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
hn
hn-ai
1moTell HN: The special .github repo name for orgs and .codex-root AGENTS.md—
↗Not briefed
rss
techmeme
1moLucid and Bolt sign an initial agreement to deploy a fleet of 25K+ robotaxis in Europe built on Lucid's affordable EV platform, set to launch as soon as 2028 (Financial Times)Financial Times: Lucid and Bolt sign an initial agreement to deploy a fleet of 25K+ robotaxis in Europe built on Lucid's affordable EV…
↗Not briefed
arxiv
arxiv-ai
1moThe Bias of Nonlinear Two-Time-scale Stochastic Approximation under Constant Step-SizesTwo-timescale stochastic approximation (TTSA) is a fundamental tool for analyzing coupled iterative algorithms in reinforcement learni…
↗Not briefed
arxiv
arxiv-ai
1moXeno-Interpretability: Investigating the Alien Minds of LLMsLarge language models are usually interpreted through concepts that humans already possess: truthfulness, refusal, deception, personal…
↗Not briefed
hn
hn-ai
1moInterest in programming languages hit a decade low after Claude Code launch—
↗Not briefed
arxiv
arxiv-ai
1moLearning Principal-Agent Contracts for Equitable Smallholder Carbon Farming under Moral Hazard and Adverse SelectionAgricultural soils are a major untapped carbon sink. Carbon farming is emerging as a promising practice for tapping this potential. Sm…
↗Not briefed
hn
hn-ai
1moTrust, but benchmark: How we let an AI agent optimize Elasticsearch—
↗Not briefed
bluesky
bluesky-ai
1moA good, short thread about the next head of the hydra to pop up in all this. As an aside, a family member is convinced that this will be t…A good, short thread about the next head of the hydra to pop up in all this. As an aside, a family member is convinced that this will…
↗Not briefed
bluesky
bluesky-ai
1moModel Railroader had it right saying there's no point to showing AI images because then they're not actual models you can makeModel Railroader had it right saying there's no point to showing AI images because then they're not actual models you can make
↗Not briefed
hn
hn-ai
1moPartPilot Open-source component intelligence for hardware engineers—
↗Not briefed
hn
hn-ai
1moNovgraph: Knowledge Graph Engine for Codebases—
↗Not briefed
bluesky
bluesky-ai
1moChinese AI tools are being used to advance the Chinese government’s political narrative on Tibet with the deep integration of AI models int…Chinese AI tools are being used to advance the Chinese government’s political narrative on Tibet with the deep integration of AI model…
↗Not briefed
hn
hn-ai
1moShow HN: Smart Mouth Billy Bass – another fishy LLM assistantHello, HN people. This project was inspired by another very similar Show HN post from a few months ago called BillAI Bass. I decided t…
↗Not briefed
bluesky
bluesky-ai
1moI have bad news for you, Chuck. It already is www.bbc.co.uk/news/article...I have bad news for you, Chuck. It already is www.bbc.co.uk/news/article...
↗Not briefed
bluesky
bluesky-ai
1moOpenAIのHugging Face攻撃事件などで調査を担当しているAI監査組織「METR」とは何なのか? https://gigazine.net/news/20260917-model-evaluation-and-threat-research/🔧 Hugging FaceOpenAIのHugging Face攻撃事件などで調査を担当しているAI監査組織「METR」とは何なのか? https://gigazine.net/news/20260917-model-evaluation-and-threat-research/
↗Not briefed
rss
techcrunch-ai
1moRival AI agents, Instinct and Meta’s Muse, both add the ability to make callsPeople can use these assistants to make restaurant reservations and cancel subscriptions
↗Not briefed
arxiv
arxiv-ai
1moSchema-Anchored Latent Reasoning for Semantic Parsing-Based Knowledge Base Question AnsweringSemantic parsing (SP)-based knowledge base question answering aims to answer natural language questions by generating executable logic…
↗Not briefed
rss
techmeme
1moComp AI, which uses AI agents to draft security policies, collect audit evidence, conduct continuous compliance monitoring, and more, raised a $34M Series A (Dominic-Madori Davis/TechCrunch)Dominic-Madori Davis / TechCrunch: Comp AI, which uses AI agents to draft security policies, collect audit evidence, conduct continuou…
↗Not briefed
hn
hn-ai
1moPeople's houses are collapsing into the ocean. FEMA gives them no other option—
