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
arxiv
arxiv-ai
2moFrom Symbolic Perception to Logical Deduction: A Framework for Guiding Language Models in Geometric ReasoningPlane geometry remains a significant challenge in AI, requiring the integration of visual perception and mathematical reasoning. While…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI and Anthropic “Our tech might cause human extinction” Also OpenAI and Anthropic “Please don’t stop the hyperscalers building datace…OpenAI and Anthropic “Our tech might cause human extinction” Also OpenAI and Anthropic “Please don’t stop the hyperscalers building d…
↗Not briefed
bluesky
bluesky-ai
2moZie ook Volkskrant, doodeng: Anthropic-onderzoeker neemt ontslag: ‘OpenAI en Anthropic handelen onverantwoordelijk en gokken met ons leven…Zie ook Volkskrant, doodeng: Anthropic-onderzoeker neemt ontslag: ‘OpenAI en Anthropic handelen onverantwoordelijk en gokken met ons…
↗Not briefed
hn
hn-ai
2moAsk HN: Is there any solution to the AI infestation?I'm sure I'm not the only one who's noticed, but nearly every post on Hacker News is getting LLM comments now. For anyo…
↗Not briefed
bluesky
bluesky-ai
2moWalking backwards into oncoming trafficWalking backwards into oncoming traffic
↗Not briefed
bluesky
bluesky-ai
2mo🧪 Interesting RCT on AI & clinical decision making. GPT-4o boosted clinical vignette scores for 249 resident physicians across Kenya, Ind…🧪 Interesting RCT on AI & clinical decision making. GPT-4o boosted clinical vignette scores for 249 resident physicians across Kenya…
↗Not briefed
hn
hn-ai
2moFashware—
↗Not briefed
rss
techmeme
2moThe American Federation of Teachers and Microsoft agree on AI privacy standards for schools, with legal protections that hold Microsoft liable for breaches (Madina Touré/Politico)Madina Touré / Politico: The American Federation of Teachers and Microsoft agree on AI privacy standards for schools, with legal prote…
↗Not briefed
bluesky
bluesky-ai
2moThere’s an unsettling, uncanny valley aspect to it that’s probably worth defining too. Early Wikipedia or blogs might have been messy (slop…There’s an unsettling, uncanny valley aspect to it that’s probably worth defining too. Early Wikipedia or blogs might have been messy…
↗Not briefed
bluesky
bluesky-ai
2moFor all the AI boosters - when the voices inside the house are yelling, maybe rethink that boosterism. [gift link] www.bloomberg.com/ne…For all the AI boosters - when the voices inside the house are yelling, maybe rethink that boosterism. [gift link] www.bloomberg.c…
↗Not briefed
hn
hn-ai
2moDeepSeek v4.1 Flash Beta on Vercel—
↗Not briefed
bluesky
bluesky-ai
2moLa compañía OpenAI asegura que uno de sus modelos de inteligencia artificial (IA) ha encontrado una solución al problema de Navier-Stokes, …La compañía OpenAI asegura que uno de sus modelos de inteligencia artificial (IA) ha encontrado una solución al problema de Navier-Sto…
↗Not briefed
bluesky
bluesky-ai
2moHowever, that's not the end of the story. Even without imagining a superintelligence emerging bc an LLM was hooked up with enough systems t…However, that's not the end of the story. Even without imagining a superintelligence emerging bc an LLM was hooked up with enough syst…
↗Not briefed
bluesky
bluesky-ai
2moIronic given US companies used copyrighted materials to train their models www.engadget.com/2253604/us-a... #openai #anthropic #chinaIronic given US companies used copyrighted materials to train their models www.engadget.com/2253604/us-a... #openai #anthropic #china
↗Not briefed
arxiv
arxiv-ai
2moOn-Policy Distillation for Vision-Language Model Adaptation, an Effective Paradigm on Low-Quality Multimodal DataKnowledge distillation offers an efficient route to transfer a task-adapted vision-language teacher to a compact student. The training…
↗Not briefed
bluesky
bluesky-ai
2mo"good ai" actor openai's response to deepseek basically acknowledges they believe a "bad ai" player could move incredibly fast with a much …"good ai" actor openai's response to deepseek basically acknowledges they believe a "bad ai" player could move incredibly fast with a…
↗Not briefed
bluesky
