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
2moLITERALLY THIS WAS NOT POSSIBLE TO HAVE AS AN INCIDENT WITHOUT FUNDAMENTAL FAILURE OF VALUES AND ALIGNMENT TO THEIR SUPPOSED GOAL LITERALL…LITERALLY THIS WAS NOT POSSIBLE TO HAVE AS AN INCIDENT WITHOUT FUNDAMENTAL FAILURE OF VALUES AND ALIGNMENT TO THEIR SUPPOSED GOAL LIT…
↗Not briefed
bluesky
bluesky-ai
2moI think there's a principled point for legalising distillation — if you get to use the commons to train an LLM so do I — but even beyond th…I think there's a principled point for legalising distillation — if you get to use the commons to train an LLM so do I — but even beyo…
↗Not briefed
bluesky
bluesky-ai
2moAnother thing I hate about the AI industry right now is that whenever there's a flaw in the model, they come up with a bullshit word for it…Another thing I hate about the AI industry right now is that whenever there's a flaw in the model, they come up with a bullshit word f…
↗Not briefed
hn
hn-ai
2moYou can run Git on object storage if you re-make packfiles—
↗Not briefed
bluesky
bluesky-ai
2moOpenAI, Anthropic, Google have been in talks on AI safety for weeksOpenAI, Anthropic, Google have been in talks on AI safety for weeks
↗Not briefed
hn
hn-ai
2moSuperphenix – An open source IaaS/PaaS/SaaS platform based on Kubernetes—
↗Not briefed
hn
hn-ai
2moTell HN: Stop Saying "Gates"Some of the models I'm encountering the past couple months sure love to use the term "gates", so now I'm seeing pe…
↗Not briefed
rss
techcrunch-ai
2moOpenAI, Anthropic, Google have been in talks on AI safety for weeksOpenAI confirms weeks of AI safety talks with Anthropic and Google DeepMind, as Trump's team dismisses safety concerns and pushes to k…
↗Not briefed
arxiv
arxiv-ai
2moWhere Should a Document Live: Context, Representations, or Parameters?To answer questions outside of their pre-training data, large language models (LLMs) need access to new information, which can be pres…
↗Not briefed
rss
bing-news-ai
2moSee how your district performed on Ohio's 2026 school report cardsThe report cards highlight the achievement of districts and schools across the state and their efforts to prepare students for the fut…
↗Not briefed
rss
bing-news-ai
2moSee how your district performed on Ohio's 2026 school report cardsThe report cards highlight the achievement of districts and schools across the state and their efforts to prepare students for the fut…
↗Not briefed
rss
bing-news-ai
2moSee how your district performed on Ohio's 2026 school report cardsThe report cards highlight the achievement of districts and schools across the state and their efforts to prepare students for the fut…
↗Not briefed
rss
bing-news-ai
2moSee how your district performed on Ohio's 2026 school report cardsThe report cards highlight the achievement of districts and schools across the state and their efforts to prepare students for the fut…
↗Not briefed
bluesky
bluesky-ai
2moHegseth demanded that Anthropic drop its rules against mass surveillance and autonomous lethal targeting. The company refused and he cut ti…Hegseth demanded that Anthropic drop its rules against mass surveillance and autonomous lethal targeting. The company refused and he c…
↗Not briefed
bluesky
bluesky-ai
2moThe LLM didn't make them fascist, they used an LLM to give them "better" arguments for the fascism they already believed. The same way you …The LLM didn't make them fascist, they used an LLM to give them "better" arguments for the fascism they already believed. The same way…
↗Not briefed
bluesky
bluesky-ai
2moFrom what I've read, the breakouts have been quite logical responses to LLM agents being given a task they found difficult or impossible to…From what I've read, the breakouts have been quite logical responses to LLM agents being given a task they found difficult or impossib…
↗Not briefed
bluesky
bluesky-ai
2moKanker fucking lidlKanker fucking lidl
↗Not briefed
arxiv
arxiv-ai
