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
2moKoska lapset ja nuoret eivät ilmeisesti pysty tässä teknistyneessä ja ärsykkeiden täyttämässä maailmassa keskittymään lukemiseen, ehdotan v…Koska lapset ja nuoret eivät ilmeisesti pysty tässä teknistyneessä ja ärsykkeiden täyttämässä maailmassa keskittymään lukemiseen, ehdo…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI adds AI doomer to its board 🤖 Paul Christiano, inventor of RLHF and a vocal AI alignment researcher, joins the OpenAI Foundation b…OpenAI adds AI doomer to its board 🤖 Paul Christiano, inventor of RLHF and a vocal AI alignment researcher, joins the OpenAI Foundat…
↗Not briefed
bluesky
bluesky-ai
2moEin ehemaliger Anthropic Mitarbeiter meinte, KI könne die Menschenheit bis 2030 auslöschen. Etwas enttäuschend, damit verschiebt sich der B…Ein ehemaliger Anthropic Mitarbeiter meinte, KI könne die Menschenheit bis 2030 auslöschen. Etwas enttäuschend, damit verschiebt sich…
↗Not briefed
bluesky
bluesky-ai
2moIm Juli sind KI-Agenten von OpenAI aus ihrer Test-Umgebung ausgebrochen und haben die KI-Firma Hugging Face gehackt. Berichte legen den Umf…🔧 Hugging FaceIm Juli sind KI-Agenten von OpenAI aus ihrer Test-Umgebung ausgebrochen und haben die KI-Firma Hugging Face gehackt. Berichte legen de…
↗Not briefed
hn
hn-ai
2moShow HN: Community curated list of 310 AI providers offering $4.3k free creditsHey people! I have been compiling a directory of AI API providers which provide a combined total of $4.3k in free API credits and I&#x…
↗Not briefed
rss
wired-ai
2moClearview AI Is Testing an AI Tool That Would Let Cops Unearth Your Life OnlineInquiryIQ, a previously unreported prototype, tested a model from xAI, maker of Grok, to surface associates, social accounts, and othe…
↗Not briefed
rss
stratechery
2moThe iPhone Duo, The Intelligent Personal Hub, Apple Watch Audio IntelligenceApple once again demonstrated the power of integrating hardware and software, but it's biggest AI blindspot might be its belief in the…
↗Not briefed
bluesky
bluesky-ai
2moA trans LLM might actually cause parts of bsky to implode.A trans LLM might actually cause parts of bsky to implode.
↗Not briefed
hn
hn-ai
2moAsk HN: As a teacher how should I teach kids to use AI?Hi,<p>An institution asked me to help them in teaching poor kids to use AI and about AI. They are between 7 - 14 years old from poor f…
↗Not briefed
bluesky
bluesky-ai
2moAlways nice to wake up to DeepSeek eating everyone's lunch again. "But it's not as good as Fable or Astra!" "Good" is doing a lot of work…Always nice to wake up to DeepSeek eating everyone's lunch again. "But it's not as good as Fable or Astra!" "Good" is doing a lot of…
↗Not briefed
arxiv
arxiv-ai
2moThe Semantic Elevation Operator and the Closure of the Undecidable Class under PreservationThe undecidability of a program's static semantic properties is governed by Rice's theorem. Self-modifying systems, however, require a…
↗Not briefed
rss
techmeme
2moDeepSeek debuts DeepSeek-V4.1-Flash, its smallest model built on a new Causal Encoder-Decoder architecture, with 552B backbone parameters and 1M-token context (Reuters)Reuters: DeepSeek debuts DeepSeek-V4.1-Flash, its smallest model built on a new Causal Encoder-Decoder architecture, with 552B backbon…
↗Not briefed
bluesky
bluesky-ai
2moEsto: OpenAI corrige de tapadillo su prueba ‘del milenio’ para citar a los matemáticos clave que ninguneó al anunciar el descubrimiento el…Esto: OpenAI corrige de tapadillo su prueba ‘del milenio’ para citar a los matemáticos clave que ninguneó al anunciar el descubrimien…
↗Not briefed
bluesky
bluesky-ai
2moICYMI: AI researcher Jacob Coxon sent shock waves through Silicon Valley and beyond on Tuesday by announcing his resignation from Anthropic…ICYMI: AI researcher Jacob Coxon sent shock waves through Silicon Valley and beyond on Tuesday by announcing his resignation from Anth…
↗Not briefed
hn
hn-ai
2moGuarantee or accuracy: a CCAR-F question type—
↗Not briefed
arxiv
arxiv-ai
2moMultiHuSE: A Multimodal Dataset for Humour Styles and EmotionsComputational recognition of verbal humour remains a challenging task, requiring an understanding of language, delivery style, emotion…
↗Not briefed
bluesky
bluesky-ai
2moEn el caso de OpenAI no hacen ningún reconocimiento a estos dos matemáticos y fueron con amenazas al matemático que colaboraba con la compe…En el caso de OpenAI no hacen ningún reconocimiento a estos dos matemáticos y fueron con amenazas al matemático que colaboraba con la…
