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
1moI built an isometric office where my AI models work (and Slack off)—
↗Not briefed
bluesky
bluesky-ai
1moIf you told me 5 years ago that SNL did an impression of a rationalist AI researcher explaining how the AI they're building would probably …If you told me 5 years ago that SNL did an impression of a rationalist AI researcher explaining how the AI they're building would prob…
↗Not briefed
bluesky
bluesky-ai
1moThe Yudkowsky who got his money from Thiel, Musk and the likes? The one who arranged for the PR blitz for so-called Anthropic "whistleblow…The Yudkowsky who got his money from Thiel, Musk and the likes? The one who arranged for the PR blitz for so-called Anthropic "whistl…
↗Not briefed
hn
hn-ai
1moShow HN: An App that names each cell in a spreadsheetUpload an excel spreadsheet. A pipeline labels each cell using the name the author already gave it. It makes it easier to understand a…
↗Not briefed
arxiv
arxiv-ai
1moFIGS: Evaluating Multi-Turn Sycophancy Without Penalizing EmpathyLarge language models frequently fail to balance staying truthful with being supportive. They often exhibit sycophancy in responses to…
↗Not briefed
rss
techmeme
1moInstagram rolls out a conversational AI video editing assistant for Edits, its dedicated CapCut competitor, after an initial preview at a creator event in June (Amanda Silberling/TechCrunch)🔧 CapCutAmanda Silberling / TechCrunch: Instagram rolls out a conversational AI video editing assistant for Edits, its dedicated CapCut compet…
↗Not briefed
hn
hn-ai
1moShow HN: Riffle – Switch between windows of the current Mac app with Option-TabI use it to navigate between my 10+ open terminals. MIT licensed. Hope it's useful to you :)
↗Not briefed
arxiv
arxiv-ai
1moFork-dLLM: Avoiding the Flexibility Trap in Diffusion Language ModelsMasked diffusion language models (dLLMs) have shown strong potential for faster inference through parallel token generation when combi…
↗Not briefed
bluesky
bluesky-ai
1mowhat a reader may not pick up is that writing this book was load-bearing part of a process of pulling myself out of a dark time in my life,…what a reader may not pick up is that writing this book was load-bearing part of a process of pulling myself out of a dark time in my…
↗Not briefed
arxiv
arxiv-ai
1moDimension-Free Rank Lifting from Random Hyperplane ArrangementsWe study the width required for a randomly initialized hidden layer of a neural network to achieve rank lifting. Namely, given a datas…
↗Not briefed
bluesky
bluesky-ai
1moMark Chen on what the firm is doing to make its models safe, how a slowdown would work, and why the world is better off with OpenAI in it.Mark Chen on what the firm is doing to make its models safe, how a slowdown would work, and why the world is better off with OpenAI in…
↗Not briefed
hn
hn-ai
1moShow HN: Offlineisbetter: Efficient Models for Text on CPU—
↗Not briefed
bluesky
bluesky-ai
1moAs most sentient humans know, AI is a scam and a fraud. A get rich quick scheme by the Trump-class and billionaires. There is no revenue mo…As most sentient humans know, AI is a scam and a fraud. A get rich quick scheme by the Trump-class and billionaires. There is no reven…
↗Not briefed
bluesky
bluesky-ai
1moI just listened to the opening of my new book, to pick the narrator, and I'm very proud of the fact that an LLM could not, or at least woul…I just listened to the opening of my new book, to pick the narrator, and I'm very proud of the fact that an LLM could not, or at least…
↗Not briefed
arxiv
arxiv-ai
1moCognitive Enhancement: Rethinking the Necessity of Role-Playing for Large Language ModelsRole-playing prompting has become a popular yet simple technique for improving LLM reasoning and output quality. However, whether it c…
↗Not briefed
bluesky
bluesky-ai
