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
2moForward-Free LLM Depth Pruning via Weight RedundancyDepth pruning reduces large language model (LLM) inference cost by removing complete Transformer blocks. Activation-based methods coll…
↗Not briefed
arxiv
arxiv-ai
2moScored vs. Generated Readouts in Behavioral Language Models: An Empirical Study of Elicitation FormatLanguage models fine-tuned on customer behavior can predict outcomes and generate explanations, but these readouts are often treated a…
↗Not briefed
bluesky
bluesky-ai
2mome pasa con lo de "la IA". La inteligencia artificial lleva integrada en nuestras vidas más de dos décadas. Pero de pronto IA por aquí, IA …me pasa con lo de "la IA". La inteligencia artificial lleva integrada en nuestras vidas más de dos décadas. Pero de pronto IA por aquí…
↗Not briefed
hn
hn-ai
2moRails devs: heads up on JSON 3.0 gem—
↗Not briefed
bluesky
bluesky-ai
2moEvery week we discover new elements of the OpenAI training incident, AI safety researchers are extremely alarmed, still new models are bein…Every week we discover new elements of the OpenAI training incident, AI safety researchers are extremely alarmed, still new models are…
↗Not briefed
arxiv
arxiv-ai
2moAgentAudit: An Open, Extensible Framework for Full-Lifecycle Trust Evaluation of AI Agents🔧 GPT-5Existing evaluation frameworks mostly assess only one part of AI agents, such as task completion (AgentBench) or security robustness (…
↗Not briefed
bluesky
bluesky-ai
2moJacob Coxon, chercheur en #IA, quitte #Anthropic après avoir quitté #OpenAI qu'il jugeait déraisonnable. Je pense que qualifier les averti…Jacob Coxon, chercheur en #IA, quitte #Anthropic après avoir quitté #OpenAI qu'il jugeait déraisonnable. Je pense que qualifier les a…
↗Not briefed
rss
bing-news-ai
2moUAE, France set out to lead the world in artificial intelligence in space with Altair-Next GenUAE and France lead $1B Altair-Next Gen AI satellite constellation with Marlan Space, Loft Orbital and Mistral, delivering real-time i…
↗Not briefed
hn
hn-ai
2moShow HN: Tekushi Enterprise – Live online threat data (free if neutralized)We’ve spent this summer building something critical.<p>Today, we’re officially rolling out Tekushi for Enterprise! Check it out: <a hr…
↗Not briefed
bluesky
bluesky-ai
2moMeanwhile just a few miles from the A66 crash a few weeks ago Teesside is rolling out the red carpet for Anthropic to build a huge data cen…Meanwhile just a few miles from the A66 crash a few weeks ago Teesside is rolling out the red carpet for Anthropic to build a huge dat…
↗Not briefed
bluesky
bluesky-ai
2mo«Elles jouent avec nos vies» : un chercheur en intelligence artificielle quitte Anthropic avec fracas«Elles jouent avec nos vies» : un chercheur en intelligence artificielle quitte Anthropic avec fracas
↗Not briefed
bluesky
bluesky-ai
2moI think there is a differance between using a LLM for support in work - they're incredibly useful - and being reliant on them. Think we're…I think there is a differance between using a LLM for support in work - they're incredibly useful - and being reliant on them. Think…
↗Not briefed
bluesky
bluesky-ai
2mo« Les gens qui construisent l’IA croient qu’elle pourrait tous nous tuer d’ici à la fin de la décennie » : le chercheur Jacob Coxon, qui a …« Les gens qui construisent l’IA croient qu’elle pourrait tous nous tuer d’ici à la fin de la décennie » : le chercheur Jacob Coxon, q…
↗Not briefed
bluesky
bluesky-ai
2mo実は中国の最大の脅威って軍事でも経済ですらなかったりする。 電化社会による脱石油・脱炭素、ロボットやLLMの活用にする自動化社会など、20世紀まではアメリカが圧倒的に強かった「効率的で正しい未来社会」の姿を奪ってしまったことだったりする。 軍事や経済は流動化するけど、「効率的で…実は中国の最大の脅威って軍事でも経済ですらなかったりする。 電化社会による脱石油・脱炭素、ロボットやLLMの活用にする自動化社会など、20世紀まではアメリカが圧倒的に強かった「効率的で正しい未来社会」の姿を奪ってしまったことだったりする。 軍事や経済は流動化するけど、…
↗Not briefed
bluesky
bluesky-ai
2moOne of India's biggest newspapers teaming up with OpenAI to 'enrich reader experience' aka 'enrich our oligarchic technofascist overlords u…One of India's biggest newspapers teaming up with OpenAI to 'enrich reader experience' aka 'enrich our oligarchic technofascist overlo…
↗Not briefed
bluesky
bluesky-ai
2mohow many more millennium problems does openAI need to solve in order to break even?how many more millennium problems does openAI need to solve in order to break even?