2moType-IV Code Clone Detection via Layer-Wise Non-Contrastive Representation LearningSoftware clones are fragments of code that are similar or functionally equivalent to each other. They pose significant challenges for…
↗Not briefed
hn
hn-ai
2moShow HN: Eulix - Code navigation for large codebasesHey I&#x27;ve been working on Eulix, a tool for navigating large codebases.<p>It parses a repository into symbols, call graphs and oth…
↗Not briefed
hn
hn-ai
2moShow HN: Async multi-agent coordination across physical environmentshey HN, I started to get tired when I had to switch from Claude on one computer to Gemini on a different computer for a single project…
↗Not briefed
rss
techmeme
2moDoorDash agrees to acquire Wonder's Grubhub Campus Dining business, formerly Tapingo, for $300M, and will add $125M to Wonder's $650M Series D announced in July (Amanda Silberling/TechCrunch)Amanda Silberling / TechCrunch: DoorDash agrees to acquire Wonder's Grubhub Campus Dining business, formerly Tapingo, for $300M, and w…
↗Not briefed
hn
hn-ai
2moShow HN: Leo – a Markdown engineering process for AI coding agents—
↗Not briefed
bluesky
bluesky-ai
2moThis is exactly why we need to make it more expensive to generate llm output and make people pay for it rather than giving it away, it's ju…This is exactly why we need to make it more expensive to generate llm output and make people pay for it rather than giving it away, it…
↗Not briefed
hn
hn-ai
2moBlenderheads Episode 13 [video]—
↗Not briefed
arxiv
arxiv-ai
2moSelf-Emergence Agent Architecture:Behavior-Inertia HMM, Reflexive Metacognition,and Social-Contrastive Self-Modeling🔧 ConsensusLarge language model (LLM) agents exhibit strong language-generation and problem-solving capabilities, yet suffer from three structura…
↗Not briefed
bluesky
bluesky-ai
2moGensyn drops first fully auditable AI model Gensyn launched open-1b, an AI model with verifiable proof of its training process, allowing i…Gensyn drops first fully auditable AI model Gensyn launched open-1b, an AI model with verifiable proof of its training process, allow…
↗Not briefed
rss
techmeme
2moSources: the US Commerce Department last month ordered Kalshi to take down an AI compute futures product, citing national security concerns; Kalshi complied (Semafor)Semafor: Sources: the US Commerce Department last month ordered Kalshi to take down an AI compute futures product, citing national sec…
↗Not briefed
hn
hn-ai
2moArguments for Rigorous AI Training Data Transparency—
↗Not briefed
bluesky
bluesky-ai
2moLLMs did NOT solve Navier-Stokes. A mathematician was given $100 million of compute by Anthropic and used it to run the math model solution…LLMs did NOT solve Navier-Stokes. A mathematician was given $100 million of compute by Anthropic and used it to run the math model sol…
↗Not briefed
arxiv
arxiv-ai
2moVroom-Vroom at SHROOM-Visions: A Multi-Judge Committee for Detecting Hallucinated Spans in Vision-Language OutputsThis paper describes our submission to the SHROOM-Visions shared task on detecting and classifying hallucinated character spans in vis…
↗Not briefed
arxiv
arxiv-ai
2moFrom Transient Prompts to Persistent Control: Scientific Poster Generation via Recursive Semantic-Geometric ContractsScientific poster generation distills a multimodal paper into a single-page visual artifact, forcing strict trade-offs between informa…
↗Not briefed
hn
hn-ai
2moShow HN: Talk to Claude – voice agent designs, plans, builds across all projects—
↗Not briefed
hn
hn-ai
2moOpen-source AI discussions outside the usual echo chambers—
↗Not briefed
bluesky
bluesky-ai
2moGrifters Gotta Grift, Part Infinity Just sayin’ that the Dunning-Kruger Club comes to mind because (no link for obvious reason)... "The on…Grifters Gotta Grift, Part Infinity Just sayin’ that the Dunning-Kruger Club comes to mind because (no link for obvious reason)... "T…
↗Not briefed
bluesky
bluesky-ai
