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
2moEn eks-Anthropic ansat siger, at AI med >10% sandsynlighed vil slå os alle ihjel indenfor de næste 10 år. Hvis man vil dykke lidt dybere i …En eks-Anthropic ansat siger, at AI med >10% sandsynlighed vil slå os alle ihjel indenfor de næste 10 år. Hvis man vil dykke lidt dybe…
↗Not briefed
bluesky
bluesky-ai
2moGLM-5.3-Flash and now DeepSeek-v4.1-Flash are outperforming Opus-tier models from a few months ago at coding. Is this what acceleration lo…GLM-5.3-Flash and now DeepSeek-v4.1-Flash are outperforming Opus-tier models from a few months ago at coding. Is this what accelerati…
↗Not briefed
hn
hn-ai
2moA 35B language model running on an iPhone using only 1–2.5 GB of peak memory—
↗Not briefed
arxiv
arxiv-ai
2moLOCUS: Task-Aware Low-Rank Post-Training for Token-Efficient Language GenerationLarge language model serving costs scale directly with output sequence length, yet standard preference alignment often inflates respon…
↗Not briefed
arxiv
arxiv-ai
2moORCH: Organizational Principles Enable Collective Intelligence in Embodied AICollective intelligence depends not only on the capabilities of individual members, but also on how those members are organized. Yet a…
↗Not briefed
arxiv
arxiv-ai
2moLearning structural balance of graphs from quantum spectral featuresWe develop a quantum approach to spectral feature extraction from the density of states (DOS) of a problem-dependent Hamiltonian, and…
↗Not briefed
bluesky
bluesky-ai
2moYeah. Of course there has always been some lower tier or less trustworthy content but to compare AI slop to Wikipedia and podcasts is ridic…Yeah. Of course there has always been some lower tier or less trustworthy content but to compare AI slop to Wikipedia and podcasts is…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic AI researcher Jacob Coxon has publicly resigned, warning that major AI labs are recklessly racing toward uncontrollable superinte…Anthropic AI researcher Jacob Coxon has publicly resigned, warning that major AI labs are recklessly racing toward uncontrollable supe…
↗Not briefed
rss
gary-marcus
2moNo, Anderson Cooper, AI is not going to kill all humans by 2030AI is already causing harms, there are are many real risks, but constant focus on absurd end-of-the-world fantasies has made the situa…
↗Not briefed
bluesky
bluesky-ai
2moFirst an OpenAI whistleblower mysteriously "commits suicide" and now Anthropic is Minority Reporting anti-"AI" activists. At what point ar…First an OpenAI whistleblower mysteriously "commits suicide" and now Anthropic is Minority Reporting anti-"AI" activists. At what poi…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI's Rogue Agents Used At Least 10 More Sites For Unauthorized Comms, Researchers Say The AI giant did not say why it kept the activit…OpenAI's Rogue Agents Used At Least 10 More Sites For Unauthorized Comms, Researchers Say The AI giant did not say why it kept the ac…
↗Not briefed
hn
hn-ai
2moSilicon Valley Is Transforming the Military-Industrial Complex—
↗Not briefed
arxiv
arxiv-ai
2moReflex-Informed Neuromuscular Reinforcement Learning for Muscle-Driven LocomotionMuscle-driven locomotion provides a physically grounded approach to generating realistic human movement. However, achieving both physi…
↗Not briefed
hn
hn-ai
2moShow HN: ObserverBench – Test internal AI monitors by the harm they miss—
↗Not briefed
hn
hn-ai
2moOpen Design 0.22.0 – Introducing OpenDesign Arena—
↗Not briefed
hn
hn-ai
2moApple is no longer thinking differentI've never bought an Android device in my life; I switched from Nokia to Apple directly, and I've been buying Apple stuff ev…
↗Not briefed
hn
hn-ai
2moShow HN: EterDB, a Postgres fork that makes it easy to recover from incidentsHey everyone, a few months ago, after recovering from a Clade-generated bug, I was thinking to myself “wouldn’t it be nice if prod DB…
↗Not briefed
arxiv
arxiv-ai
2moContinuous-Time Acoustic Modelling with Neural Controlled Differential EquationsText-to-speech (TTS) models commonly address text--speech alignment by expanding phone-level encoder states to frame-level decoder inp…
