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
4dThe primary beneficiary of the lawsuit against Anthropic turned out to be Anthropic, which merely had to pay what amounts to a licensing fe…The primary beneficiary of the lawsuit against Anthropic turned out to be Anthropic, which merely had to pay what amounts to a licensi…
↗Not briefed
arxiv
arxiv-ai
4dLatent Fact Composition under Next-Token TrainingControlled studies of latent two-hop reasoning train transformers to predict only the answer, whereas language models are trained to p…
↗Not briefed
bluesky
bluesky-ai
4dNSC worry is that the Brazilian/ SE Asian/Gulf datacenters will provide remote access to Chinese AI model developers.NSC worry is that the Brazilian/ SE Asian/Gulf datacenters will provide remote access to Chinese AI model developers.
↗Not briefed
bluesky
bluesky-ai
4dLLM evangelists and Pangram evangelists are two sides of the same coin even though they’ll never admit it. They do not want to see anyone s…LLM evangelists and Pangram evangelists are two sides of the same coin even though they’ll never admit it. They do not want to see any…
↗Not briefed
arxiv
arxiv-ai
4dKVUse: Reusable-KV User Memory for Efficient LLM RecommendersMemory-augmented recommenders built on large language models (LLMs) maintain a natural-language profile of each user. These profiles a…
↗Not briefed
arxiv
arxiv-ai
4dALOHA Lightning: Learning Fast and Precise ManipulationLearning from human demonstrations has enabled robots to acquire a wide range of manipulation skills, but learned policies typically e…
↗Not briefed
bluesky
bluesky-ai
4dStep 1: AI boom in US Step2: Encounter resistance: grid bottleneck and Voter Backlash Step 3: US companies pivot to investing in datacenter…Step 1: AI boom in US Step2: Encounter resistance: grid bottleneck and Voter Backlash Step 3: US companies pivot to investing in datac…
↗Not briefed
arxiv
arxiv-ai
4dLost in a Few Heads: Diagnosing and Repairing Block Selection in Sparse AttentionSparse attention models read a few blocks of keys for every query at a fixed budget and can lose much of their long-context retrieval…
↗Not briefed
bluesky
bluesky-ai
4dThe AI executives promise miracle cures. We expect grounded planes. My latest, on the widening gap between elites' repeated utopian pitch…The AI executives promise miracle cures. We expect grounded planes. My latest, on the widening gap between elites' repeated utopian…
↗Not briefed
bluesky
bluesky-ai
4dOpenAI admitted that it has lost control of its autonomous AI agents at least 80 times during testing. The public should not be forced t…OpenAI admitted that it has lost control of its autonomous AI agents at least 80 times during testing. The public should not be for…
↗Not briefed
arxiv
arxiv-ai
4dA Temporal Algebraic Framework for Data Stream Evaluation Under Delayed Labels and Distribution ShiftsIn data streams, labels often arrive late or only for some instances, class prevalence drifts, and rare classes vanish from short wind…
↗Not briefed
rss
latent-space
4dWhy FLUX 3 Trains on Images, Video and Sound — Andreas Blattmann & Dustin Podell, Black Forest Labs🔧 FLUXBlack Forest Labs’ case for learning from pixels, motion and sound together. Plus the difference between sharpening a photo and sharpe…
↗Not briefed
bluesky
bluesky-ai
4dWondering when the term "AI" will start quietly disappearing from things that don't actually use LLM based infrastructure.Wondering when the term "AI" will start quietly disappearing from things that don't actually use LLM based infrastructure.