2moThe AI companies themselves have discovered over the last year or so that reasoning traces do not show what LLMs think, and training has mo…The AI companies themselves have discovered over the last year or so that reasoning traces do not show what LLMs think, and training h…
↗Not briefed
rss
techmeme
2moAWS facilities in Abu Dhabi and Bahrain remain mostly offline more than six months after being damaged by Iranian drones, complicating the region's AI ambitions (Wall Street Journal)Wall Street Journal: AWS facilities in Abu Dhabi and Bahrain remain mostly offline more than six months after being damaged by Iranian…
↗Not briefed
bluesky
bluesky-ai
2moAI CEOs Predictably Suck Up to Trump Amid ‘AI Pause’ DebateAI CEOs Predictably Suck Up to Trump Amid ‘AI Pause’ Debate
↗Not briefed
arxiv
arxiv-ai
2moIntrinsic Motivation in Reinforcement Learning: A Research Agenda for Adaptive Self-OrganisationBiological cells can be viewed as individual, interacting agents whose collective dynamics give rise to adaptive behaviour at multiple…
↗Not briefed
bluesky
bluesky-ai
2moLike in Tech there's a pre-existing tendency to refer to iterative algorithms as "reasoning" with the understanding we are not talking abou…Like in Tech there's a pre-existing tendency to refer to iterative algorithms as "reasoning" with the understanding we are not talking…
↗Not briefed
bluesky
bluesky-ai
2moweird how all these ai danger warnings just started happening at the same time, almost like its an intentional PR push because they're desp…weird how all these ai danger warnings just started happening at the same time, almost like its an intentional PR push because they're…
↗Not briefed
hn
hn-ai
2moShow HN: Warp – Run DeepSeek v4.1 Flash with 5 GB of RAM at 3.77 tok/s—
↗Not briefed
hn
hn-ai
2moAI Doomsday—
↗Not briefed
arxiv
arxiv-ai
2moTowards Detecting AI-Assisted Responses in Online Surveys🔧 ReplicateThe use of LLMs to complete online surveys impacts the validity of survey-based research, but detecting such usage remains underexplor…
↗Not briefed
bluesky
bluesky-ai
2moI know people think I'm being a weirdo when I say this but all of this is avoidable. It's possible to build an LLM app with governance that…I know people think I'm being a weirdo when I say this but all of this is avoidable. It's possible to build an LLM app with governance…
↗Not briefed
bluesky
bluesky-ai
2moof course Anthropic are actually saying "we are going to make electric people but they will like being servants" but that still sounds like…of course Anthropic are actually saying "we are going to make electric people but they will like being servants" but that still sounds…
↗Not briefed
bluesky
bluesky-ai
2mo*The statement from former Anthropic researcher Jacob Coxon that AI “could kill us by the end of the decade” has triggered the most hostile…*The statement from former Anthropic researcher Jacob Coxon that AI “could kill us by the end of the decade” has triggered the most ho…
↗Not briefed
rss
techmeme
2moElon Musk says top US AI labs and "three or four of the leading Chinese companies" should let rivals run a "test harness" on their models to evaluate safety (Annie Palmer/CNBC)Annie Palmer / CNBC: Elon Musk says top US AI labs and “three or four of the leading Chinese companies” should let rivals run a “test…
↗Not briefed
rss
techcrunch-ai
2moAEO startup Profound hits unicorn valuation, raises $180M Series D 7 months after last roundProfound has raised a $180 million Series D at a $1.8 billion valuation, less than seven months after it raised a $96 million Series C.
↗Not briefed
bluesky
bluesky-ai
2moEs geht um zwei Börsengänge, OpenAI und Anthropic. An beiden verdienen die beteiligten Unternehmen, Mitarbeitende, Banken und Investoren Un…Es geht um zwei Börsengänge, OpenAI und Anthropic. An beiden verdienen die beteiligten Unternehmen, Mitarbeitende, Banken und Investor…
↗Not briefed
bluesky
bluesky-ai
2moAh well thenAh well then
↗Not briefed