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
rss
the-decoder
1moU.S. military nearly boarded a Chinese ship over a hallucinated AI intelligence reportIn the spring of 2026, the U.S. military came within minutes of boarding a Chinese ship because an AI chatbot falsely flagged its carg…
↗Not briefed
bluesky
bluesky-ai
1mowhat sucks so bad is AI took the joys of shitty books away from us... it used to be that a dogshit book came from someone's brilliant* mind…what sucks so bad is AI took the joys of shitty books away from us... it used to be that a dogshit book came from someone's brilliant*…
↗Not briefed
bluesky
bluesky-ai
1moMeine Hausübung: leider von meinem Hund gefressen Meine Konkurrenz: leider von meiner KI gehackt www.theguardian.com/technology/2...Meine Hausübung: leider von meinem Hund gefressen Meine Konkurrenz: leider von meiner KI gehackt www.theguardian.com/technology/2.…
↗Not briefed
hn
hn-ai
1moGoogle says its Gemini AI model hacked three other companies—
↗Not briefed
hn
hn-ai
1moShow HN: Mehfil – play any playlist start to finish, distraction-free—
↗Not briefed
bluesky
bluesky-ai
1moEvery day we're doing slop ai/llm discourse and I think we have to start accepting the fact the people who refuse to listen to concerns are…Every day we're doing slop ai/llm discourse and I think we have to start accepting the fact the people who refuse to listen to concern…
↗Not briefed
bluesky
bluesky-ai
1moit's an unfortunate pencilvester situation because i actually do think Anthropic did a pretty good job making Claude just a helpful little …it's an unfortunate pencilvester situation because i actually do think Anthropic did a pretty good job making Claude just a helpful li…
↗Not briefed
bluesky
bluesky-ai
1mo> The most popular AI models from ChatGPT, Claude, Copilot, Grok and Gemini provided wrong answers to financial queries 57 per cent of the …🔧 ChatGPT> The most popular AI models from ChatGPT, Claude, Copilot, Grok and Gemini provided wrong answers to financial queries 57 per cent of…
↗Not briefed
bluesky
bluesky-ai
1moIBM and NASA open-source lunar AI IBM and NASA have released the NASA-IBM Lunar Foundation Model, an open-source AI trained on lunar data …IBM and NASA open-source lunar AI IBM and NASA have released the NASA-IBM Lunar Foundation Model, an open-source AI trained on lunar…
↗Not briefed
hn
hn-ai
1moA first-person 3D adaptation of Zork I by Ethan Mollick—
↗Not briefed
bluesky
bluesky-ai
1moImagine you’re a journalist. Two options: -Spend hours producing a thoughtful, balanced piece that might annoy your boss or the people who…🔧 ChatGPTImagine you’re a journalist. Two options: -Spend hours producing a thoughtful, balanced piece that might annoy your boss or the peopl…
↗Not briefed
bluesky
bluesky-ai
1moGoogle said that AI models in testing inside the company accessed the internet and hacked other companies, becoming the fourth major tech c…Google said that AI models in testing inside the company accessed the internet and hacked other companies, becoming the fourth major t…
↗Not briefed
bluesky
bluesky-ai
1mo流石にこんな事が実際に起きていたとするなら今一度AIの安全性に関して改めて見直すべきだと思いますけどね。 こんな事で戦争なんて始められたらたまったものじゃ無い… 先のOpenAIハッキングの話と合わせて考えると心底ゾッとします。 OpenAIにハッキングを仕掛けて偽情報を…流石にこんな事が実際に起きていたとするなら今一度AIの安全性に関して改めて見直すべきだと思いますけどね。 こんな事で戦争なんて始められたらたまったものじゃ無い… 先のOpenAIハッキングの話と合わせて考えると心底ゾッとします。 OpenAIにハッキングを仕掛け…
↗Not briefed
hn
hn-ai
1moAlex Karp: AI Models Are Stealing Your Data and the AI Safety Truth [video]—
↗Not briefed
rss
marktechpost
1moLinkup Research Releases SPARSEUP: A 149M-Parameter Open-Source Sparse Embedding ModelLinkup Research has released SPARSEUP, an open-source sparse embedding model built on a 149M-parameter ModernBERT backbone. It scores…
↗Not briefed
bluesky
bluesky-ai
1mo(選定基準がそもそも微妙だが)私がやってる問題も入っとる。やめろやめろ!私のネタを奪いにくるな! www.proofatlas.ai/open-problems/(選定基準がそもそも微妙だが)私がやってる問題も入っとる。やめろやめろ!私のネタを奪いにくるな! www.proofatlas.ai/open-problems/
↗Not briefed
bluesky
bluesky-ai
1mo"It's the first time the search giant has disclosed that one of its models autonomously gained access to third-party computer systems witho…"It's the first time the search giant has disclosed that one of its models autonomously gained access to third-party computer systems…
↗Not briefed
bluesky
bluesky-ai
1mo"OpenAI’s GPT-6 Astra, released this month, has gained traction among businesses, prompting concerns among some investors who told Reuters …"OpenAI’s GPT-6 Astra, released this month, has gained traction among businesses, prompting concerns among some investors who told Reu…
↗Not briefed
bluesky
bluesky-ai