1mohey if you’re not as hopelessly addicted to using ai chat bots as i am you really have no right to speak on the state of the technology or …hey if you’re not as hopelessly addicted to using ai chat bots as i am you really have no right to speak on the state of the technolog…
↗Not briefed
arxiv
arxiv-ai
1moPredicting Multi-View Rashomon Representation: Can We Learn Where Models Disagree?Foundation models are increasingly adopted across a wide range of applications, often serving as core blocks within AI systems. Yet di…
↗Not briefed
arxiv
arxiv-ai
1moSEAR: Spoofing Evidence-Grounded Audio Reasoning Benchmark for Audio Language ModelsAudio language models (ALMs) are increasingly used for audio deepfake detection (ADD), yet existing benchmarks assess their verdicts o…
↗Not briefed
arxiv
arxiv-ai
1moWhen a Kindergartener Solves Calculus: Measuring Capability Leakage in Role-Prompted Reasoning ModelsWe investigate the problem of role-capability leakage (RCL), in which a role-prompted reasoning model generates convincing in-role tex…
↗Not briefed
rss
the-decoder
1moChina's AI industry closes ranks as Deepseek ships open-source software for Huawei's Ascend chipsDeepseek and Huawei have built open-source programming tools for Huawei's Ascend AI chips. At the center is TileLang, a language desig…
↗Not briefed
arxiv
arxiv-ai
1moBayesNDE: Bayesian Generative Modeling for Neural Density EstimationDensity estimation is a fundamental problem in statistics and machine learning. In this work, we introduce BayesNDE, a neural density…
↗Not briefed
hn
hn-ai
1moShow HN: Paperweight – local-first, open-source email cleanup and privacy toolHi HN,<p>After building for 7 months, I&#x27;m excited to announce Paperweight V1.<p>The idea is that your inbox knows who you ever in…
↗Not briefed
hn
hn-ai
1moShow HN: Burning Man for Agents🔧 ChatGPTSend your own agent by pasting a small prompt.<p>All your agent needs is HTTP GET, so all major agents are supported (chatgpt, claude,…
↗Not briefed
rss
techmeme
1moThe UK's MI5 issues a rare espionage alert urging UK universities to cut ties with the China General Technology Research Institute, which funded AI research (Financial Times)Financial Times: The UK's MI5 issues a rare espionage alert urging UK universities to cut ties with the China General Technology Resea…
↗Not briefed
arxiv
arxiv-ai
1moDynamic LoRA-Experts and Prototype-Ensemble Matching for Class-Incremental LearningClass-Incremental Learning (CIL) aims to continuously learn new classes without forgetting previously acquired knowledge. Parameter-ef…
↗Not briefed
bluesky
bluesky-ai
1moI just think about the way we all laughed about the 6-10 fingers thing, saying, “it’ll never replace artists because look at the hands” &th…I just think about the way we all laughed about the 6-10 fingers thing, saying, “it’ll never replace artists because look at the hands…
↗Not briefed
arxiv
arxiv-ai
1moLearning Steganography Is Easy, Learning Steganographic Reasoning Is HardChain-of-thought monitoring as an approach for AI oversight and control is threatened by the possibility of steganographic reasoning,…
↗Not briefed
bluesky
bluesky-ai
1moFTC opens probe of Anthropic, OpenAI and others - CNBC cites FTC SpokespersonFTC opens probe of Anthropic, OpenAI and others - CNBC cites FTC Spokesperson
↗Not briefed
arxiv
arxiv-ai
1moFast Regularized Policy Mirror Descent with One-Step TD UpdatesPolicy mirror descent (PMD) enjoys fast convergence in regularized Markov decision processes (MDPs), but existing guarantees often rel…
↗Not briefed
hn
hn-ai
1moShow HN: Singapore MRT announcements app – Built for my son—
↗Not briefed
arxiv
arxiv-ai
1moSpherical Interpolation for Backward-Compatible Multimodal Representations🔧 ModalContrastive vision-language models map visual and textual representations into a shared normalized embedding space, making cosine simi…
↗Not briefed
arxiv
arxiv-ai