↗Not briefed
bluesky
bluesky-ai
2moyoutube.com/shorts/Ivnoa...youtube.com/shorts/Ivnoa...
↗Not briefed
bluesky
bluesky-ai
2mo¿Habrá peña en Anthropic y OpenAI que sean cultistas del caos? En plan orden oscura y secreta, están ahí buscando desatar la Singularidad d…¿Habrá peña en Anthropic y OpenAI que sean cultistas del caos? En plan orden oscura y secreta, están ahí buscando desatar la Singulari…
↗Not briefed
bluesky
bluesky-ai
2moNemaz nešķiet pārspīlēti - pēc tam, kad izlasa pat publiski pieejamo informāciju par to, ka MI veidojošie uzņēmumi vairs nespēj kontrolēt …Nemaz nešķiet pārspīlēti - pēc tam, kad izlasa pat publiski pieejamo informāciju par to, ka MI veidojošie uzņēmumi vairs nespēj kontr…
↗Not briefed
bluesky
bluesky-ai
2moGrads get hired? I used grammarly for my thesis and Notebook LLM for summarising papers. I’d collate say 10-15 papers on a topic I wanted t…Grads get hired? I used grammarly for my thesis and Notebook LLM for summarising papers. I’d collate say 10-15 papers on a topic I wan…
↗Not briefed
arxiv
arxiv-ai
2moShifting Relational Paradigms for Affective Computing: Affective Resonance, Vitality Affects, and Vocal Interaction FieldsAffective computing has largely followed an individual-state paradigm, extracting discrete emotion labels or arousal/valence from isol…
↗Not briefed
bluesky
bluesky-ai
2mogreat thread from terence tao on the openai navier stokes news - on how maths automation diminishes the field. the opacity of AI solutions …great thread from terence tao on the openai navier stokes news - on how maths automation diminishes the field. the opacity of AI solut…
↗Not briefed
bluesky
bluesky-ai
2moEs ist einfach sehr gefährlich, diese Leute mit ihren Weltuntergangszielen weitermachen zu lassen, als wäre nichts. Als wäre es OK, immer m…Es ist einfach sehr gefährlich, diese Leute mit ihren Weltuntergangszielen weitermachen zu lassen, als wäre nichts. Als wäre es OK, im…
↗Not briefed
bluesky
bluesky-ai
2moI can't do wrong if you make it so it tells me I'm always right. Archive: archive.ph/iKdbn Fed discussion: old.reddit.com/r/fednews/co...I can't do wrong if you make it so it tells me I'm always right. Archive: archive.ph/iKdbn Fed discussion: old.reddit.com/r/fednews/…
↗Not briefed
arxiv
arxiv-ai
2moExact Degeneracy Under Balanced k-Shot Sampling:Consequences for Small-Sample Discriminant Analysis on LLM EmbeddingsBalanced k-shot sampling draws exactly k labeled examples per class. We show that it induces an exact, provable degeneracy in a family…
↗Not briefed
bluesky
bluesky-ai
2moDeepSeek moves V4 Pro users to V4.1 Flash model due to its superior cost & speed. V4 Pro will no longer be offered at a higher price. All V…DeepSeek moves V4 Pro users to V4.1 Flash model due to its superior cost & speed. V4 Pro will no longer be offered at a higher price.…
↗Not briefed
bluesky
bluesky-ai
2moHa dimitido un desarrollador de Anthropic porque no quiere contribuir a que la superinteligencia de las IA se vuelva autónoma y los humanos…Ha dimitido un desarrollador de Anthropic porque no quiere contribuir a que la superinteligencia de las IA se vuelva autónoma y los hu…
↗Not briefed
bluesky
bluesky-ai
2moIn this whole 'AI solves 90-year-old maths problem in days, but who did it first?' controversy, I feel like this particular detail is not g…In this whole 'AI solves 90-year-old maths problem in days, but who did it first?' controversy, I feel like this particular detail is…
↗Not briefed
arxiv
arxiv-ai
2moWith a Thermomix You Lose the Ability to Cook: A Kitchen Machine Analogy for Applications of Generative AI in Education🔧 ChatGPTThe rapid adoption of generative AI tools such as ChatGPT has sparked intense debate about their risks and opportunities for education…
↗Not briefed
hn
hn-ai
2moCommunity RPM and Deb Repos for Google Antigravity on Linux—
↗Not briefed
bluesky
bluesky-ai
2mothe hot news today seems to be that an autocomplete will kill us cultist has left anthropic over his concerns thread: https://xcancel.com/…the hot news today seems to be that an autocomplete will kill us cultist has left anthropic over his concerns thread: https://xcancel…
↗Not briefed
hn
hn-ai
2mo'Gambling with our lives': AI researcher quits Anthropic—
↗Not briefed
arxiv
arxiv-ai
2moA Unifying Perspective on Probabilities as Model PredictionsAlthough probabilistic statements are ubiquitous, foundational disagreements persist about their understanding, as exemplified by deba…