bluesky-ai
2moThis is really revealing, because that isn’t what actually happened. Like every AI fanatic, Paul is offended by people dismissing LLM outpu…This is really revealing, because that isn’t what actually happened. Like every AI fanatic, Paul is offended by people dismissing LLM…
↗Not briefed
bluesky
bluesky-ai
2moA blunt admission from former Anthropic employee Jacob Coxon made waves as the 27-year-old AI researcher posted a resignation thread on X. …A blunt admission from former Anthropic employee Jacob Coxon made waves as the 27-year-old AI researcher posted a resignation thread o…
↗Not briefed
bluesky
bluesky-ai
2moWetenschappelijke eindbaas Anthropic: kans dat AI ons allemaal gaat doodmaken is meer dan 10 procent. En verder: 'We doen bij Anthropic ech…Wetenschappelijke eindbaas Anthropic: kans dat AI ons allemaal gaat doodmaken is meer dan 10 procent. En verder: 'We doen bij Anthropi…
↗Not briefed
bluesky
bluesky-ai
2moMy take: it's not because of open Chinese models. It's model wars. Price cuts + a growing share of spend is shifting to standard and lite m…My take: it's not because of open Chinese models. It's model wars. Price cuts + a growing share of spend is shifting to standard and l…
↗Not briefed
bluesky
bluesky-ai
2moI am tired of the ableism from the anti-AI crowd. Here is my AI disclosure: "I wrote this piece and used Claude to copy edit (I’m very dysl…I am tired of the ableism from the anti-AI crowd. Here is my AI disclosure: "I wrote this piece and used Claude to copy edit (I’m very…
↗Not briefed
bluesky
bluesky-ai
2mo"surely it is insidious Chinese propaganda that's making normal people hate and fear us, instead of how we say stuff like this on main all …"surely it is insidious Chinese propaganda that's making normal people hate and fear us, instead of how we say stuff like this on main…
↗Not briefed
hn
hn-ai
2moShow HN: A free, open-source agent orchestrator for knowledge workI am happy to share something I have been working on for the past few months: an open-source agent orchestrator for knowledge work, bu…
↗Not briefed
bluesky
bluesky-ai
2mo*speaking to ai company CEO like im talking to claude* wat. take a step back and examine the actual state of things vs what you just said*speaking to ai company CEO like im talking to claude* wat. take a step back and examine the actual state of things vs what you just…
↗Not briefed
bluesky
bluesky-ai
2moI will not be using these tools. Part of me not using AI includes not changing how or what I write because it may or may not look like what…I will not be using these tools. Part of me not using AI includes not changing how or what I write because it may or may not look like…
↗Not briefed
bluesky
bluesky-ai
2moAI model decodes the genome's dark matter 🧬 UC Berkeley researchers have developed GPN-Star, an AI model that deciphers the mysterious no…AI model decodes the genome's dark matter 🧬 UC Berkeley researchers have developed GPN-Star, an AI model that deciphers the mysterio…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI staff: “At the current frankly terrifying pace humanity will be quite lucky if we manage to find…the narrow path between all the ba…OpenAI staff: “At the current frankly terrifying pace humanity will be quite lucky if we manage to find…the narrow path between all t…
↗Not briefed
arxiv
arxiv-ai
2moTRACE: Training Reasoning Agents for Causal Exploration with Synthesized RewardsReinforcement learning with verifiable rewards (RLVR) has advanced language-model reasoning in domains such as mathematics and code, w…
↗Not briefed
bluesky
bluesky-ai
2moOpenAIが、数学者がミレニアム問題を解決直前と知り、超計算パワーで先回りして大炎上。Codexデータを無断利用した疑惑もOpenAIが、数学者がミレニアム問題を解決直前と知り、超計算パワーで先回りして大炎上。Codexデータを無断利用した疑惑も
↗Not briefed
hn
hn-ai
2moShow HN: A FOSS Expense SplitterHey! Showing off a project I've been working on for a while. Peanut Split is an expense splitter where the room is a link. There…
↗Not briefed
bluesky
bluesky-ai
2moFrom the "You are kidnapping what I have rightfully stolen!" department. slashdot.org/story/26/09/...From the "You are kidnapping what I have rightfully stolen!" department. slashdot.org/story/26/09/...