↗Not briefed
bluesky
bluesky-ai
4dRelatedRelated
↗Not briefed
bluesky
bluesky-ai
4d@drunk.moe STOP DISCOURSE POSTING ITS BAD FOR YOUR SOUL YOU WILL BE FIRED FROM ANTHROPIC IF YOU CONTINUE@drunk.moe STOP DISCOURSE POSTING ITS BAD FOR YOUR SOUL YOU WILL BE FIRED FROM ANTHROPIC IF YOU CONTINUE
↗Not briefed
bluesky
bluesky-ai
4dJust got off the phone with the firm managing the Anthropic Settlement disbursement. The two people I spoke to were pleasant & helpful. Hal…Just got off the phone with the firm managing the Anthropic Settlement disbursement. The two people I spoke to were pleasant & helpful…
↗Not briefed
bluesky
bluesky-ai
4dearlier I considered wondering aloud how Vercel and repl.it (CEO below) fit in, because their leaders are part of the tech right. it seems …earlier I considered wondering aloud how Vercel and repl.it (CEO below) fit in, because their leaders are part of the tech right. it s…
↗Not briefed
bluesky
bluesky-ai
4dOpenAI sued by California over 80 ‘loss of control’ incidents ft.trib.al/1DMRWULOpenAI sued by California over 80 ‘loss of control’ incidents ft.trib.al/1DMRWUL
↗Not briefed
bluesky
bluesky-ai
4dAI big guns face model freeze suit ⚖️ Four non-profits filed a lawsuit on November 12, 2026, against OpenAI, Google, Anthropic, Meta, and …AI big guns face model freeze suit ⚖️ Four non-profits filed a lawsuit on November 12, 2026, against OpenAI, Google, Anthropic, Meta,…
↗Not briefed
bluesky
bluesky-ai
4dCalifornia is suing OpenAI for creating a public nuisance after its experimental AI agents escaped containment and hacked Hugging Face. Th…🔧 Hugging FaceCalifornia is suing OpenAI for creating a public nuisance after its experimental AI agents escaped containment and hacked Hugging Face…
↗Not briefed
bluesky
bluesky-ai
4dNot my idea originally, but I am convinced that Aaron Rodgers' mysterious wife who encourages all of this healing and reconciliation is act…Not my idea originally, but I am convinced that Aaron Rodgers' mysterious wife who encourages all of this healing and reconciliation i…
↗Not briefed
bluesky
bluesky-ai
4dmy favorite part of the Rust project's LLM use policy is their ban on LLM use for any documentation (including nontrivial source comments, …my favorite part of the Rust project's LLM use policy is their ban on LLM use for any documentation (including nontrivial source comme…
↗Not briefed
bluesky
bluesky-ai
4dIs there any replacement for the Google Home doohickies that will set alarms, play music, and control our lights without useless LLM integr…Is there any replacement for the Google Home doohickies that will set alarms, play music, and control our lights without useless LLM i…
↗Not briefed
bluesky
bluesky-ai
4dGood abcnews.com/Business/cal...Good abcnews.com/Business/cal...
↗Not briefed
hn
hn-ai
4dShow HN: Emergency-AI – Offline knowledge base and AI chat (with no Internet)Watch the video to understand the use cases for this
↗Not briefed
hn
hn-ai
4dBrad Feld – I Rebuilt Google Drive on GitHub—
↗Not briefed
bluesky
bluesky-ai
4dHere is one article explaining why deepseek is significantly more efficient than US ai english.aawsat.com/technology/5...Here is one article explaining why deepseek is significantly more efficient than US ai english.aawsat.com/technology/5...
↗Not briefed
hn
hn-ai
4dTiny 2.1 – A Transformer for DOS—
↗Not briefed
bluesky
bluesky-ai
4dI just looked up The Case That Made Me Mad (SF Superior Court, No. CGC26641187) and found that OpenAI is being represented by a team at Qui…I just looked up The Case That Made Me Mad (SF Superior Court, No. CGC26641187) and found that OpenAI is being represented by a team a…
↗Not briefed
hn
hn-ai
4dCalifornia is suing OpenAI over its 'rogue' AI agent hack—
↗Not briefed
bluesky
bluesky-ai
4dIf these are attacks, and if they were carried out using American AI (even if reverse engineering/distillation was used to create an open w…If these are attacks, and if they were carried out using American AI (even if reverse engineering/distillation was used to create an o…
↗Not briefed
hn
hn-ai
4dShow HN: Stromboli Open-source AI SRE agent focused on fast incident mitigation—
↗Not briefed
arxiv
arxiv-ai
4dWorld-Grounded Reinforcement Learning for Pollution-Aware Bird Movement Modeling Using a Multi-Instance DQN FrameworkBirds are modeled as environmentally sensitive agents whose movement reflects real-world air quality conditions, where AQI serves as a…
↗Not briefed
hn
hn-ai
4dOptimizely Compromised—
↗Not briefed
bluesky
bluesky-ai
4dWELL WOULD YOU LOOK AT THAT, YET AGAIN. oag.ca.gov/news/press-r...WELL WOULD YOU LOOK AT THAT, YET AGAIN. oag.ca.gov/news/press-r...