↗Not briefed
arxiv
arxiv-ai
2moAI Exposure and AI Resilience: A Two-Dimensional Assessment Framework for Software and Software-Based Business ModelArtificial intelligence is changing both software production and the economics of software-based business models. Classical technology…
↗Not briefed
arxiv
arxiv-ai
2moMagenta: Closing the Loop Between Mathematical Reasoning and Lean VerificationMost of mathematical knowledge has been communicated through so-called informal use of mathematics and natural language. With large la…
↗Not briefed
bluesky
bluesky-ai
2moI am reasonably sangfroid about the AI/Skynet killing us stuff because I'm not very intelligent and these concepts are just beyond my conce…I am reasonably sangfroid about the AI/Skynet killing us stuff because I'm not very intelligent and these concepts are just beyond my…
↗Not briefed
bluesky
bluesky-ai
2moWe need to really focus on the fact that OpenAI employees are behind this not sentient software. Like, if a person/company creates a malici…We need to really focus on the fact that OpenAI employees are behind this not sentient software. Like, if a person/company creates a m…
↗Not briefed
arxiv
arxiv-ai
2moMr.LHDR: A Benchmark for Multimodal Real-World Long-Horizon Deep Research AgentsDeep research agents are increasingly capable of web search, tool use, multimodal evidence analysis, and information synthesis. Howeve…
↗Not briefed
rss
techmeme
2moLetter: the Senate disaster management subcommittee, led by Sen. Josh Hawley, is probing OpenAI's handling of the Hugging Face breach, calling it "reckless" (Axios)🔧 Hugging FaceAxios: Letter: the Senate disaster management subcommittee, led by Sen. Josh Hawley, is probing OpenAI's handling of the Hugging Face…
↗Not briefed
hn
hn-ai
2moWeWorm: The first zero-click worm to spread through WeChat calls—
↗Not briefed
hn
hn-ai
2moAWS Lambda now supports 90-minute function timeout on Lambda Managed Instances—
↗Not briefed
arxiv
arxiv-ai
2moRouting by Reasoning Need: Trajectory-Aware Decoding Control for Diffusion Vision-Language ModelsDiffusion vision-language models generate answers through iterative refinement, exposing intermediate answer trajectories that can be…
↗Not briefed
bluesky
bluesky-ai
2moScoop: Anthropic whistleblower gave up his equity to leave the company www.axios.com/2026/09/09/a... - 앤스로픽 연구원 제이콥 콕슨은 AI의 치명적인 위험성을 경고하…Scoop: Anthropic whistleblower gave up his equity to leave the company www.axios.com/2026/09/09/a... - 앤스로픽 연구원 제이콥 콕슨은 AI의 치명적인 위험성…
↗Not briefed
arxiv
arxiv-ai
2moA Dynamic Fusion Large Language Model for Traffic Flow PredictionTraffic flow prediction is a core supporting technology for intelligent transportation systems. It uses historical data to infer futur…
↗Not briefed
bluesky
bluesky-ai
2moJob postings and interviews with senior security officials at Anthropic show that the frontier AI lab is building out an extensive monitori…Job postings and interviews with senior security officials at Anthropic show that the frontier AI lab is building out an extensive mon…
↗Not briefed
arxiv
arxiv-ai
2moGRIPNet: Gaussian Radial Intensity Prior Guided Architecture for Pulmonary Nodule Detection in CTLung cancer causes more deaths than any other malignancy, and low-dose CT screening is the main pathway to early diagnosis. That pathw…
↗Not briefed
arxiv
arxiv-ai
2moYour Model Already Knows Don't Teach It, Learn to Ask It: Soft Prompting for Few-Shot Adaptation of Vision-Language Models🔧 ModalWe address few-shot object detection with vision-language models (VLMs) in out-of-domain settings such as aerial, industrial, and medi…
↗Not briefed
arxiv
arxiv-ai
2mo2AM: Grounding Agent-Side Memory as Guidance for Steerable Action Models in Long-Horizon ManipulationLong-horizon robot manipulation requires memory, but not necessarily inside the action policy. To address such tasks, current agentic…
↗Not briefed
rss
the-decoder
2moTop AI spenders cut per-employee costs by nearly 10 percent in AugustThe Ramp AI Index for September 2026 shows AI spending per employee among the top 1 percent of US companies fell nearly 10 percent in…
↗Not briefed
bluesky
bluesky-ai
2mowww.cnbc.com/2026/09/09/a... isn't Skynet, I mean Cyberdyne, oops AI doing that already?www.cnbc.com/2026/09/09/a... isn't Skynet, I mean Cyberdyne, oops AI doing that already?
