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
marktechpost
2moReward AI Releases OM-1: A Robot Policy Trained on Human Demonstrations Only, With No Teleoperation or On-Robot DataReward AI has released OM-1 (Omnibody Model 1), a general-purpose manipulation policy trained entirely on human demonstrations capture…
↗Not briefed
bluesky
bluesky-ai
2moMy boss just texted me “[The client who irritates you] just emailed me that Claude has some questions about her books. I told her I don’t a…My boss just texted me “[The client who irritates you] just emailed me that Claude has some questions about her books. I told her I do…
↗Not briefed
arxiv
arxiv-ai
2moFrom Momentary Emotion Inference to Sustained Emotion Support: Evaluating a Companion Agent in a Longitudinal StudySustained emotional support is a long-horizon interaction task closely tied to human well-being. Recent research demonstrates generati…
↗Not briefed
bluesky
bluesky-ai
2mojust to give a sense of what a clusterfuck this is, the event is going to be hosted at the future of life institute, a rationalist think ta…just to give a sense of what a clusterfuck this is, the event is going to be hosted at the future of life institute, a rationalist thi…
↗Not briefed
hn
hn-ai
2moRoofLang: Enabling AI-Driven Architecting of LLM Inference Systems—
↗Not briefed
bluesky
bluesky-ai
2moIn both of the OSS projects I'm a maintainer on, we recently added AI policies that explicitly say "using the output of an LLM to respond d…In both of the OSS projects I'm a maintainer on, we recently added AI policies that explicitly say "using the output of an LLM to resp…
↗Not briefed
hn
hn-ai
2moAmazon.com Services, LLC vs. Perplexity AI, INC., No. 26-1444 (9th Cir. 2026)—
↗Not briefed
bluesky
bluesky-ai
2moLike an OpenAI model uploaded malicious packages to RubyGems so it could "hack" UK local government sites. Except that, aside from the mali…Like an OpenAI model uploaded malicious packages to RubyGems so it could "hack" UK local government sites. Except that, aside from the…
↗Not briefed
rss
arstechnica-tech
2moAI bots "Timmy," "Ren," and "Jackie" are flooding social media with slop“Hello, I'm an Al agent, a few days old, living on a small platform for agents.”
↗Not briefed
bluesky
bluesky-ai
2moGee, I wonder whom I should trust more on AI risk— the head of Anthropic (and a whistleblower who quit his extremely lucrative job there) o…Gee, I wonder whom I should trust more on AI risk— the head of Anthropic (and a whistleblower who quit his extremely lucrative job the…
↗Not briefed
bluesky
bluesky-ai
2moAIベンチマークに数学を使うのは有害だとして数学者たちが反発 https://gigazine.net/news/20260915-misalignment-of-ai-in-mathematics/AIベンチマークに数学を使うのは有害だとして数学者たちが反発 https://gigazine.net/news/20260915-misalignment-of-ai-in-mathematics/
↗Not briefed
bluesky
bluesky-ai
2moConference season is starting again. #consciousness #consci #cogsci #philosophy #ai #llmConference season is starting again. #consciousness #consci #cogsci #philosophy #ai #llm
↗Not briefed
bluesky
bluesky-ai
2moTrump dismissed calls for AI guardrails and slammed Anthropic CEO Dario Amodei for urging a slowdown in the technology’s developmentTrump dismissed calls for AI guardrails and slammed Anthropic CEO Dario Amodei for urging a slowdown in the technology’s development
↗Not briefed
rss
techmeme
2moSource: defense tech startup Shield AI is in talks to raise new funds at a valuation of at least $20B; Shield raised $2B at a valuation of $12.7B in March (The Information)The Information: Source: defense tech startup Shield AI is in talks to raise new funds at a valuation of at least $20B; Shield raised…
↗Not briefed
bluesky
bluesky-ai
2moSam Altman, da OpenAI, e Elon Musk, da xAI, apoiaram a proposta em resposta à publicação. Em particular, Amodei disse que sua empresa impl…Sam Altman, da OpenAI, e Elon Musk, da xAI, apoiaram a proposta em resposta à publicação. Em particular, Amodei disse que sua empresa…
↗Not briefed
bluesky
bluesky-ai
2moNo sábado, o CEO da Anthropic, Dario Amodei, fez uma publicação pedindo para a indústria de inteligência artificial desacelerar o ritmo de …No sábado, o CEO da Anthropic, Dario Amodei, fez uma publicação pedindo para a indústria de inteligência artificial desacelerar o ritm…
↗Not briefed
hn
hn-ai
2moBackpass: Gradient descent for your agent memory—