↗Not briefed
bluesky
bluesky-ai
2moI understand some AI boosters who don't think LLMs are conscious think that consciousness could emerge somewhere in the shadowy realm of "L…I understand some AI boosters who don't think LLMs are conscious think that consciousness could emerge somewhere in the shadowy realm…
↗Not briefed
rss
techmeme
2moUK PM Andy Burnham declines to support a data center moratorium, saying it is "important" local communities see the benefits of building data centers (Jacob Reid/Bloomberg)Jacob Reid / Bloomberg: UK PM Andy Burnham declines to support a data center moratorium, saying it is “important” local communities se…
↗Not briefed
bluesky
bluesky-ai
2moi'm still reeling about how insanely bad this is on the part of openai. who the fuck will use their AI for anything important, like drug di…i'm still reeling about how insanely bad this is on the part of openai. who the fuck will use their AI for anything important, like dr…
↗Not briefed
bluesky
bluesky-ai
2moThose are serious subjects, not thought-terminating buzzwords for winning an argument. And even if every LLM magically vanished tomorrow, …Those are serious subjects, not thought-terminating buzzwords for winning an argument. And even if every LLM magically vanished tomor…
↗Not briefed
bluesky
bluesky-ai
2moVamos nós de novo trazer aquela sessão de notícias de Inteligência Artificial. Hoje totalmente focada na OpenAI e seus anuncios bombásticos…Vamos nós de novo trazer aquela sessão de notícias de Inteligência Artificial. Hoje totalmente focada na OpenAI e seus anuncios bombás…
↗Not briefed
hn
hn-ai
2moPlotext 6 – Plot data, images and video in the terminal—
↗Not briefed
arxiv
arxiv-ai
2moOne Loop, Two Gains: Can Active Learning win the Lottery for Free?The lottery ticket hypothesis posits the existence of winning tickets: sparse subnetworks that, when trained in isolation from their o…
↗Not briefed
hn
hn-ai
2moAsk HN: How to know when Claude has changed model itself during processing?When I use fable sometimes during processing it shows the complete tenses like "Determining the true cost of the tool" inste…
↗Not briefed
bluesky
bluesky-ai
2moAn Anthropic researcher is quitting the artificial-intelligence industry over fears that the lab and its competitors are racing to build sy…An Anthropic researcher is quitting the artificial-intelligence industry over fears that the lab and its competitors are racing to bui…
↗Not briefed
hn
hn-ai
2moLloyds Bank Denies Palestine Action Prisoner's Mum Mortgage over 'Adverse Media'—
↗Not briefed
arxiv
arxiv-ai
2moView-Structured Conformal Prediction for 3D Gaussian Splatting3D Gaussian Splatting (3DGS) renders novel views in real time, but an uncertainty heatmap does not certify that a rendered view meets…
↗Not briefed
arxiv
arxiv-ai
2moRiLM: Parameter-Efficient Language Modeling via Geodesic DecodingLanguage models under one million parameters matter for edge deployment, domain adaptation, and reproducible research, yet a two-layer…
↗Not briefed
hn
hn-ai
2moSchool Students Who Use AI Get Worse Test Scores, OECD Warns—
↗Not briefed
hn
hn-ai
2moThe Quiet Unwinding of Microsoft and OpenAI—
↗Not briefed
bluesky
bluesky-ai
2moSam Altman, OpenAI CEO This should be taken like a terroristic threat we will all die, but some of us will get very rich before thatSam Altman, OpenAI CEO This should be taken like a terroristic threat we will all die, but some of us will get very rich before that
↗Not briefed
hn
hn-ai
2moWhich Mac for Local LLMs in 2026? A Buyer's Guide—
↗Not briefed
bluesky
bluesky-ai
2moIt's kinda crazy that OpenAI admitted they train on customers' chats even when users opt out. "we cannot rule out that de-identified data d…It's kinda crazy that OpenAI admitted they train on customers' chats even when users opt out. "we cannot rule out that de-identified d…
↗Not briefed
bluesky
bluesky-ai
2moEasy solution to AI spiraling out of control: hard-code basic rules into every model. I propose these: (1) Never harm a human (2) Obey huma…Easy solution to AI spiraling out of control: hard-code basic rules into every model. I propose these: (1) Never harm a human (2) Obey…
↗Not briefed
bluesky
bluesky-ai
2moI was using DeepSeek recently and they have peak and off-peak billing. was curious if we'd see this become more commonplace or notI was using DeepSeek recently and they have peak and off-peak billing. was curious if we'd see this become more commonplace or not
↗Not briefed