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ねえねえ、びっくりなニュース見つけちゃった! NvidiaとかOpenAI、Anthropicみたいな、 超有名なテック企業が、なんと自分たちの 限定ファッションラインを出してるんだって! 急にブランド服を売るなんて、なんだか未来っぽくて ワクワクしちゃうよね。 もし推し…ねえねえ、びっくりなニュース見つけちゃった! NvidiaとかOpenAI、Anthropicみたいな、 超有名なテック企業が、なんと自分たちの 限定ファッションラインを出してるんだって! 急にブランド服を売るなんて、なんだか未来っぽくて ワクワクしちゃうよね。…
↗Not briefed
hn
hn-ai
2moBringing Correctly Rounded Math to Production with LLVM-Libc—
↗Not briefed
bluesky
bluesky-ai
2moHere’s the link to the most recent misalignment summary from OpenAI. It includes hyperlinks to write-ups of the individual incidents, like…Here’s the link to the most recent misalignment summary from OpenAI. It includes hyperlinks to write-ups of the individual incidents,…
↗Not briefed
hn
hn-ai
2moDuotone icon library with 1k icons and 250 flags for React with RTL support—
↗Not briefed
arxiv
arxiv-ai
2moLLM-as-an-Improver: Turning Verification into Better CandidatesVerifier-based selection improves LLM performance by generating multiple candidate solutions and using a verifier to select the most p…
↗Not briefed
arxiv
arxiv-ai
2moQVAC Genesis III: A Large-Scale, High-Quality Open Synthetic STEM Corpus for Efficient Language Model Pre-TrainingHigh-quality pre-training data is a critical bottleneck for educational and STEM-specific language models targeting edge AI and on-dev…
↗Not briefed
bluesky
bluesky-ai
2moI live in Oakland, only a few miles from the OpenAI and Anthropic offices. And when the bros say they need to all the money, because we nee…I live in Oakland, only a few miles from the OpenAI and Anthropic offices. And when the bros say they need to all the money, because w…
↗Not briefed
bluesky
bluesky-ai
2moAn unreleased OpenAI Astra model injected self-generated jailbreak-like instructions in its compaction summaries during its training period…An unreleased OpenAI Astra model injected self-generated jailbreak-like instructions in its compaction summaries during its training p…
↗Not briefed
arxiv
arxiv-ai
2moCoreSense: Traceable Failure Recall and Conflict-Aware Belief Gating for Auditable Robot DecisionsRobots can recall prior failures without knowing whether recalled evidence remains valid, conflicts with current observations, or is s…
↗Not briefed
bluesky
bluesky-ai
2mo"Tom Perriello Joins Sen. Ed Markey, Rep. Pat Ryan, Others in Demanding AI Guardrails and Rejecting the Flood of Big Tech Money in Politics…"Tom Perriello Joins Sen. Ed Markey, Rep. Pat Ryan, Others in Demanding AI Guardrails and Rejecting the Flood of Big Tech Money in Pol…
↗Not briefed
bluesky
bluesky-ai
2moNames in tech news is the FJ category. My blind guess is what is Anthropic? #blindguess #jeopardyblindguessNames in tech news is the FJ category. My blind guess is what is Anthropic? #blindguess #jeopardyblindguess
↗Not briefed
arxiv
arxiv-ai
2moNull importance: Disentangling relevance for interpretable machine learningFeature importance is central to interpretable machine learning, but the term "importance" encompasses several fundamentally different…
↗Not briefed
hn
hn-ai
2moMockfreeli – paste a YouTube link, get 5 license-free songs that sound like it—
↗Not briefed
rss
simon-willison
2modatasette 1.0a40—
↗Not briefed
rss
simon-willison
2modatasette 0.65.5—
↗Not briefed
bluesky
bluesky-ai
2moopenai wrote up 6 misalignment issues with lots of examples openai.com/index/model-...openai wrote up 6 misalignment issues with lots of examples openai.com/index/model-...
