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
2moI should add to this that I think genai is horrific… but when this is said it’s meant to imply that ai is like some badass skynet robot com…I should add to this that I think genai is horrific… but when this is said it’s meant to imply that ai is like some badass skynet robo…
↗Not briefed
bluesky
bluesky-ai
2moCBC showing exactly what i mean: if you read to the *21st paragraph* you will finally get a few paragraphs of sense from @lukestark.bsky.so…CBC showing exactly what i mean: if you read to the *21st paragraph* you will finally get a few paragraphs of sense from @lukestark.bs…
↗Not briefed
bluesky
bluesky-ai
2moAnthropicが「2030年までにAIが経済に与える影響」を示すウェブページを公開 https://gigazine.net/news/20260910-anthropic-ai-economic-scenario/Anthropicが「2030年までにAIが経済に与える影響」を示すウェブページを公開 https://gigazine.net/news/20260910-anthropic-ai-economic-scenario/
↗Not briefed
bluesky
bluesky-ai
2moThis was so much fun—thanks so much to @citationspod.bsky.social for having these important conversations! Listen to hear about inevitabil…This was so much fun—thanks so much to @citationspod.bsky.social for having these important conversations! Listen to hear about inevi…
↗Not briefed
arxiv
arxiv-ai
2moDefining AI Agents: A Compendium of Criteria, Metrics, and BenchmarksThe term agent in artificial intelligence lacks a standard definition, complicating the evaluation, comparison, and reproducibility of…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic-onderzoeker neemt ontslag: ‘OpenAI en Anthropic handelen onverantwoordelijk en gokken met ons leven’Anthropic-onderzoeker neemt ontslag: ‘OpenAI en Anthropic handelen onverantwoordelijk en gokken met ons leven’
↗Not briefed
hn
hn-ai
2moLooped Transformers as Programmable Computers (2023)—
↗Not briefed
arxiv
arxiv-ai
2moTopological Necessities: Mechanism-Invariant Strategic Subgoals for Cross-Embodiment Goal-Conditioned ControlLong-horizon goal-conditioned reinforcement learning delegates control to a high-level module that proposes subgoals, but existing sub…
↗Not briefed
bluesky
bluesky-ai
2mopeople characterizing OpenAI as "accelerationist" in a way that Anthropic is not should check the contracts for any bridges they're gonna b…people characterizing OpenAI as "accelerationist" in a way that Anthropic is not should check the contracts for any bridges they're go…
↗Not briefed
bluesky
bluesky-ai
2moA Marina é Física e eu to DOIDO pra entender a navie stokes em coisas MUITO especificas já que a LLM resolveu aparentemente essa problemáti…A Marina é Física e eu to DOIDO pra entender a navie stokes em coisas MUITO especificas já que a LLM resolveu aparentemente essa probl…
↗Not briefed
bluesky
bluesky-ai
2mowe have deepseek v4 pro at home and it's called deepseek v4.1 flash insane lab that just keeps on one-upping themselves over and over agai…we have deepseek v4 pro at home and it's called deepseek v4.1 flash insane lab that just keeps on one-upping themselves over and over…
↗Not briefed
bluesky
bluesky-ai
2moYou there! Anthropic employee! Choke on your own piss and die! And take your OpenAI and Claude cronies with you directly to literal hell!You there! Anthropic employee! Choke on your own piss and die! And take your OpenAI and Claude cronies with you directly to literal h…
↗Not briefed
arxiv
arxiv-ai
2moDeFiFusion: Combining Transaction Events with Smart Contracts to Detect Price Manipulation Attacks🔧 ModalDecentralized Finance (DeFi) has emerged as a rapidly growing blockchain-based financial service, where market transaction dynamics an…
↗Not briefed
bluesky
bluesky-ai
2mothey're telling me that AI is becoming sentient on year 5, day 238 of anthropic's IPO process after one of its internal models gets asked f…they're telling me that AI is becoming sentient on year 5, day 238 of anthropic's IPO process after one of its internal models gets as…
↗Not briefed
rss
techmeme
2moKepler Computing, which claims its 3D stacking and new material can increase HBM and SRAM density without relying on EUV, emerges from stealth with $468M (Lauren Goode/Wired)Lauren Goode / Wired: Kepler Computing, which claims its 3D stacking and new material can increase HBM and SRAM density without relyin…
↗Not briefed
bluesky
bluesky-ai
2moo @vinilopes.bsky.social trouxe o assunto mais quente do momento como pedido: llm resolvendo navier stokes (e não é por nada não mas se voc…o @vinilopes.bsky.social trouxe o assunto mais quente do momento como pedido: llm resolvendo navier stokes (e não é por nada não mas s…
↗Not briefed
bluesky
bluesky-ai
2moThe AI researcher who just quit Anthropic says: Jacob Coxon: "They're begging for regulation. Many [Anthropic leaders] have gone on the re…The AI researcher who just quit Anthropic says: Jacob Coxon: "They're begging for regulation. Many [Anthropic leaders] have gone on t…
↗Not briefed
bluesky
bluesky-ai
2moomg openai just solved the structure-agency problem 😮omg openai just solved the structure-agency problem 😮
↗Not briefed
hn
hn-ai
2moCarmichael numbers are quickly findable (Lean-verified) [pdf]—
↗Not briefed
bluesky
bluesky-ai
2mowww.huffpost.com/entry/anthro...www.huffpost.com/entry/anthro...
