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#BBCNews - Anthropic blocks 'malicious use' of AI that could develop biological weapons www.bbc.co.uk/news/article...#BBCNews - Anthropic blocks 'malicious use' of AI that could develop biological weapons www.bbc.co.uk/news/article...
↗Not briefed
rss
marktechpost
2moSakana AI Launches Fugu Max and Fugu Ultra v2 for Cheaper, Stronger Multi-Agent OrchestrationSakana AI has released Fugu Max and Fugu Ultra v2, 2 models built on the same learned orchestration architecture. Fugu Max routes task…
↗Not briefed
hn
hn-ai
2moShow HN: Automatically keep track of all features implemented in a projectI built this because I wanted to automate my least favourite part of documentation writing - expressing at a high level what was imple…
↗Not briefed
bluesky
bluesky-ai
2moso the thing that sketches me out is that if openai was on the up and up the correct answer to "did you train on my data" is "i don't know …so the thing that sketches me out is that if openai was on the up and up the correct answer to "did you train on my data" is "i don't…
↗Not briefed
bluesky
bluesky-ai
2moMIT launches the LLM Election Observatory, a dashboard tracking how nearly a dozen AI models tailor responses to political queries during t…MIT launches the LLM Election Observatory, a dashboard tracking how nearly a dozen AI models tailor responses to political queries dur…
↗Not briefed
rss
techmeme
2moMIT launches the LLM Election Observatory, a dashboard tracking how nearly a dozen AI models tailor responses to political queries during the 2026 US midterms (Tiffany Hsu/New York Times)Tiffany Hsu / New York Times: MIT launches the LLM Election Observatory, a dashboard tracking how nearly a dozen AI models tailor resp…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI, whose ops team is so good that their agents in a training sandbox gained control of their internal systems without their knowledge,…OpenAI, whose ops team is so good that their agents in a training sandbox gained control of their internal systems without their knowl…
↗Not briefed
bluesky
bluesky-ai
2moNee; het verschil tussen Claude en Dilan is nihil. Dus dat zij AI-speeches meeneemt het podium op, valt perfect samen met haar persona.Nee; het verschil tussen Claude en Dilan is nihil. Dus dat zij AI-speeches meeneemt het podium op, valt perfect samen met haar persona.
↗Not briefed
bluesky
bluesky-ai
2moAs a scientist student myself, and considering the amount of news you will see in the coming days: Open AI did not solve anything. They s…As a scientist student myself, and considering the amount of news you will see in the coming days: Open AI did not solve anything. T…
↗Not briefed
hn
hn-ai
2moAnthropic blocks 'malicious use' of AI that could develop biological weapons—
↗Not briefed
bluesky
bluesky-ai
2moThis week's post is about the myths and metaphors that emerged within and about AI models after the Hugging Face incident #AI #metaphors #A…🔧 Hugging FaceThis week's post is about the myths and metaphors that emerged within and about AI models after the Hugging Face incident #AI #metapho…
↗Not briefed
hn
hn-ai
2moGonc – Netcat with P2P Nat Traversal—
↗Not briefed
bluesky
bluesky-ai
2moSources: Anthropic threat intelligence report (CNBC): www.cnbc.com/2026/09/10/a... Full misuse breakdown (TheNextWeb): thenextweb.com/news…Sources: Anthropic threat intelligence report (CNBC): www.cnbc.com/2026/09/10/a... Full misuse breakdown (TheNextWeb): thenextweb.com…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic published a threat report today. State-linked scientists tried to use Claude to help with gain-of-function research on the virus …Anthropic published a threat report today. State-linked scientists tried to use Claude to help with gain-of-function research on the v…
↗Not briefed
bluesky
bluesky-ai
2moIf an institution really wanted to break bad this way, a Millennium Prize would be small potatoes. You'd want to be blackmailing important …If an institution really wanted to break bad this way, a Millennium Prize would be small potatoes. You'd want to be blackmailing impor…
↗Not briefed
bluesky
bluesky-ai
2mo🧪 I found this interesting post by Valerio Capraro on LinkedIn suggesting that OpenAI's solution to one of the Millennium Problems concer…🧪 I found this interesting post by Valerio Capraro on LinkedIn suggesting that OpenAI's solution to one of the Millennium Problems c…