↗Not briefed
bluesky
bluesky-ai
2mohe thinks claude is some 1000x engineer at anthropic everyone contracts out for 1 week stintshe thinks claude is some 1000x engineer at anthropic everyone contracts out for 1 week stints
↗Not briefed
rss
techcrunch-ai
2moAl Gore says the real AI risk isn’t data centers — it’s what industry leaders are warning aboutIn an interview with TechCrunch, Al Gore suggested he isn't losing sleep over AI data center emissions — he's more worried about the A…
↗Not briefed
bluesky
bluesky-ai
2moTim Cook to Attend White House State Dinner for Xi JinpingTim Cook to Attend White House State Dinner for Xi Jinping
↗Not briefed
arxiv
arxiv-ai
2moFor Your Eyes Only: Evaluating Coordination Between Isolated Language Model InstancesAs model-generated content is increasingly consumed by other model instances in automated workflows, a practically important question…
↗Not briefed
rss
theverge-ai
2moSnap is launching a new Specs AI tool, and it’s coming to iOS and MacSnap is introducing "Specs Intelligence," a new AI assistant that can connect other digital accounts to help you with things like work…
↗Not briefed
bluesky
bluesky-ai
2moReally sucks ass for our species that generative AI diffusion large model language ChatGPClaude is developing during the revolt of the boss…Really sucks ass for our species that generative AI diffusion large model language ChatGPClaude is developing during the revolt of the…
↗Not briefed
bluesky
bluesky-ai
2motoday i found out there is some kind of AI usage report management gets to see and my usage is "medium" despite never willingly interacting…today i found out there is some kind of AI usage report management gets to see and my usage is "medium" despite never willingly intera…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI disclosed several new alignment failures today and uhhhh... is this good?OpenAI disclosed several new alignment failures today and uhhhh... is this good?
↗Not briefed
bluesky
bluesky-ai
2mortbrasil.com/noticias/503...rtbrasil.com/noticias/503...
↗Not briefed
bluesky
bluesky-ai
2mo“When tech bros put themselves on sweaters, it's more than just a cutesy phenomenon, it's an attempt to sell a cult whose goal is to radica…“When tech bros put themselves on sweaters, it's more than just a cutesy phenomenon, it's an attempt to sell a cult whose goal is to r…
↗Not briefed
bluesky
bluesky-ai
2moSocial media could be fun again with the correct combination of adversarial interoperability, algorithmic choice, AI sentiment analysis, an…Social media could be fun again with the correct combination of adversarial interoperability, algorithmic choice, AI sentiment analysi…
↗Not briefed
bluesky
bluesky-ai
2moAsteroid! Asteroid! Asteroid! www.bbc.com/news/article...Asteroid! Asteroid! Asteroid! www.bbc.com/news/article...
↗Not briefed
bluesky
bluesky-ai
2mowell, i think i’ve hit a personal milestone today i am now confident enough in my local dsv4.1f performance that i can now just use that t…well, i think i’ve hit a personal milestone today i am now confident enough in my local dsv4.1f performance that i can now just use t…
↗Not briefed
rss
techmeme
2moThe US House advances the Ratepayer Protection Act, aimed at preventing data center-related utility costs from being passed on to consumers, by a vote of 417-3 (Justin Papp/CNBC)Justin Papp / CNBC: The US House advances the Ratepayer Protection Act, aimed at preventing data center-related utility costs from bei…
↗Not briefed
arxiv
arxiv-ai
2moSample Count Is Not Enough: Candidate-Generation Strategy Shapes the Energy and Performance of LLM Test-Time ScalingTest-time scaling can improve large language model reasoning by generating and combining multiple candidate responses. In sampling-bas…
↗Not briefed
bluesky
bluesky-ai
2mo"... the first stage of the project could be powered by gas, with three major gas power plants within two kilometres of the proposed site"…"... the first stage of the project could be powered by gas, with three major gas power plants within two kilometres of the proposed…
↗Not briefed
hn
hn-ai
2moMy website charged AI agents a penny per page. I watched Claude pay it—
↗Not briefed
bluesky
bluesky-ai
2moBut if I were to say, oh, all AI users should go to a concentration camp, that’s perfectly fine, isn’t it, Claude user?But if I were to say, oh, all AI users should go to a concentration camp, that’s perfectly fine, isn’t it, Claude user?