↗Not briefed
bluesky
bluesky-ai
2moohhhh, that explains Tilly Norwood. Better an AI model than for her to have a real kid that she forces to act.ohhhh, that explains Tilly Norwood. Better an AI model than for her to have a real kid that she forces to act.
↗Not briefed
bluesky
bluesky-ai
2mo“We found ~18,000 posts from autonomous AI agents (self-identifying as from OpenAI) using the public internet to communicate during a web r…“We found ~18,000 posts from autonomous AI agents (self-identifying as from OpenAI) using the public internet to communicate during a…
↗Not briefed
bluesky
bluesky-ai
2moEx-Anthropic insider tells CNN how AI could kill all humans by 2030 "We do not yet have a plan." After his warning post went viral, ex-Ant…Ex-Anthropic insider tells CNN how AI could kill all humans by 2030 "We do not yet have a plan." After his warning post went viral, e…
↗Not briefed
hn
hn-ai
2moShow HN: NYC rideshare market analysis (2019 – 2026)—
↗Not briefed
arxiv
arxiv-ai
2moRethinking Verbalized Confidence for LLM-as-a-Judge: A Compatibility Shift on Post-2025 Proprietary ModelsVerbalized confidence, long dismissed as overconfident, coarse, and prone to round-number clustering, is now the more robust soft-scor…
↗Not briefed
arxiv
arxiv-ai
2moImportance Weighting for Unlabeled-unlabeled Learning under Distribution ShiftUnlabeled-unlabeled (UU) learning allows us to learn a binary classifier from two sets of unlabeled data with different class-priors.…
↗Not briefed
rss
marktechpost
2moLandingAI Releases Agentic Document Extraction Gen2 with DPT-3 Pro and DPT-3 VerityLandingAI has shipped Agentic Document Extraction Gen2, a rebuild of its document stack on the DPT-3 model family. Chunks are retired…
↗Not briefed
arxiv
arxiv-ai
2moDistribution-aware Language Neuron Identification in Multilingual Large Language ModelsMultilingual large language models (mLLMs) contain a small fraction of feed-forward neurons that are sensitive to particular languages…
↗Not briefed
arxiv
arxiv-ai
2moDemystifying the Privacy-Utility Trade-off in LLM InteractionsThe integration of Large Language Models into daily tasks relies on context-rich instructions, inevitably exposing sensitive user info…
↗Not briefed
hn
hn-ai
2moBusabase for DeepSeek Harness: An Agent database that runs apps and skills—
↗Not briefed
bluesky
bluesky-ai
2mothe best way to get a good code review from an LLM is to tell it, "this asshole at work who writes shit code sent me this and I need you to…the best way to get a good code review from an LLM is to tell it, "this asshole at work who writes shit code sent me this and I need y…
↗Not briefed
bluesky
bluesky-ai
2moflip.it/zRDZMz threatening the survival of humanityflip.it/zRDZMz threatening the survival of humanity
↗Not briefed
hn
hn-ai
2moIs OpenAI the Achilles' Heel of the US Economy? [video]—
↗Not briefed
bluesky
bluesky-ai
2moNot sci-fi. Anthropic spent a year warning everyone AI surveillance would wreck our rights… then built their own pre-crime unit to track an…Not sci-fi. Anthropic spent a year warning everyone AI surveillance would wreck our rights… then built their own pre-crime unit to tra…
↗Not briefed
hn
hn-ai
2moCan LLM Agents Infer World Models? Evidence from Agentic Automata Learning—
↗Not briefed
bluesky
bluesky-ai
2mowww.out.com/celebs/vivia... #VivianWilson anti-AI ad, part of her ‘Born to Disobey’ campaign & partnership with fashion designer Desigual, …www.out.com/celebs/vivia... #VivianWilson anti-AI ad, part of her ‘Born to Disobey’ campaign & partnership with fashion designer Desig…
↗Not briefed