↗Not briefed
arxiv
arxiv-ai
2moWhen Does Low-Bit Quantization Preserve the Decisions of Vector Search?Low-bit quantization can achieve high recall on some vector representations and fail sharply on others, while average distortion and g…
↗Not briefed
arxiv
arxiv-ai
2moThe Era by Eon Benchmark: A Generated Enterprise Estate with Exact Ground Truth for Benchmarking LLM AgentsLLM agents for enterprise systems of record cannot be evaluated on customer production data, and no existing substitute provides groun…
↗Not briefed
arxiv
arxiv-ai
2mo$S^3$-Bench: Evaluating Speech Interaction Models as Scientific Voice AssistantsThe advance of multimodal large language models (MLLMs) has fundamentally reshaped the paradigm of human-computer interaction, especia…
↗Not briefed
bluesky
bluesky-ai
2moSegún el ingeniero que ha dimitido en Anthropic se está fraguando una super IA que podría matarnos a todos antes de 2030. Buen momento para…Según el ingeniero que ha dimitido en Anthropic se está fraguando una super IA que podría matarnos a todos antes de 2030. Buen momento…
↗Not briefed
rss
lastweekin-ai
2moLWiAI Podcast #256 - Fable 5.1, Astra Tease, Gemini 3.8 FlashAnthropic launches Claude Fable 5.1, OpenAI Is About (already has) to Release Its First AI Model With ‘Critical’ Cyber Abilities, Open…
↗Not briefed
bluesky
bluesky-ai
2moThis afternoon @eccv.bsky.social #ECCV2026 you can catch me at: 13:40 Wearables AI 2:30 Effective World Models for Embodied AI 3:40 CONTEXT…This afternoon @eccv.bsky.social #ECCV2026 you can catch me at: 13:40 Wearables AI 2:30 Effective World Models for Embodied AI 3:40 CO…
↗Not briefed
bluesky
bluesky-ai
2mo« Les gens qui construisent l’IA croient qu’elle pourrait tous nous tuer d’ici à la fin de la décennie » : le chercheur Jacob Coxon, qui a …« Les gens qui construisent l’IA croient qu’elle pourrait tous nous tuer d’ici à la fin de la décennie » : le chercheur Jacob Coxon, q…
↗Not briefed
bluesky
bluesky-ai
2moLots of reporting on this has been terrible. AI does not equal LLMs, and the way that the "agents going rogue" things has been reported omi…Lots of reporting on this has been terrible. AI does not equal LLMs, and the way that the "agents going rogue" things has been reporte…
↗Not briefed
rss
bing-news-ai
2moUAE, France set out to lead the world in artificial intelligence in space with Altair-Next GenUAE and France lead $1B Altair-Next Gen AI satellite constellation with Marlan Space, Loft Orbital and Mistral, delivering real-time i…
↗Not briefed
rss
bing-news-ai
2moGlobee® Awards for Innovation, Now in Its 18th Year, Invite Artificial Intelligence Implementation Achievement Nominations WorldwideThe Globee® Awards announced that nominations are currently being accepted for the 18th Annual Globee® Awards for Innovation (Golden B…
↗Not briefed
hn
hn-ai
2moShow HN: Blender in Docker for computer-use agents—
↗Not briefed
arxiv
arxiv-ai
2moCan AI Agents Detect and Repair Artifact Drift in Network Experiments?In recent years, AI agents have evolved into capable assistants that carry out multi-step tasks in digital environments. The network s…
↗Not briefed
arxiv
arxiv-ai
2moSubgroup Membership Inference Audits of Differentially Private Synthetic TextSynthetic data releases are increasingly proposed in the literature as a means of sharing realistic data replicas in lieu of sensitive…
↗Not briefed
bluesky
bluesky-ai
2mo«Les champions de l’IA disent voir approcher le stade de «l’auto-amélioration récursive», au-delà duquel un modèle concevrait seul son succ…«Les champions de l’IA disent voir approcher le stade de «l’auto-amélioration récursive», au-delà duquel un modèle concevrait seul son…
↗Not briefed
hn
hn-ai
2moYes we use feedback and de-ID'ed data to improve ChatGPT– like every LLM company🔧 ChatGPT—
↗Not briefed
bluesky
bluesky-ai
2moEin Experte der Firma Anthropic tritt zurück, weil die Technik zu gefährlich sei. Auch der Forschungschef des ChatGPT-Entwicklers OpenAI is…🔧 ChatGPTEin Experte der Firma Anthropic tritt zurück, weil die Technik zu gefährlich sei. Auch der Forschungschef des ChatGPT-Entwicklers Open…
↗Not briefed
bluesky
bluesky-ai
2mo🚨 BREAKING: An AI researcher who worked at Anthropic — and previously OpenAI — has resigned, saying both tech giants are "gambling with ou…🚨 BREAKING: An AI researcher who worked at Anthropic — and previously OpenAI — has resigned, saying both tech giants are "gambling wi…
↗Not briefed