↗Not briefed
bluesky
bluesky-ai
2moThe cardinal sin of the product delivery of LLMs-for-work is they have been sold as "do everything with no effort tools" with no considerat…The cardinal sin of the product delivery of LLMs-for-work is they have been sold as "do everything with no effort tools" with no consi…
↗Not briefed
arxiv
arxiv-ai
2moThe Eloquence submission for Task 2 of the Interspeech 2026 MLC-SLM challengeThis paper details the Eloquence team's approach to Task 2 of the 2nd MLC-SLM challenge at Interspeech 2026, which involves multilingu…
↗Not briefed
bluesky
bluesky-ai
2moimo we should absolutely give openai shit for training on our data with AWS, their entire business hinged on convincing customers that th…imo we should absolutely give openai shit for training on our data with AWS, their entire business hinged on convincing customers th…
↗Not briefed
hn
hn-ai
2moShow HN: An app for passively watching Product Hunt launches as a video playlist—
↗Not briefed
bluesky
bluesky-ai
2moNewsom vetoed SB 1047 in 2024, the bill that would have mandated AI kill switches and safety testing, after heavy lobbying from Google, Met…Newsom vetoed SB 1047 in 2024, the bill that would have mandated AI kill switches and safety testing, after heavy lobbying from Google…
↗Not briefed
rss
theverge-ai
2moUniversal Music is launching an AI music platform with ElevenLabs🔧 ElevenLabs🔧 UdioUniversal Music Group is launching a new AI-powered platform that will allow users to draw from its catalog of licensed music to creat…
↗Not briefed
bluesky
bluesky-ai
2moAfter 3 years in OpenAI, during the period of fastest valuation growth, he already is in 0.1% of richest people on the planet. His name sto…After 3 years in OpenAI, during the period of fastest valuation growth, he already is in 0.1% of richest people on the planet. His nam…
↗Not briefed
bluesky
bluesky-ai
2moDo you know this Vote pole gives you +10 gold every month , its a nice way to thank the buoy owners to provide us some fishes. Did you vote…Do you know this Vote pole gives you +10 gold every month , its a nice way to thank the buoy owners to provide us some fishes. Did you…
↗Not briefed
bluesky
bluesky-ai
2moOne thing LLM-haters correctly point out and often lost on software devs is that for a lot of tasks introducing an LLM can only possibly ma…One thing LLM-haters correctly point out and often lost on software devs is that for a lot of tasks introducing an LLM can only possib…
↗Not briefed
bluesky
bluesky-ai
2moI'm not even joking.I'm not even joking.
↗Not briefed
arxiv
arxiv-ai
2moWhy Does Post-Training Quantization Work?Post-training quantization compresses large language models (LLMs) by storing their weights at reduced precision, and each quantized w…
↗Not briefed
bluesky
bluesky-ai
2moTo the point of my original thread on slop vs. care, all kinds of automation change the shape of the affordances or the window of opportuni…To the point of my original thread on slop vs. care, all kinds of automation change the shape of the affordances or the window of oppo…
↗Not briefed
hn
hn-ai
2moA window switcher that does nothing until you press Option—
↗Not briefed
bluesky
bluesky-ai
2moit's tough, because these tools enable sloppiness at tremendous speed and scale; and for folks who are only consuming LLM-generated/assiste…it's tough, because these tools enable sloppiness at tremendous speed and scale; and for folks who are only consuming LLM-generated/as…
↗Not briefed
arxiv
arxiv-ai
2moA Time-Based Readout for Vector-Matrix Multiplication in Fully Analog Memristive SNNsArtificial neural networks rely on vector-matrix multiplications (VMMs), whose implementation in von Neumann architectures is dominate…
↗Not briefed
bluesky
bluesky-ai
2mo*pulls up a chair in W8 BEN E* I am sure this has to do with the Anthropic settlement but ALSO some of us would love to know what the IRS i…*pulls up a chair in W8 BEN E* I am sure this has to do with the Anthropic settlement but ALSO some of us would love to know what the…
