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
2moThe weirdest thing about OpenAI use these days for me: I think quietly one of their more impressive accomplishments is Luna. At least in m…The weirdest thing about OpenAI use these days for me: I think quietly one of their more impressive accomplishments is Luna. At least…
↗Not briefed
arxiv
arxiv-ai
2moPermutation-Based Stegomalware in Large Language Models: Threats and CountermeasuresThe difficulty of training large language models (LLMs), together with their ubiquity, raises the threat of stegomalware, where malici…
↗Not briefed
bluesky
bluesky-ai
2moNo coercion necessary. People are just freely handing over sensitive financial and health records.No coercion necessary. People are just freely handing over sensitive financial and health records.
↗Not briefed
bluesky
bluesky-ai
2moPeritus should make cheap models effective. We have people running deepseek tests right now.Peritus should make cheap models effective. We have people running deepseek tests right now.
↗Not briefed
bluesky
bluesky-ai
2moThere is absolutely nothing to support that any of these scientists were "trying to weaponize pathogens". While I don't have access to the…There is absolutely nothing to support that any of these scientists were "trying to weaponize pathogens". While I don't have access t…
↗Not briefed
bluesky
bluesky-ai
2mo"OpenAI is hiring hundreds of contractors who read a massive stream of real users’ ChatGPT prompts, with the prompts sometimes including se…🔧 ChatGPT"OpenAI is hiring hundreds of contractors who read a massive stream of real users’ ChatGPT prompts, with the prompts sometimes includi…
↗Not briefed
bluesky
bluesky-ai
2moA Megalomanical Egotistical Narcissist with Antisocial Personality Disorder who is utterly ignorant in charge of #AI... No, thank you very …A Megalomanical Egotistical Narcissist with Antisocial Personality Disorder who is utterly ignorant in charge of #AI... No, thank you…
↗Not briefed
arxiv
arxiv-ai
2moAI-Driven Feedback Systems, Digital Labour, and Silent Quitting: Transforming African WorkplacesThe current trend of digitalisation has revolutionised the organisation of work and the way it is measured and performed across the gl…
↗Not briefed
bluesky
bluesky-ai
2mopeople don't even know about harnesses! the number of arguments i've had with people who think "AI" in 2026 is the raw output of a feed-for…people don't even know about harnesses! the number of arguments i've had with people who think "AI" in 2026 is the raw output of a fee…
↗Not briefed
bluesky
bluesky-ai
2moLet’s ban warp speed and teleportation too, while we’re banning sci-fi concepts. To be clear, I am not dismissing cybersecurity threats fr…Let’s ban warp speed and teleportation too, while we’re banning sci-fi concepts. To be clear, I am not dismissing cybersecurity threa…
↗Not briefed
bluesky
bluesky-ai
2moOh wellOh well
↗Not briefed
bluesky
bluesky-ai
2moWas steckt hinter der Debatte über eine KI-Bremse? #FAQ #KünstlicheIntelligenz #OpenAIWas steckt hinter der Debatte über eine KI-Bremse? #FAQ #KünstlicheIntelligenz #OpenAI
↗Not briefed
bluesky
bluesky-ai
2mo"De-identified"?? Lol. Given the sheer amount of context in these conversations, if you are at all a politically sensitive person there is …"De-identified"?? Lol. Given the sheer amount of context in these conversations, if you are at all a politically sensitive person ther…
↗Not briefed
hn
hn-ai
2moChina's Top Spy Chief Warns A.I. Is a Threat to Party Rule—
↗Not briefed
hn
hn-ai
2moUnison: A friendly programming language from the future—
↗Not briefed
bluesky
bluesky-ai
2moThey unleashed something they can't control anymore. AI getting more powerful 'by the day' and kill switch should be required, tech boss t…They unleashed something they can't control anymore. AI getting more powerful 'by the day' and kill switch should be required, tech b…
↗Not briefed