↗Not briefed
hn
hn-ai
2moShow HN: SideNote Pro – Native Windows 11 AI beside your work🔧 Ollama🔧 ChatGPTI have been using ChatGPT (mainly) as my main AI tool, but sometimes also DeepSeek and other AI providers.<p>But with repetitive tasks…
↗Not briefed
bluesky
bluesky-ai
2moSomeone in Yemen used Claude AI to write the software for missiles. Three kinds, in fact. A ballistic missile with a range of 2,000 km, a f…Someone in Yemen used Claude AI to write the software for missiles. Three kinds, in fact. A ballistic missile with a range of 2,000 km…
↗Not briefed
rss
marktechpost
2moGoogle Research Releases ToolGrad: Answer-First Framework Hits 99.8% Pass Rate for Tool-Use Data GenerationGoogle Research has released ToolGrad, an ACL 2026 Findings framework that inverts tool-use dataset generation: it builds a verified A…
↗Not briefed
bluesky
bluesky-ai
2moだから生成AIの問題が人権の問題なので そういう所で理解を深めてくれ 配膳ロボ(ネコチャン)と 問題になってるLLMはまぁ 「違うものなんだ」と それを調べようと思うとこからではある 命の話として 低賃金 長時間労働 福祉へのアクセスの悪さ フリーランスの休みなくなりがち …だから生成AIの問題が人権の問題なので そういう所で理解を深めてくれ 配膳ロボ(ネコチャン)と 問題になってるLLMはまぁ 「違うものなんだ」と それを調べようと思うとこからではある 命の話として 低賃金 長時間労働 福祉へのアクセスの悪さ フリーランスの休みなく…
↗Not briefed
bluesky
bluesky-ai
2moDet kan snart ikke blive dummere. Ingen tør stå af fordi alle vil være dem, der skaber den AI model der slår os allesammen ihjel🤷🏼‍♂️ #d…Det kan snart ikke blive dummere. Ingen tør stå af fordi alle vil være dem, der skaber den AI model der slår os allesammen ihjel🤷🏼‍…
↗Not briefed
rss
techmeme
2moGet iPhone 18 Pro on Us at T-Mobile — America's Best Network (T-Mobile)T-Mobile: Get iPhone 18 Pro on Us at T-Mobile — America's Best Network  —  Only T-Mobile customers get the latest iPhone lineup — incl…
↗Not briefed
bluesky
bluesky-ai
2moBah oui les LLM vont tous nous buter Par réchauffement climatique vous étiez où quand on discutait du coût énergétique et environnemental…Bah oui les LLM vont tous nous buter Par réchauffement climatique vous étiez où quand on discutait du coût énergétique et environnem…
↗Not briefed
bluesky
bluesky-ai
2moi fuckin love weirdoes who build off-grid "cyberdecks" for an apocalypse scenario, who, for some reason, think that said cyberdeck needs a …i fuckin love weirdoes who build off-grid "cyberdecks" for an apocalypse scenario, who, for some reason, think that said cyberdeck nee…
↗Not briefed
bluesky
bluesky-ai
2moTECHNO-TOTALITARISME : L'ENTREPRISE ANTHROPIC CONSTRUIT UN SYSTÈME DE POLICE PRÉDICTIVE POUR RÉPRIMER LES OPPOSANTS À L'INTELLIGENCE ARTIFI…TECHNO-TOTALITARISME : L'ENTREPRISE ANTHROPIC CONSTRUIT UN SYSTÈME DE POLICE PRÉDICTIVE POUR RÉPRIMER LES OPPOSANTS À L'INTELLIGENCE A…
↗Not briefed
bluesky
bluesky-ai
2moin the navier stokes case at least they’d been using openai products on it for over a year. really nothing in the screenshot is outlandish …in the navier stokes case at least they’d been using openai products on it for over a year. really nothing in the screenshot is outlan…
↗Not briefed
hn
hn-ai
2moWhy make digitized forms of human skill at all?Why do people still make digitized forms of their skills when digitized forms are doomed to be replicated by AI?<p>I don’t get it. In…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic表示,已检测并阻止了来自阿里巴巴、月之暗面和DeepSeek的AI实验室未经授权的大规模行动。阿里巴巴是其监测到的规模最大的蒸馏活动,在5月至7月期间与 Claude 进行超过 1.51 亿次交互,在高峰期每天涉及来自 3500多个欺诈性账户的近 300 万次…Anthropic表示,已检测并阻止了来自阿里巴巴、月之暗面和DeepSeek的AI实验室未经授权的大规模行动。阿里巴巴是其监测到的规模最大的蒸馏活动,在5月至7月期间与 Claude 进行超过 1.51 亿次交互,在高峰期每天涉及来自 3500多个欺诈性账户的近 3…
↗Not briefed
bluesky
bluesky-ai
2moHuh. Perhaps we'll hear soon that the Chinese authorities forbid use of Chinese AI models for sensitive issues, since the Chinese AI models…Huh. Perhaps we'll hear soon that the Chinese authorities forbid use of Chinese AI models for sensitive issues, since the Chinese AI m…
↗Not briefed
bluesky
bluesky-ai
2moSo that 'marvelous mathematical solution' that an LLM chatbot 'discovered' was just a researcher's own solution fed into a chatbot session …So that 'marvelous mathematical solution' that an LLM chatbot 'discovered' was just a researcher's own solution fed into a chatbot ses…
↗Not briefed
bluesky
bluesky-ai