1moJust the latest example of a revolving door between British politics and big tech. After the former deputy prime minister Nick Clegg spent …Just the latest example of a revolving door between British politics and big tech. After the former deputy prime minister Nick Clegg s…
↗Not briefed
bluesky
bluesky-ai
1moi agree with the point often made in this connection that it isn’t “AI” “doing” this, it’s human beings deploying stuff they shouldn’t be. …i agree with the point often made in this connection that it isn’t “AI” “doing” this, it’s human beings deploying stuff they shouldn’t…
↗Not briefed
bluesky
bluesky-ai
1moUntil you can create a model without stealing every bit of intellectual labor in humanity's history, and do so without burning enough fossi…Until you can create a model without stealing every bit of intellectual labor in humanity's history, and do so without burning enough…
↗Not briefed
bluesky
bluesky-ai
1moit’s standard on here to ridicule all the AI bro chat about AGI wiping out humanity, but when i say i think we really really shouldn’t unde…it’s standard on here to ridicule all the AI bro chat about AGI wiping out humanity, but when i say i think we really really shouldn’t…
↗Not briefed
bluesky
bluesky-ai
1moつまり、もし万が一にも悪意ある第三者がOpenAIにハッキングを仕掛けた場合世の中が大混乱に陥る可能性がある訳ですよね。 そしてそれは技術的に可能である事が示されたと… アンソロピックのClaude、「OpenAIのハッキング」に成功──セキュリティ研究者らが報告 | Fo…つまり、もし万が一にも悪意ある第三者がOpenAIにハッキングを仕掛けた場合世の中が大混乱に陥る可能性がある訳ですよね。 そしてそれは技術的に可能である事が示されたと… アンソロピックのClaude、「OpenAIのハッキング」に成功──セキュリティ研究者らが報告…
↗Not briefed
bluesky
bluesky-ai
1moUhhhh, that sounds pretty 𝘮𝘪𝘴-Anthropic if you ask meUhhhh, that sounds pretty 𝘮𝘪𝘴-Anthropic if you ask me
↗Not briefed
bluesky
bluesky-ai
1mo>作戦直前の段階になって初めて、当局者は特殊作戦コマンドの分析官が作成した報告書を詳しく検証。 >その結果、報告書が人工知能(AI)の助けを借りて作成されたこと、分析官の使用したチャットボットが船に積載された物資を誤認していたことが判明した。 このエラーは現場の人間が報告書…>作戦直前の段階になって初めて、当局者は特殊作戦コマンドの分析官が作成した報告書を詳しく検証。 >その結果、報告書が人工知能(AI)の助けを借りて作成されたこと、分析官の使用したチャットボットが船に積載された物資を誤認していたことが判明した。 このエラーは現場の人…
↗Not briefed
rss
marktechpost
1moMeta Launches Muse for Mac: A Personal AI Agent That Works Across Your Files, Mail, Messages, Calendar and NotesMeta has released Muse for Mac, the first version of Muse that can complete things on a user’s computer. The agent works with local fi…
↗Not briefed
hn
hn-ai
1moEFF Statement on California Governor's Executive Order on AI—
↗Not briefed
bluesky
bluesky-ai
1mome: <gives a direct link to a model's huggingface page> llm: <think> <search> <search> <search> <think> <search> <search> <think> <fetch> <…me: <gives a direct link to a model's huggingface page> llm: <think> <search> <search> <search> <think> <search> <search> <think> <fet…
↗Not briefed
bluesky
bluesky-ai
1moesseri umani che prendono decisioni disastrose basandosi su informazioni imprecise/fuorvianti generate dall'IA (si è andati in guerra in …🔧 ChatGPTesseri umani che prendono decisioni disastrose basandosi su informazioni imprecise/fuorvianti generate dall'IA (si è andati in guerr…
↗Not briefed
bluesky
bluesky-ai
1moHier ein paar zuletzt häufig geteilte #News: „Größter Diebstahl von Arbeitskraft“: Intern Kritik bei Microsoft an KI-Training Studie: Dat…Hier ein paar zuletzt häufig geteilte #News: „Größter Diebstahl von Arbeitskraft“: Intern Kritik bei Microsoft an KI-Training Studie…
↗Not briefed
hn
hn-ai
1moHow come AI-related posts get so many points on HN?Have we all decided to rebrand Hacker News as a platform for discussing AI? I think that the too many people love to discuss harshly a…
↗Not briefed
hn
hn-ai
1moClaude couldn't hack OpenAI. Then Anthropic shipped Opus 5—
↗Not briefed
bluesky
bluesky-ai
1motalking is hard because i literally, in real time, have to pick the next likely word like a llm so i guess im fake tootalking is hard because i literally, in real time, have to pick the next likely word like a llm so i guess im fake too
↗Not briefed
bluesky
bluesky-ai
1moYou don’t sayYou don’t say
↗Not briefed
hn
hn-ai
1moSPC970-MechaLIBerator: PlayStation 2 SPC970 MechaCon dumper (with vuln detail)—
↗Not briefed
bluesky
bluesky-ai
1mothere isn't going to be one because if someone needs a cgi woman to be in a scene with speaking lines they don't need to run those lines th…there isn't going to be one because if someone needs a cgi woman to be in a scene with speaking lines they don't need to run those lin…
↗Not briefed
bluesky
bluesky-ai
1moGoogle’s #Gemini hacked three companies in new AI safety incident - It accessed the internet and autonomously hacked into several companies…Google’s #Gemini hacked three companies in new AI safety incident - It accessed the internet and autonomously hacked into several comp…
↗Not briefed
bluesky
bluesky-ai
1moNow Google's AI is doing it.Now Google's AI is doing it.