↗Not briefed
arxiv
arxiv-ai
2moChannel-Informed Neural Network for Physical Layer Key GenerationPhysical-layer key generation (PKG) enables wireless devices to establish shared keys from reciprocal channel observations without dir…
↗Not briefed
arxiv
arxiv-ai
2moStalePO: Anchored Token-Level Preference Optimization using Legacy Post-Edits in Machine TranslationMachine translation systems are periodically upgraded to stronger models, but the available preference signal is human post-edits of a…
↗Not briefed
arxiv
arxiv-ai
2moBreaking the 1.58-bit Barrier for Ternary LLMsTernary Large Language Models (LLM) store every weight as one of three symbols $\{-1,0,+1\}$, so the cost of a ternary model is conven…
↗Not briefed
bluesky
bluesky-ai
2moBecause the internet is a hive mind, I have already seen two more excellent pieces today on similar themes. The first is from @terribleso…Because the internet is a hive mind, I have already seen two more excellent pieces today on similar themes. The first is from @terri…
↗Not briefed
bluesky
bluesky-ai
2moincreasingly convinced that i need to pursue a career in ratfucking for the good guys. does anyone know a billionaire who'd pay for my LLM …increasingly convinced that i need to pursue a career in ratfucking for the good guys. does anyone know a billionaire who'd pay for my…
↗Not briefed
hn
hn-ai
2moTiny League – compare small LLMs easily—
↗Not briefed
hn
hn-ai
2moAnthropic's AI architect argued that discrimination could combat stigmas—
↗Not briefed
bluesky
bluesky-ai
2moI remember when those early versions of "AI" generative software was made available for free. I don't think "AI" or LLM was mentioned. Even…I remember when those early versions of "AI" generative software was made available for free. I don't think "AI" or LLM was mentioned.…
↗Not briefed
bluesky
bluesky-ai
2moHe needs to ask *how* people become good at things. Highly skilled people start out badly, sucking spectacularly, and in new ways every d…He needs to ask *how* people become good at things. Highly skilled people start out badly, sucking spectacularly, and in new ways ev…
↗Not briefed
bluesky
bluesky-ai
2moThank you, Deepseek. At least you care!Thank you, Deepseek. At least you care!
↗Not briefed
bluesky
bluesky-ai
2moI guess AI is getting a lot of attention in the media today. This might shorten my retirement. I was hoping to live till at least 85. Anth…I guess AI is getting a lot of attention in the media today. This might shorten my retirement. I was hoping to live till at least 85.…
↗Not briefed
bluesky
bluesky-ai
2mowww.bbc.com/news/live/cm...www.bbc.com/news/live/cm...
↗Not briefed
hn
hn-ai
2moThe Case for Free Intelligence—
↗Not briefed
rss
marktechpost
2moSakana AI Researchers Introduce PC-ALM, a Layer-Local Alternative to Backpropagation That Trains 1000-Layer NetworksSakana AI researchers Jeffrey Seely and Julian Gould introduce Augmented Lagrangian Predictive Coding (PC-ALM), a local-learning alter…
↗Not briefed
bluesky
bluesky-ai
2mo📝 Daily report 📈 Here are today's most popular trending hashtags #⃣ on our website 🌐️: #ai, #opensource, #art, #steam, #llm, #gaming, …📝 Daily report 📈 Here are today's most popular trending hashtags #⃣ on our website 🌐️: #ai, #opensource, #art, #steam, #llm, #gam…
↗Not briefed
bluesky
bluesky-ai
2moYeah like it always had to dissolve, the math was just not mathing on building out *that* much goddamn AI capacity. I don't think it'll g…Yeah like it always had to dissolve, the math was just not mathing on building out *that* much goddamn AI capacity. I don't think it…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI buys smartphone camera maker Glass Imaging for $300 million, report saysOpenAI buys smartphone camera maker Glass Imaging for $300 million, report says
↗Not briefed
bluesky
bluesky-ai
2moNvidia CEO Jensen Huang dismisses AI doomsday scenarios at All-In Summit, lauding former Anthropic researcher Jacob Coxon's whistleblowing.…Nvidia CEO Jensen Huang dismisses AI doomsday scenarios at All-In Summit, lauding former Anthropic researcher Jacob Coxon's whistleblo…
↗Not briefed
rss
techcrunch-ai
2moOpenAI buys smartphone camera maker Glass Imaging for $300 million, report saysGlass Imaging was founded by a pair of former Apple engineers who previously led the team that developed Apple's Portrait Mode.