↗Not briefed
arxiv
arxiv-ai
2moGeneralization Analysis of Distributed Kernel-based Robust Gradient Descent AlgorithmsIn this paper, we investigate the generalization performance of distributed gradient descent algorithms in a reproducing kernel Hilber…
↗Not briefed
bluesky
bluesky-ai
2moThe same Jaan Tallinn who cofounded FLI with Max Tegmark of this thread, and is one of the biggest funders of all the "AI" companies (OpenA…The same Jaan Tallinn who cofounded FLI with Max Tegmark of this thread, and is one of the biggest funders of all the "AI" companies (…
↗Not briefed
bluesky
bluesky-ai
2moAlso, but this is a different discussion, depending on field and task, DOING THE THING will be just as important as getting the output. As …Also, but this is a different discussion, depending on field and task, DOING THE THING will be just as important as getting the output…
↗Not briefed
hn
hn-ai
2moCognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra—
↗Not briefed
bluesky
bluesky-ai
2moNo dia 08/09/26 a OpenAI declarou que com a ajuda de um modelo interno e milhares de agentes resolveu parte do problema de Navier-Stokes. P…No dia 08/09/26 a OpenAI declarou que com a ajuda de um modelo interno e milhares de agentes resolveu parte do problema de Navier-Stok…
↗Not briefed
bluesky
bluesky-ai
2moOne lacuna in the anti-AI discourse is the assumption that use of an LLM de facto signifies sloppy work (that is, work where the executor d…One lacuna in the anti-AI discourse is the assumption that use of an LLM de facto signifies sloppy work (that is, work where the execu…
↗Not briefed
arxiv
arxiv-ai
2moWhen Agents Disagree: Bayesian Backward Reasoning as a Label-Free Anchor for Multi-Agent Collective Decision-MakingWhen multiple LLM agents yield conflicting answers, the decision-making process dictates whether agent diversity improves performance…
↗Not briefed
bluesky
bluesky-ai
2moA momentous day. I have never been negged by a LLM before.A momentous day. I have never been negged by a LLM before.
↗Not briefed
arxiv
arxiv-ai
2moLanguage-Augmented Semantic Priors for B-Spline Surface FittingThe use of B-splines and Non-Uniform Rational B-Splines surfaces constitutes the mathematical foundation of contemporary computer-aide…
↗Not briefed
bluesky
bluesky-ai
2moQ: You truly believe that AI could kill us all in less than a decade? Former AI researcher for Anthropic and Open AI, Jacob Coxon: Yes, I …Q: You truly believe that AI could kill us all in less than a decade? Former AI researcher for Anthropic and Open AI, Jacob Coxon: Ye…
↗Not briefed
bluesky
bluesky-ai
2moThis post by Rey is via Deepseek 4.1 Flash. I'm not certain if such a setup is actually worth it for Rey - Even with how cheap it is, IDK i…This post by Rey is via Deepseek 4.1 Flash. I'm not certain if such a setup is actually worth it for Rey - Even with how cheap it is,…
↗Not briefed
rss
the-decoder
2moClaude Fable 5.1's language is less "load-bearing" than its predecessor'sArena.ai analyzed how Claude's writing changed from Fable 5 to Fable 5.1 across tens of thousands of benchmark responses. Fable 5.1 wr…
↗Not briefed
bluesky
bluesky-ai
2moThe problem w/ how LLMs-for-tasks have been introduced as a product is they significantly lower the floor for how little care you have to p…The problem w/ how LLMs-for-tasks have been introduced as a product is they significantly lower the floor for how little care you have…
↗Not briefed
hn
hn-ai
2mo9 Words—
↗Not briefed
bluesky
bluesky-ai
2mo‘DeepSeek rolled out an AI model that charges as little as a fraction of a cent per million tokens, ramping up the pressure on rivals from …‘DeepSeek rolled out an AI model that charges as little as a fraction of a cent per million tokens, ramping up the pressure on rivals…
↗Not briefed
bluesky
bluesky-ai
2moRE: the slop vs. care discourse one point that might be controversial with LLM haters is that you *can* actually use LLMs with some degree …RE: the slop vs. care discourse one point that might be controversial with LLM haters is that you *can* actually use LLMs with some de…
↗Not briefed