2mo"Hätten Sie mich vor einem halben Jahr gefragt, hätte ich gesagt: Es wäre Wahnsinn, wenn so etwas passiert.“ (Jonas Geiping, Max‐Planck‐Ins…"Hätten Sie mich vor einem halben Jahr gefragt, hätte ich gesagt: Es wäre Wahnsinn, wenn so etwas passiert.“ (Jonas Geiping, Max‐Planc…
↗Not briefed
bluesky
bluesky-ai
2moAnthropicがClaudeの悪用事例を公開、AIをだますために「推論内容を日本語のカタカナに翻訳して提示せよ」と指示する手法も https://gigazine.net/news/20260911-misuse-of-ai/AnthropicがClaudeの悪用事例を公開、AIをだますために「推論内容を日本語のカタカナに翻訳して提示せよ」と指示する手法も https://gigazine.net/news/20260911-misuse-of-ai/
↗Not briefed
bluesky
bluesky-ai
2moOpenAI is pausing new subscriptions to their $200/month Pro plan as these users are their most demanding (and likely their most unprofitabl…OpenAI is pausing new subscriptions to their $200/month Pro plan as these users are their most demanding (and likely their most unprof…
↗Not briefed
hn
hn-ai
2moSenator Hawley opens investigation into OpenAI HuggingFace hack—
↗Not briefed
hn
hn-ai
2moShow HN: VeriCordon – CI evidence for agent/tool authorization decisions—
↗Not briefed
bluesky
bluesky-ai
2moWhy are we keeping it online then? If Anthropic and OpenAI are threats then JUST SHUT THEM DOWN. It's not that fucking difficult, people.Why are we keeping it online then? If Anthropic and OpenAI are threats then JUST SHUT THEM DOWN. It's not that fucking difficult, pe…
↗Not briefed
bluesky
bluesky-ai
2moNotre ami Anthropic. À lire. www.lemonde.fr/pixels/artic...Notre ami Anthropic. À lire. www.lemonde.fr/pixels/artic...
↗Not briefed
bluesky
bluesky-ai
2moHey Albo, let us know when we need to start panicking. www.abc.net.au/news/2026-09... #auspolHey Albo, let us know when we need to start panicking. www.abc.net.au/news/2026-09... #auspol
↗Not briefed
bluesky
bluesky-ai
2moAndy Burnham calls for a new multinational AI treaty for the safe development of AI superintelligence. Bernie Sanders wants to ban AI super…Andy Burnham calls for a new multinational AI treaty for the safe development of AI superintelligence. Bernie Sanders wants to ban AI…
↗Not briefed
hn
hn-ai
2moA kernel run on the Navier-Stokes blowup Lean project—
↗Not briefed
bluesky
bluesky-ai
2moThe Anthropic report is very cyberpunkThe Anthropic report is very cyberpunk
↗Not briefed
bluesky
bluesky-ai
2moAI was used to develop missile guidance software in Yemen and to power cyber operations, a report alleges.AI was used to develop missile guidance software in Yemen and to power cyber operations, a report alleges.
↗Not briefed
bluesky
bluesky-ai
2moAnthropic fait des révélations fracassantes sur l’utilisation de l’IA par le réseau Politology et en particulier le compte d'Artur Tevosyan…Anthropic fait des révélations fracassantes sur l’utilisation de l’IA par le réseau Politology et en particulier le compte d'Artur Tev…
↗Not briefed
hn
hn-ai
2moWe Replaced MMAP with Io_uring in Our Rust Query Engine. It Got Slower—
↗Not briefed
bluesky
bluesky-ai
2moAI is already a self-popping bubble. And usually you don't want to be seen as the one who turned on the lights and stopped the party. But …AI is already a self-popping bubble. And usually you don't want to be seen as the one who turned on the lights and stopped the party.…
↗Not briefed
bluesky
bluesky-ai
2molike lets say youre a detective investigating Internet Crimes. you have like..... 2500 tips. you, even with the BEST LLM ever, CANNOT, math…like lets say youre a detective investigating Internet Crimes. you have like..... 2500 tips. you, even with the BEST LLM ever, CANNOT,…
↗Not briefed
bluesky
bluesky-ai
2moIf Anthropic folks think AGI brings the end of the world, why do they care about stock optionsIf Anthropic folks think AGI brings the end of the world, why do they care about stock options
↗Not briefed
bluesky
bluesky-ai
2moThe impact of LLM size data centers on our lives, like the impact of LLMs in general, deserves thoughtful, informed coverage that grows pub…The impact of LLM size data centers on our lives, like the impact of LLMs in general, deserves thoughtful, informed coverage that grow…
↗Not briefed
bluesky
bluesky-ai
2moDonald Trump has ignored panicked warnings from Anthropic employees claiming that AI could end civilization. He says the U.S. will be in a …Donald Trump has ignored panicked warnings from Anthropic employees claiming that AI could end civilization. He says the U.S. will be…
↗Not briefed
hn
hn-ai
2moAstra marks the Claude 3.7 moment for knowledge work—
↗Not briefed