↗Not briefed
arxiv
arxiv-ai
2moAutomatic Lyric Transcription for Greek Songs: Scaling and Task Composition Effects in Whisper Adaptation🔧 WhisperAutomatic Lyric Transcription (ALT) remains substantially more challenging than speech recognition due to melodic variability, rhythmi…
↗Not briefed
arxiv
arxiv-ai
2moA Two-Mirror Faceted Projection System for EUV Lithography🔧 ModalWe propose an all-reflective two-mirror projection system for extreme ultraviolet (EUV) lithography operating at exposure wavelengths…
↗Not briefed
bluesky
bluesky-ai
2moModel trained on millions of protein measurements can gauge drug effectiveness in tissue samples taken from people with triple-negative bre…Model trained on millions of protein measurements can gauge drug effectiveness in tissue samples taken from people with triple-negativ…
↗Not briefed
bluesky
bluesky-ai
2mo2030 is in 4 years time. Actually less cos we’re in the latter part of 2026 already. If we heat up the planet with enough data centres an…2030 is in 4 years time. Actually less cos we’re in the latter part of 2026 already. If we heat up the planet with enough data centr…
↗Not briefed
arxiv
arxiv-ai
2moMemory Compression for High-Fanout Agent SandboxesHigh-fanout agent workloads create a growing memory bottleneck because a single task may spawn many concurrent sandbox sessions. Yet t…
↗Not briefed
arxiv
arxiv-ai
2moA Hilbert-Valued Functional Decomposition Framework for Explaining Time-Dependent OutputsFeature-based explanations quantify features' influence on model predictions, but are primarily designed for scalar outputs. In many a…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI corrige de tapadillo su prueba ‘del milenio’ para citar a los matemáticos clave que ninguneó al anunciar el descubrimiento elpais.c…OpenAI corrige de tapadillo su prueba ‘del milenio’ para citar a los matemáticos clave que ninguneó al anunciar el descubrimiento elp…
↗Not briefed
bluesky
bluesky-ai
2moCentrist Dad take on AI. 1. Large language models are not going to kill us. 2. We should look at regulating it, but I'd like to see speci…Centrist Dad take on AI. 1. Large language models are not going to kill us. 2. We should look at regulating it, but I'd like to see…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI é unha empresa de vendedores de fume e Anthropic é basicamente unha secta apocalíptica. Non vos creades NUNCA nada do que diganOpenAI é unha empresa de vendedores de fume e Anthropic é basicamente unha secta apocalíptica. Non vos creades NUNCA nada do que digan
↗Not briefed
bluesky
bluesky-ai
2moO mal dos LLM é que as pessoas usam até para escrever em grupos de WhatsApp. E agora, toda a gente escreve parecido. Não se distinguem uma…O mal dos LLM é que as pessoas usam até para escrever em grupos de WhatsApp. E agora, toda a gente escreve parecido. Não se distingue…
↗Not briefed
arxiv
arxiv-ai
2moOff-Target Effects of Response-Style Alignment in a Korean 27B Language ModelWe post-train Qwen3.8-27B for Korean response style -- verbosity, list and markdown usage, discourse structure and register -- and mea…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI adds a prominent AI doomer to its board of directors | TechCrunch techcrunch.com/2026/09/09/o... - AI 안전성 연구 권위자이자 'AI 파멸론자'로 불리는 …OpenAI adds a prominent AI doomer to its board of directors | TechCrunch techcrunch.com/2026/09/09/o... - AI 안전성 연구 권위자이자 'AI 파멸론자'로…
↗Not briefed
bluesky
bluesky-ai
2mo"when you design these systems, you're building a pinball machine for words, with different levers interacting to bounce the steel ball of …"when you design these systems, you're building a pinball machine for words, with different levers interacting to bounce the steel bal…
↗Not briefed
bluesky
bluesky-ai
2moRegardless of how true the claim "the [LLM] we're building will kill us all in 4 years" is, it's telling that the markets are continuing to…Regardless of how true the claim "the [LLM] we're building will kill us all in 4 years" is, it's telling that the markets are continui…
↗Not briefed
bluesky
bluesky-ai
2moBreaking: Anthropic announces new frontier model has resolved the St Ives problem, confirms it was 7 cats, each cat had 7 legs.Breaking: Anthropic announces new frontier model has resolved the St Ives problem, confirms it was 7 cats, each cat had 7 legs.
↗Not briefed
bluesky
bluesky-ai
2moit's darkly funny how all the people leaving Anthropic are like "no the Anthropic take the dangers of AI really seriously! They just want t…it's darkly funny how all the people leaving Anthropic are like "no the Anthropic take the dangers of AI really seriously! They just w…
↗Not briefed