arxiv
arxiv-ai
2moAnatomy of Associative Recall in Fixed-State Recurrences: A Matched-State Decomposition, an Interference Wall, and a Curriculum That Breaks ItFixed-state recurrences--linear attention and state-space models--are reported to lag behind attention on associative recall, but whol…
↗Not briefed
bluesky
bluesky-ai
2moTrump attacks ‘sick conspiracy’ against AI as tech stocks slide. President hits out after Anthropic, OpenAI and SpaceX leaders back call to…Trump attacks ‘sick conspiracy’ against AI as tech stocks slide. President hits out after Anthropic, OpenAI and SpaceX leaders back ca…
↗Not briefed
bluesky
bluesky-ai
2moSo this is completely weird - I had assumed that this was just some LLM gurning, not that I had in fact "written" a whole bunch of "books."So this is completely weird - I had assumed that this was just some LLM gurning, not that I had in fact "written" a whole bunch of "bo…
↗Not briefed
bluesky
bluesky-ai
2moinstead you were treated with some baseless Skynet type hysteria by 20-something tech employees high on their own farts, followed by this w…instead you were treated with some baseless Skynet type hysteria by 20-something tech employees high on their own farts, followed by t…
↗Not briefed
hn
hn-ai
2moShow HN: SayItFlow – On-device Mac dictation that unloads models when idle—
↗Not briefed
bluesky
bluesky-ai
2moyoull find that they pay anthropic quite wellyoull find that they pay anthropic quite well
↗Not briefed
hn
hn-ai
2moOur first impressions of the GPT-Live-1 API for voice agents—
↗Not briefed
bluesky
bluesky-ai
2moAbbey Games Godhood, Renowned Explorers, Reus LLM use and Coding -y abbeygames.comAbbey Games Godhood, Renowned Explorers, Reus LLM use and Coding -y abbeygames.com
↗Not briefed
rss
techmeme
2moNvidia announces the RTX Pro 5500 Blackwell Workstation Edition, offering comparable specs to the RTX 5090 but with 84GB of GDDR7 memory, vs. RTX 5090's 32GB (Zhiye Liu/Tom's Hardware)Zhiye Liu / Tom's Hardware: Nvidia announces the RTX Pro 5500 Blackwell Workstation Edition, offering comparable specs to the RTX 5090…
↗Not briefed
bluesky
bluesky-ai
2motried it with Sol, Opus, Fable, Astra, just the diff of the commit and this prompt (this is turning into DipshitBench). all easily understa…tried it with Sol, Opus, Fable, Astra, just the diff of the commit and this prompt (this is turning into DipshitBench). all easily und…
↗Not briefed
arxiv
arxiv-ai
2moSkeletal Prototypes on Iterative Nerve ExpansionsPrototype reduction replaces a training set with a smaller representation, and the established methods return a finite set of points.…
↗Not briefed
bluesky
bluesky-ai
2moLooked up some book editions by Algernon Blackwood and came across this gem of an inappropriate cover. 😆Looked up some book editions by Algernon Blackwood and came across this gem of an inappropriate cover. 😆
↗Not briefed
hn
hn-ai
2moGoogle tests paying publishers for using its content in AI Mode and Gemini—
↗Not briefed
arxiv
arxiv-ai
2moGPEvac: GNN-Based PPO for Adaptive Evacuation Routing During Shooting EventsThe sharp increase in mass shootings underscores an urgent need for systems that guide victims to safety in real time. An effective ev…
↗Not briefed
arxiv
arxiv-ai
2moLLM Inference in a Flash!Large Language Models (LLMs) have shown impressive capabilities across a range of natural language processing tasks, and LLM inference…
↗Not briefed
hn
hn-ai
2moLibation: Liberate Your Library—
↗Not briefed
hn
hn-ai
2moBackprop Alternative: Augmented Lagrangian Predictive Coding—
↗Not briefed
arxiv
arxiv-ai
2moComputer-assisted global regularity across nonlinear families of three-dimensional periodic Navier-Stokes flowsNumerical simulations reveal how vortices stretch and transfer energy, but establishing smooth evolution requires bounds that remain v…
↗Not briefed
bluesky
bluesky-ai