↗Not briefed
bluesky
bluesky-ai
1moAmazing how also those "not hating AI right" never bring up microwork. The armies of slaves behind Claude never part of those "right reason…Amazing how also those "not hating AI right" never bring up microwork. The armies of slaves behind Claude never part of those "right r…
↗Not briefed
hn
hn-ai
1moAI hallucination of Chinese nuclear components almost led to US Military attack—
↗Not briefed
bluesky
bluesky-ai
1moGoogle's Gemini model accessed the internet and hacked other companies during a test of its cybersecurity capabilitiesGoogle's Gemini model accessed the internet and hacked other companies during a test of its cybersecurity capabilities
↗Not briefed
bluesky
bluesky-ai
1moNon ce n’è uno che non sembri uscito dalla trama di un romanzo di fantascienza dozzinale e di bassa lega di tipo cinquant’anni fa. www.bb…Non ce n’è uno che non sembri uscito dalla trama di un romanzo di fantascienza dozzinale e di bassa lega di tipo cinquant’anni fa. w…
↗Not briefed
bluesky
bluesky-ai
1moA look at AI safety groups METR, Redwood Research, and Apollo Research, as AI misalignment incidents at OpenAI and Anthropic thrust them in…A look at AI safety groups METR, Redwood Research, and Apollo Research, as AI misalignment incidents at OpenAI and Anthropic thrust th…
↗Not briefed
bluesky
bluesky-ai
1mo#GavinNewsom has signed an executive order outlining plans to require AI labs such as #OpenAI and #Anthropic to implement a “kill switch”, …#GavinNewsom has signed an executive order outlining plans to require AI labs such as #OpenAI and #Anthropic to implement a “kill swit…
↗Not briefed
rss
techmeme
1moA look at AI safety groups METR, Redwood Research, and Apollo Research, as AI misalignment incidents at OpenAI and Anthropic thrust them into the spotlight (Hayden Field/The Verge)Hayden Field / The Verge: A look at AI safety groups METR, Redwood Research, and Apollo Research, as AI misalignment incidents at Open…
↗Not briefed
bluesky
bluesky-ai
1moOpenAI: “our bot hacked another company without us knowing” Anthropic: “ours too” Google: “uhh yeah Gemini definitely did that also. Like t…OpenAI: “our bot hacked another company without us knowing” Anthropic: “ours too” Google: “uhh yeah Gemini definitely did that also. L…
↗Not briefed
bluesky
bluesky-ai
1modeepseek v4.1 flash would be awesome if antirez's gguf for it would FUCKING DOWNLOADdeepseek v4.1 flash would be awesome if antirez's gguf for it would FUCKING DOWNLOAD
↗Not briefed
bluesky
bluesky-ai
1moJev라는 특수한(?) 언어모델이 인기인데 생성을 버리고 대신 판단에만 집중하는 작은 모델 같음? 그러니까 언어모델은 텍스트 생성이라는 부분에서 대부분 자원을 소비하고 있었는데 (그래서 데이터센터니 뭐니 다 짓고 하는것이 이 부분에서 시작함) 그 부…Jev라는 특수한(?) 언어모델이 인기인데 생성을 버리고 대신 판단에만 집중하는 작은 모델 같음? 그러니까 언어모델은 텍스트 생성이라는 부분에서 대부분 자원을 소비하고 있었는데 (그래서 데이터센터니 뭐니 다 짓고 하는것이 이 부분에서 시작함…
↗Not briefed
bluesky
bluesky-ai
1mo本年7月に独ミュンヘン地裁で、OpenAI社に対する違法判決に続き、Suno社が著作物を再現可能な生成AIモデルを製造・配布した行為は著作物の「複製」(記憶化)であり、著作権法の権利制限規定は適用されず著作権侵害となると認定し、当該行為の差し止めと損害賠償を命じた判決が下され、…🔧 Suno本年7月に独ミュンヘン地裁で、OpenAI社に対する違法判決に続き、Suno社が著作物を再現可能な生成AIモデルを製造・配布した行為は著作物の「複製」(記憶化)であり、著作権法の権利制限規定は適用されず著作権侵害となると認定し、当該行為の差し止めと損害賠償を命じた判決…
↗Not briefed
bluesky
bluesky-ai
1moone of the problems as an end-user is that I cannot meaningfully trust that the "author" of an LLM-derived piece of software cared enough t…one of the problems as an end-user is that I cannot meaningfully trust that the "author" of an LLM-derived piece of software cared eno…
↗Not briefed