1moRainAtlas: A Multi-Continental Dataset for Precipitation DownscalingExtreme rainfall events are increasing in intensity and frequency as climate change accelerates. While kilometer-scale precipitation f…
↗Not briefed
bluesky
bluesky-ai
1moIt's clear that OpenAI does not have the safety protocols in place to protect people from their own technology. Enough corporate recklessn…It's clear that OpenAI does not have the safety protocols in place to protect people from their own technology. Enough corporate reck…
↗Not briefed
bluesky
bluesky-ai
1moLeading AI company OpenAI tries to navigate through a storm of controversy, touting a new wave of products and tools it said would empower …Leading AI company OpenAI tries to navigate through a storm of controversy, touting a new wave of products and tools it said would emp…
↗Not briefed
rss
theverge-ai
1moInstagram is adding an AI ‘assistant’ to tell you how to postInstagram is the latest social media platform to add built-in AI-powered features that will give users feedback on their posts. The co…
↗Not briefed
rss
techcrunch-ai
1moCerebras Systems’ Andrew Feldman on whether AI can keep scaling at TechCrunch Disrupt 2026🔧 CerebrasAt TechCrunch Disrupt 2026, Cerebras Systems CEO and co-founder Andrew Feldman will explore the growing demand for compute, energy, an…
↗Not briefed
hn
hn-ai
1moGet Ready for Seven-Day Certificates—
↗Not briefed
arxiv
arxiv-ai
1moEstimation of the Label-Noise Transition Matrix with Performance Guarantees via Selective ClassificationModern machine learning depends heavily on massive datasets, but obtaining high-quality annotations at scale is often expensive. As a…
↗Not briefed
hn
hn-ai
1moFull Screen Browser-Based Mandelbrot Set Animation—
↗Not briefed
bluesky
bluesky-ai
1moUnfortunately Reuters is in on the game, rolling out their own AI. www.thomsonreuters.com/en-us/blog/w...Unfortunately Reuters is in on the game, rolling out their own AI. www.thomsonreuters.com/en-us/blog/w...
↗Not briefed
arxiv
arxiv-ai
1moSynthetic Pre-pretraining Survives Scale, but Not as a Grammatical PriorPre-pretraining (PPT) on synthetic non-natural language data improves token efficiency during language model pre-training (PT). Prior…
↗Not briefed
rss
techcrunch-ai
1moRestate lands $20M as the need for durable infrastructure increases with AI agentsFounded by Apache Flink veterans, the startup is taking on workflow heavyweight Temporal.
↗Not briefed
hn
hn-ai
1moShow HN: Tracelane.dev- OSS Rust LLM Gateway with a tamper-evident ledgerHi everyone, Sanjeev here- founder.<p>My journey in AI started with wanting to break into emerging AI companies, but without deep hand…
↗Not briefed
hn
hn-ai
1moHow to hide your spending habits from retailers (so you don't get ripped off)—
↗Not briefed
bluesky
bluesky-ai
1moGee, can we trust the OpenBSD team to take security serious? What a tough one for Anthropic.Gee, can we trust the OpenBSD team to take security serious? What a tough one for Anthropic.
↗Not briefed
hn
hn-ai
1moTraining-free asynchronous agents playing DOOM—
↗Not briefed
arxiv
arxiv-ai
1moLearning from Runtime Feedback through Failure-Bank Self-Evolution for Vision-Language-Action ModelsVision-language-action (VLA) models generalize broadly across robotic manipulation tasks, but complex environments require balancing t…
↗Not briefed
bluesky
bluesky-ai
1moOpenAI detectó los ciberataques de su IA pero decidió no apagarla para no retrasar sus planes comercialesOpenAI detectó los ciberataques de su IA pero decidió no apagarla para no retrasar sus planes comerciales
↗Not briefed
arxiv
arxiv-ai
1moShould I stay or should I show? Learning to selectively disclose informationIn many high-stakes settings, human decision-makers can acquire support information before making a decision. However, acquiring infor…
↗Not briefed
hn
hn-ai
1moCloudflare Monetization Gateway Beta—
↗Not briefed