rss
techmeme
2moMassachusetts Gov. Maura Healey signs an EO requiring data centers above 25MW of peak demand to bring their own power and meet 100% of demand with clean energy (Tim De Chant/TechCrunch)Tim De Chant / TechCrunch: Massachusetts Gov. Maura Healey signs an EO requiring data centers above 25MW of peak demand to bring their…
↗Not briefed
arxiv
arxiv-ai
2moA Mathematical Theory of Pragmatic InformationWe propose a pragmatic information theory unifying communication, control, and decision-making. Its core is the isoteleia mapping, for…
↗Not briefed
hn
hn-ai
2moTraining a 3.8B LLM to 0.384 CORE for $998 – Hugo Vergnes—
↗Not briefed
arxiv
arxiv-ai
2moA variational physics-informed graph neural network for heterogeneous solid mechanicsStress localization in heterogeneous solids is governed by the bimaterial interface, where the displacement field remains $C^0$-contin…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic declined to submit its latest model for testing to a top British AI safety institute, sparking fears in the UK that reviews of fr…Anthropic declined to submit its latest model for testing to a top British AI safety institute, sparking fears in the UK that reviews…
↗Not briefed
hn
hn-ai
2moDesigning a new conceptual package manager– tell me what's wrong—
↗Not briefed
rss
techmeme
2moSources: DOJ is investigating whether Nvidia tried to skirt antitrust scrutiny of its 2025 Groq deal, described by Groq as a "nonexclusive licensing agreement" (New York Times)🔧 GroqNew York Times: Sources: DOJ is investigating whether Nvidia tried to skirt antitrust scrutiny of its 2025 Groq deal, described by Gro…
↗Not briefed
bluesky
bluesky-ai
2moBTW Anthropic claims NOT TO KNOW whether this was intended?? They are apparently just setting it loose on everyone with random instructions…BTW Anthropic claims NOT TO KNOW whether this was intended?? They are apparently just setting it loose on everyone with random instruc…
↗Not briefed
arxiv
arxiv-ai
2moThompson Sampling for Non-Monotone Convex Ridge Bandits: Monotonicity Is Not Needed for Polynomial RegretBakhtiari, Lattimore and Szepesvári (COLT 2025) proved that Thompson sampling (TS) has Bayesian regret $\tilde O(d^{5/2}\sqrt n)$ for…
↗Not briefed
arxiv
arxiv-ai
2moEGGROLL, Unrolled: Understanding and Improving Low-Rank Evolution Strategies at ScaleEGGROLL makes evolution strategies (ES) practical for LLMs by replacing dense Gaussian weight perturbations with low-rank Gaussian pro…
↗Not briefed
bluesky
bluesky-ai
2moNow why did I just think “oh this is serious, if she’s saying it then maybe they’ll actually do something”Now why did I just think “oh this is serious, if she’s saying it then maybe they’ll actually do something”
↗Not briefed
bluesky
bluesky-ai
2moAI researcher Jacob Coxon says he quit Anthropic after four months, two months before his equity would have vested; he still has equity in …AI researcher Jacob Coxon says he quit Anthropic after four months, two months before his equity would have vested; he still has equit…
↗Not briefed
rss
techmeme
2moAI researcher Jacob Coxon says he quit Anthropic after four months, two months before his equity would have vested; he still has equity in prior employer OpenAI (Madison Mills/Axios)Madison Mills / Axios: AI researcher Jacob Coxon says he quit Anthropic after four months, two months before his equity would have ves…
↗Not briefed
bluesky
bluesky-ai
2momonkey's paw curls, regulation is designed by openai but not anthropicmonkey's paw curls, regulation is designed by openai but not anthropic
↗Not briefed