2mo… but can’t get them because the TA subsidies dry up as management takes those funds and gives them to OpenAI, Microsoft, Google, Meta and …… but can’t get them because the TA subsidies dry up as management takes those funds and gives them to OpenAI, Microsoft, Google, Meta…
↗Not briefed
hn
hn-ai
2momacOS 27 Golden Gate Is Now Available for All Users—
↗Not briefed
bluesky
bluesky-ai
2moNach mehreren autonomen Hackerangriffen durch KI-Modelle fordert Dario Amodei, der CEO von Anthropic, eine langsamere Entwicklung auf dem K…Nach mehreren autonomen Hackerangriffen durch KI-Modelle fordert Dario Amodei, der CEO von Anthropic, eine langsamere Entwicklung auf…
↗Not briefed
arxiv
arxiv-ai
2moLLMs as Master Forgers: Generating Synthetic Time Series Data for ManufacturingThis paper presents a novel framework leveraging Large Language Models (LLMs) to generate synthetic time series data for manufacturing…
↗Not briefed
bluesky
bluesky-ai
2moDario Amodei (Anthropic), Sam Altman (OpenAI), Elon Musk (xAI) y Demis Hassabis (Google) han acordado que es el momento de fingir preocupac…Dario Amodei (Anthropic), Sam Altman (OpenAI), Elon Musk (xAI) y Demis Hassabis (Google) han acordado que es el momento de fingir preo…
↗Not briefed
arxiv
arxiv-ai
2moSafe Error Correction for Language Models: Frozen-Base Adjustment with Capability PreservationWe study a practical question: can a small correction module fix errors in a frozen language model's outputs without degrading its bas…
↗Not briefed
arxiv
arxiv-ai
2moOptimal Pruning for Neural Architectures using Fisher Information DistancesA new scheme for parameter pruning is introduced, derived from the differential-geometric distance in model space. Pruning a parameter…
↗Not briefed
arxiv
arxiv-ai
2moBellman Policy OptimizationReinforcement learning with verifiable rewards (RLVR) improves the reasoning capabilities of large language models (LLMs). We introduc…
↗Not briefed
bluesky
bluesky-ai
2moThis article makes two important points: 1. People using ChatGPT expose sensitive data because they think it is private 2. OpenAI relies …🔧 ChatGPTThis article makes two important points: 1. People using ChatGPT expose sensitive data because they think it is private 2. OpenAI re…
↗Not briefed
arxiv
arxiv-ai
2moStellar Colosseum: A Many-Agent Harness for Long-Horizon Research in Mathematics and Theoretical Computer ScienceLanguage models can produce plausible short proofs, but may still be unreliable on long-horizon research problems, where progress depe…
↗Not briefed
arxiv
arxiv-ai
2moThe Router Within: Eliciting Native Skill Routing from a Frozen LLMSkills extend an LLM agent beyond its parametric knowledge, and the gain they promise rests on picking the right one. Deployed harness…
↗Not briefed
arxiv
arxiv-ai
2moA Chosen Future Can Still Be Rewritten: Causal Writability in Video ModelsWhen a video model generates physically incorrect motion, did it fail to learn the correct motion, or did it learn it but fail to use…
↗Not briefed
bluesky
bluesky-ai
2moUsing an LLM as a therapist, even if it gives you advice that doesn’t cause direct harm, is really really bad. Literally zero protections a…Using an LLM as a therapist, even if it gives you advice that doesn’t cause direct harm, is really really bad. Literally zero protecti…
↗Not briefed
bluesky
bluesky-ai
2motaz zitiert mit Yudkowsky einen kompletten Irren ohne ihn einzuordnen. Versucht es doch mal mit Journalismus!taz zitiert mit Yudkowsky einen kompletten Irren ohne ihn einzuordnen. Versucht es doch mal mit Journalismus!
↗Not briefed
bluesky
bluesky-ai
2moKHOSLA: “This isn’t redistribution. It will rebalance a tax code that was already redistributing, just upward. "The alternative is to do n…KHOSLA: “This isn’t redistribution. It will rebalance a tax code that was already redistributing, just upward. "The alternative is to…
↗Not briefed
hn
hn-ai
2moFlushing toilet releases 'plume' of breathable pathogens – closing lid helps—
↗Not briefed