↗Not briefed
bluesky
bluesky-ai
1mojust endless sensationalized hyperbolic clickbait headlines candidly informing the public they're all about to die at the hands of auto-evo…just endless sensationalized hyperbolic clickbait headlines candidly informing the public they're all about to die at the hands of aut…
↗Not briefed
bluesky
bluesky-ai
1moGoogle, Nvidia and Anthropic want Emerald AI to find space on the grid for more data centersGoogle, Nvidia and Anthropic want Emerald AI to find space on the grid for more data centers
↗Not briefed
bluesky
bluesky-ai
1mo@alexhanna.bsky.social @emilymbender.bsky.social 2 artefacts for your show: Suleyman starting out sensible on consciousness, then using "t…@alexhanna.bsky.social @emilymbender.bsky.social 2 artefacts for your show: Suleyman starting out sensible on consciousness, then usi…
↗Not briefed
arxiv
arxiv-ai
1moModel-based Bootstrap for Offline Policy Evaluation in Tabular Reinforcement LearningOffline policy evaluation (OPE) is crucial in high-stakes reinforcement learning applications, where new policies must be assessed rel…
↗Not briefed
bluesky
bluesky-ai
1moIk weet nooit goed wat ik moet maken van de verhalen die uit OpenAI en Anthropic naar buiten komen. Los van de reële capaciteiten van de …Ik weet nooit goed wat ik moet maken van de verhalen die uit OpenAI en Anthropic naar buiten komen. Los van de reële capaciteiten va…
↗Not briefed
bluesky
bluesky-ai
1mo@hikikomorphism.bsky.social : tech-wise, methods of routing configuration and automation, firewalls, etc. non-tech wise: LLM nerding on Su…@hikikomorphism.bsky.social : tech-wise, methods of routing configuration and automation, firewalls, etc. non-tech wise: LLM nerding…
↗Not briefed
bluesky
bluesky-ai
1mosorry the llm killed your pc usage ability I hope it's recoverablesorry the llm killed your pc usage ability I hope it's recoverable
↗Not briefed
hn
hn-ai
1moSuppress vulnerabilities applying Kubernetes context to scans—
↗Not briefed
bluesky
bluesky-ai
1moAs long as the LLM is just a chat, the damage that "hallucination" can do is limited. But basically Anthropic and OpenAI gave these models …As long as the LLM is just a chat, the damage that "hallucination" can do is limited. But basically Anthropic and OpenAI gave these mo…
↗Not briefed
rss
techcrunch-ai
1moGoogle, Nvidia and Anthropic want Emerald AI to find space on the grid for more data centersA new coalition that includes Google, Nvidia, Anthropic and Emerald AI wants to find 100 GW of grid capacity for new data centers.
↗Not briefed
rss
the-decoder
1moAn OpenAI model kept slipping prompt injections into its own notes, and researchers still aren't sure whyOpenAI is publishing a framework for systematically reporting AI misalignment and launching it with six reports. In one case an unrele…
↗Not briefed
hn
hn-ai
1moShow HN: I built a simple couples app that explores a new way of connecting them🔧 WhisperHello HN.<p>I built my first serious mobile (OS - Android) application, Lune. Before I tell you about it, I wanna tell you the backsto…
↗Not briefed
bluesky
bluesky-ai
1moローカルLLM利用のクリニックが話題 院長自ら環境を構築 「医療機器より安い」は本当か https://www.itmedia.co.jp/aiplus/article/2609/17/2000001587/ローカルLLM利用のクリニックが話題 院長自ら環境を構築 「医療機器より安い」は本当か https://www.itmedia.co.jp/aiplus/article/2609/17/2000001587/
↗Not briefed
rss
techmeme
1moThe US SEC unveils a five-year "Innovation Exemption" to free platforms that facilitate blockchain and tokenized stock trading from many stock exchange rules (Hannah Lang/Reuters)Hannah Lang / Reuters: The US SEC unveils a five-year “Innovation Exemption” to free platforms that facilitate blockchain and tokenize…
↗Not briefed
bluesky
bluesky-ai
1moOkay, fine, I didn't want to, but let me say just this about the "AI warning": Specifically the Anthropic CEO said a few months ago AI comp…Okay, fine, I didn't want to, but let me say just this about the "AI warning": Specifically the Anthropic CEO said a few months ago AI…
↗Not briefed
bluesky
bluesky-ai