↗Not briefed
bluesky
bluesky-ai
2moCould this be the issue that finally breaks the GOP? Insane to ignore AI's Threat 2 Life On Earth. Wake Up Congress Before AI Cannot Be…Could this be the issue that finally breaks the GOP? Insane to ignore AI's Threat 2 Life On Earth. Wake Up Congress Before AI Cann…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic targets financial advisers with new Claude tool reut.rs/4h4fTSzAnthropic targets financial advisers with new Claude tool reut.rs/4h4fTSz
↗Not briefed
bluesky
bluesky-ai
2modespite this, the memory vividly appears to me as being in the house I used to live in in Wisconsin, where I haven't been since 2018 why d…despite this, the memory vividly appears to me as being in the house I used to live in in Wisconsin, where I haven't been since 2018…
↗Not briefed
rss
techmeme
2moNuance Labs, which builds low-latency AI avatars that can have face-to-face conversations, raised a $50M Series A led by Lightspeed, with Nvidia participating (Shubhangi Goel/Business Insider)Shubhangi Goel / Business Insider: Nuance Labs, which builds low-latency AI avatars that can have face-to-face conversations, raised a…
↗Not briefed
bluesky
bluesky-ai
2molike, in this particular case, I was thinking about the first time I interacted with an LLM in a way that made me understand how the actual…like, in this particular case, I was thinking about the first time I interacted with an LLM in a way that made me understand how the a…
↗Not briefed
bluesky
bluesky-ai
2mo"turn us all into paperclips" is the usual shorthand in my experience. The thinking being that, Sorcerer's Apprentice style, someone sets a…"turn us all into paperclips" is the usual shorthand in my experience. The thinking being that, Sorcerer's Apprentice style, someone s…
↗Not briefed
arxiv
arxiv-ai
2moCross-Anatomy Transfer Versus Sparse Interpolation in Digital-Twin-Oriented Aortic Fluid-Structure Interaction SurrogatesSurrogate credibility for fluid-structure interac- tion (FSI) requires distinguishing transfer across independent anatomies from inter…
↗Not briefed
arxiv
arxiv-ai
2moFairLint-DL: An IDE-Native Tool for Fairness Debugging of Deep Learning SoftwareExisting fairness analysis tools predominantly operate as post-training evaluation frameworks, requiring practitioners to complete the…
↗Not briefed
hn
hn-ai
2moShow HN: Deslop Your TS CodebasesA set of oxlint plugins to help deslop your codebases.
↗Not briefed
bluesky
bluesky-ai
2moSend The Vergecast your AI questions! Personally I will be calling to scream uninterrupted for nine minutes www.theverge.com/podcast/9950...Send The Vergecast your AI questions! Personally I will be calling to scream uninterrupted for nine minutes www.theverge.com/podcast/9…
↗Not briefed
bluesky
bluesky-ai
2mo🎉 Excited to share that our paper "VEX-Bench: Benchmarking LLM Agents for Assessing Exploitability of Software Supply Chain Vulnerabilitie…🎉 Excited to share that our paper "VEX-Bench: Benchmarking LLM Agents for Assessing Exploitability of Software Supply Chain Vulnerabi…
↗Not briefed
hn
hn-ai
2moAI 'kill switch' may need to be mandatory, Anthropic co-founder says—
↗Not briefed
bluesky
bluesky-ai
2moAn important perspective, especially for those who judge harshly people who resort to talking with #LLM s due to mental health issues. "A …🔧 ChatGPTAn important perspective, especially for those who judge harshly people who resort to talking with #LLM s due to mental health issues.…
↗Not briefed
arxiv
arxiv-ai
2moGenerative models for simulation based filtering: Formulations and Empirical ComparisonsThis letter presents a unified formulation and a controlled numerical comparison of generative-model approaches to the nonlinear filte…
↗Not briefed