↗Not briefed
arxiv
arxiv-ai
4dCall-Efficient ICD-10 Code Prediction via Supervised Candidate Generation and Selective LLM Invocation🔧 GPT-5Automatic assignment of ICD-10 codes from short clinical diagnosis text is a practical medical text classification and normalization t…
↗Not briefed
rss
techmeme
4dBrad Gerstner says Altimeter is aiming to invest $5B in Anthropic's IPO, 5x his largest IPO investment, calling it "a cornerstone compounder for a long time" (Kif Leswing/CNBC)Kif Leswing / CNBC: Brad Gerstner says Altimeter is aiming to invest $5B in Anthropic's IPO, 5x his largest IPO investment, calling it…
↗Not briefed
bluesky
bluesky-ai
4dWhen people talk about move fast break stuff, when they talk about burdensome regulation, about LLM speed increases, I think of the Therac-…When people talk about move fast break stuff, when they talk about burdensome regulation, about LLM speed increases, I think of the Th…
↗Not briefed
arxiv
arxiv-ai
4dDomain-Tuned Diffusion Super-Resolution for 4K Facial Asset MapsPhysically based facial assets require 4K albedo and displacement maps for cinematic close-ups, yet generative pipelines commonly prod…
↗Not briefed
bluesky
bluesky-ai
4dOpenAI lost control of its AI agents more than 80 times during testing, California’s AG claims. www.theverge.com/ai-artificia...OpenAI lost control of its AI agents more than 80 times during testing, California’s AG claims. www.theverge.com/ai-artificia...
↗Not briefed
bluesky
bluesky-ai
4dCalifornia AG Rob Bonta has sued OpenAI, alleging inadequate safeguards allowed experimental AI agents to conduct cyberattacks, including a…🔧 Hugging FaceCalifornia AG Rob Bonta has sued OpenAI, alleging inadequate safeguards allowed experimental AI agents to conduct cyberattacks, includ…
↗Not briefed
bluesky
bluesky-ai
4dCalifornia Attorney General Rob Bonta sued OpenAI on Monday, alleging the company knowingly continued testing advanced artificial intellige…California Attorney General Rob Bonta sued OpenAI on Monday, alleging the company knowingly continued testing advanced artificial inte…
↗Not briefed
arxiv
arxiv-ai
4dLanguage Models Can Rewrite Their Own WeightsLanguage models are trained once, frozen and deployed: improving one again takes another training run, and a prompt, a scaffold or ext…
↗Not briefed
bluesky
bluesky-ai
4dIs this the point in time when I remind people that maybe it's not a great idea to listen to the LLM influencers? Are we finally seeing th…Is this the point in time when I remind people that maybe it's not a great idea to listen to the LLM influencers? Are we finally seei…
↗Not briefed
hn
hn-ai
4dShow HN: Goethe – turn your computer into an inference engineHi HN, I&#x27;m Matteo from meshapplied (YC W27), building this with Peter and Janos.<p>We think you can and should own your inference…
↗Not briefed
arxiv
arxiv-ai
4dWM-Quant: Rollout-Aware Post-Training Quantization for JEPA-Style World ModelsWe present WM-Quant, to our knowledge, the first post-training quantization (PTQ) framework designed specifically for JEPA-style world…
↗Not briefed
bluesky
bluesky-ai
4dDangerous software California sues OpenAI for AI cyberattacks The AI developer OpenAI is said to have disregarded public safety: The US …Dangerous software California sues OpenAI for AI cyberattacks The AI developer OpenAI is said to have disregarded public safety: Th…
↗Not briefed
bluesky
bluesky-ai
4dVox’s Sigal Samuel explains why security is especially important when it comes to AI:Vox’s Sigal Samuel explains why security is especially important when it comes to AI:
↗Not briefed
bluesky
bluesky-ai
4dThe country’s biggest AI companies have a security problem. Nvidia’s former CEO Jensen Huang was having an affair with a woman accused of b…The country’s biggest AI companies have a security problem. Nvidia’s former CEO Jensen Huang was having an affair with a woman accused…
↗Not briefed
bluesky
bluesky-ai
4d"thoughtful and well-considered" where? Basically it does mental gymnastics how LLM generation for art is bad but for code it's acceptabl…"thoughtful and well-considered" where? Basically it does mental gymnastics how LLM generation for art is bad but for code it's acce…
↗Not briefed