1moWe need to increase scientific literacy and understanding, not reduce it with bad LLM summaries. In my line of work the biggest challenge i…We need to increase scientific literacy and understanding, not reduce it with bad LLM summaries. In my line of work the biggest challe…
↗Not briefed
bluesky
bluesky-ai
1moKing Charles warns of 'existential danger' of AI falling into wrong handsKing Charles warns of 'existential danger' of AI falling into wrong hands
↗Not briefed
bluesky
bluesky-ai
1moit was funny that nobody in the tech press last week thought to ask this 27-year-old anthropic doomsday "whistleblower" if he voted for the…it was funny that nobody in the tech press last week thought to ask this 27-year-old anthropic doomsday "whistleblower" if he voted fo…
↗Not briefed
hn
hn-ai
1moShow HN: Proxy-benchmark – is it the proxy, the browser, or your machine?—
↗Not briefed
bluesky
bluesky-ai
1moalso the “umm ackshually AI is forced on you by being integrated into every technology now so you’re a hypocrite” argument is so dumb. you …🔧 ChatGPTalso the “umm ackshually AI is forced on you by being integrated into every technology now so you’re a hypocrite” argument is so dumb.…
↗Not briefed
bluesky
bluesky-ai
1moAnyway, I'm grateful to the Globe -- and to reporter Aidan Ryan -- for writing this story. And for getting the ball rolling on a timely con…Anyway, I'm grateful to the Globe -- and to reporter Aidan Ryan -- for writing this story. And for getting the ball rolling on a timel…
↗Not briefed
bluesky
bluesky-ai
1moAstra GPT6 was playing Minecraft on livestream. A “creeper” came and blew up its chest of valuables. The model then spent “the next sever…Astra GPT6 was playing Minecraft on livestream. A “creeper” came and blew up its chest of valuables. The model then spent “the next…
↗Not briefed
bluesky
bluesky-ai
1mo"If someone posts a photo on Instagram, they're doing it to be social on Instagram. They're not saying, 'hey, I'd love to train this AI mod…"If someone posts a photo on Instagram, they're doing it to be social on Instagram. They're not saying, 'hey, I'd love to train this A…
↗Not briefed
bluesky
bluesky-ai
1mo2) les LLM ne font pas que démultiplier les émissions de la tech, des acteurs majeurs de l'industrie de l'IA comme Mistral AI marchent main…2) les LLM ne font pas que démultiplier les émissions de la tech, des acteurs majeurs de l'industrie de l'IA comme Mistral AI marchent…
↗Not briefed
bluesky
bluesky-ai
1moOr maybe drop me an email if you have questions about one of my papers. I would gladly answer questions. Maybe, just maybe talking with co…Or maybe drop me an email if you have questions about one of my papers. I would gladly answer questions. Maybe, just maybe talking wi…
↗Not briefed
bluesky
bluesky-ai
1moYea it’s so SO obvious people with ideas like this have never actually once read a scientific paper or talked to anyone who does on a regul…Yea it’s so SO obvious people with ideas like this have never actually once read a scientific paper or talked to anyone who does on a…
↗Not briefed
bluesky
bluesky-ai
1moHaving a "newsroom" of AI "reporters" also raises lots of other questions. To start: why is everyone -- apparently -- white? upriseri.com…Having a "newsroom" of AI "reporters" also raises lots of other questions. To start: why is everyone -- apparently -- white? upriser…
↗Not briefed
hn
hn-ai
1moIt's time for cryptography to get its own abstraction layer—
↗Not briefed
bluesky
bluesky-ai
1moyou know what's worse than a paper? an open ended conversational UI backed by an LLMyou know what's worse than a paper? an open ended conversational UI backed by an LLM
↗Not briefed
hn
hn-ai
1moShow HN: Intent Based Infrastructure on Top of TerraformHey everyone! I built a platform that works on top of Terraform to expose high level constructs you want to create in your infrastruct…
↗Not briefed
bluesky
bluesky-ai
1moTelling a LLM connected to a coding harness to try and complete an ExploitGym problem isn't a bunch of autonomous agents deciding to do an …Telling a LLM connected to a coding harness to try and complete an ExploitGym problem isn't a bunch of autonomous agents deciding to d…
↗Not briefed