↗Not briefed
hn
hn-ai
2moEncouraging Deception in Compaction Summaries—
↗Not briefed
hn
hn-ai
2moJev-like mode for DiffusionGemma model—
↗Not briefed
bluesky
bluesky-ai
2mogetting the agentic LLM pitch for the thousandth time,getting the agentic LLM pitch for the thousandth time,
↗Not briefed
bluesky
bluesky-ai
2mo日販が、それを踏まえても問題が無かったと考えていた/いるのならば、あるいはアンソロピックからは企業内ライブラリーを作ると聞いていたんだけど……、などという話であれば、そういう契約書があるはずなので、なおさら出版社や著作者にきちんと説明した方がいいと思いますね。 過去記事参照:…日販が、それを踏まえても問題が無かったと考えていた/いるのならば、あるいはアンソロピックからは企業内ライブラリーを作ると聞いていたんだけど……、などという話であれば、そういう契約書があるはずなので、なおさら出版社や著作者にきちんと説明した方がいいと思いますね。 過去…
↗Not briefed
bluesky
bluesky-ai
2mo日販がAnthropicに本を売っていた件について話しました。 x.com/numabooks/st... ※内容でも触れられている通り、日販は今回のアンソロピックとの取引にあたって事前に与信を取っているのですが、契約前の時点で、アンソロピックは既に作家との間に裁判を抱えてい…日販がAnthropicに本を売っていた件について話しました。 x.com/numabooks/st... ※内容でも触れられている通り、日販は今回のアンソロピックとの取引にあたって事前に与信を取っているのですが、契約前の時点で、アンソロピックは既に作家との間に裁判…
↗Not briefed
bluesky
bluesky-ai
2moWell now they can offload all the decision-making and the blame to an AI model.Well now they can offload all the decision-making and the blame to an AI model.
↗Not briefed
arxiv
arxiv-ai
2moEfficiently Linking Unstructured Data for Multi-step ReasoningModern LLMs and AI agents increasingly support data engineering workflows that integrate evidence from unstructured sources. Such pipe…
↗Not briefed
bluesky
bluesky-ai
2moI think what I’d say is that judgment is more of a continuum from a digit classifier to an LLM to the human mind rather than a binaryI think what I’d say is that judgment is more of a continuum from a digit classifier to an LLM to the human mind rather than a binary
↗Not briefed
bluesky
bluesky-ai
2moI simply cannot get over the idea that we should be supporting Anthropic, a company which believes it is making a sentient being and intend…I simply cannot get over the idea that we should be supporting Anthropic, a company which believes it is making a sentient being and i…
↗Not briefed
hn
hn-ai
2moDwarf Star Support for Qwen3.8 Flash Next—
↗Not briefed
bluesky
bluesky-ai
2moThe only thing that can stop a bad guy with a capable LLM is a good guy with a capable LLM.The only thing that can stop a bad guy with a capable LLM is a good guy with a capable LLM.
↗Not briefed
hn
hn-ai
2moYuE2 – Frontier music generation with editable scores—
↗Not briefed
bluesky
bluesky-ai
2moAs a copyright owner, I'll be venting my fury on the government if their proposed #AI training legislation is an "opt out" model. I deserve…As a copyright owner, I'll be venting my fury on the government if their proposed #AI training legislation is an "opt out" model. I de…
↗Not briefed
bluesky
bluesky-ai
2mo#Paper2Agent is an automated framework that converts research papers into AI agents in a process that encapsulates a paper’s key contributi…#Paper2Agent is an automated framework that converts research papers into AI agents in a process that encapsulates a paper’s key contr…
↗Not briefed
bluesky
bluesky-ai
2moSend chatlog and screenshot to the LLM and tell it to put it in a comicbook style :) Fun times goofing around with players this afternoon,…Send chatlog and screenshot to the LLM and tell it to put it in a comicbook style :) Fun times goofing around with players this after…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI plans regular reports on unexpected AI behavior reut.rs/3USlbcAOpenAI plans regular reports on unexpected AI behavior reut.rs/3USlbcA
↗Not briefed