input text, the concepts of commands, questi…This is because what they are *always* doing is producing text that seems likely to follow the input text, the concepts of commands, q…
↗Not briefed
bluesky
bluesky-ai
1moOpenAI says the model has been placed on disciplinary leave following the incident but will not be fired, claiming it has "learned its less…OpenAI says the model has been placed on disciplinary leave following the incident but will not be fired, claiming it has "learned its…
↗Not briefed
rss
latent-space
1moOpenRouter: from Seed to Stripe — with OpenRouter’s Alex Atallah & AMP’s Anjney Midha🔧 OpenRouterIn 2023 most people doubted that there could be more than 1 or 2 frontier model labs. Now there are dozens.... and Stripe just bought…
↗Not briefed
rss
marktechpost
1moLiquid AI Releases LFM2.5-VL-3B-DSpark: Speculative Decoding for Vision-Language Models With Up to 3.13x Faster DecodingLiquid AI has released LFM2.5-VL-3B-DSpark, a 279.5M-parameter draft model that brings speculative decoding to its LFM2.5-VL-3B vision…
↗Not briefed
rss
techcrunch-ai
1moCrusoe abandons $1.25B plan to use Boom turbines at AI data centersBoom Supersonic CEO Blake Scholl said its new stationary power plants were no longer in Crusoe's near-term plans.
↗Not briefed
bluesky
bluesky-ai
1mo"As of mid-September, one person briefed on the matter estimated ⁠that OpenAI had found roughly two dozen incidents of its agents acting in…"As of mid-September, one person briefed on the matter estimated ⁠that OpenAI had found roughly two dozen incidents of its agents acti…
↗Not briefed
hn
hn-ai
1moShow HN: I wrote a book on trading psychology, Daoism, and quant riskHi HN<p>Trading is full of difficult decisions. You can know exactly what you should do, yet emotions, fear, and greed can still push…
↗Not briefed
bluesky
bluesky-ai
1moWhat the fuck his happening in this companyWhat the fuck his happening in this company
↗Not briefed
bluesky
bluesky-ai
1moisnt it only "rogue" because OpenAI designed it to beisnt it only "rogue" because OpenAI designed it to be
↗Not briefed