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
3dPitching a script where the Bad Guy is a leather jacket wearing tech CEO threatening to destabilise governments with mass LLM-generated dis…Pitching a script where the Bad Guy is a leather jacket wearing tech CEO threatening to destabilise governments with mass LLM-generate…
↗Not briefed
bluesky
bluesky-ai
3dThis is Claude this is about Claude, and the answer is to strip apart Anthropic and not use the product they create at all.This is Claude this is about Claude, and the answer is to strip apart Anthropic and not use the product they create at all.
↗Not briefed
bluesky
bluesky-ai
3dteam schumer & do nothing centrists are in part trying to fasttrack BAAJA so they can get in on the anthropic IPO btwteam schumer & do nothing centrists are in part trying to fasttrack BAAJA so they can get in on the anthropic IPO btw
↗Not briefed
bluesky
bluesky-ai
3dYou just wait until OpenAI releases Red Motivation and Blue Motivation updatesYou just wait until OpenAI releases Red Motivation and Blue Motivation updates
↗Not briefed
hn
hn-ai
3dPeer, a skill that makes Claude answer like a colleague, tested blind—
↗Not briefed
hn
hn-ai
3dShow HN: Llx – Fast, Efficient GPT-6.2 Sol and Astra Coding via ChatGPT🔧 ChatGPT—
↗Not briefed
bluesky
bluesky-ai
3d“The report uncovered a strong tendency for AI models to be sycophantic, meaning the model is more interested in pleasing the user than sta…“The report uncovered a strong tendency for AI models to be sycophantic, meaning the model is more interested in pleasing the user tha…
↗Not briefed
bluesky
bluesky-ai
3dOpenAI releases GPT-6.2 Astra to regain performance lead from Claude Fable 5.2OpenAI releases GPT-6.2 Astra to regain performance lead from Claude Fable 5.2
↗Not briefed
bluesky
bluesky-ai
3dAs Wall Street waits for earnings from the “central bank of AI,” Nvidia is flashing a warning on earnings quality. The Beneish model, which…As Wall Street waits for earnings from the “central bank of AI,” Nvidia is flashing a warning on earnings quality. The Beneish model,…
↗Not briefed
arxiv
arxiv-ai
3dIncremental Roseland: Batch-wise Landmark Diffusion via Kernelized Sequential Karhunen--Loève UpdateLandmark diffusion (Roseland) computes a diffusion-map-style spectral embedding from the singular value decomposition of a degree-norm…
↗Not briefed
hn
hn-ai
3dHow I Got My Laser Eye Injury (2024)—
↗Not briefed
bluesky
bluesky-ai
3dI dunno, simple unskippable instructions like this can cascade into catastrophic outcomes.I dunno, simple unskippable instructions like this can cascade into catastrophic outcomes.
↗Not briefed
bluesky
bluesky-ai
3dso it needs to be >=21 days from S-1 to IPO. Anthropic seems to be counting from their first *draft*. They release final prices *tomorrow* …so it needs to be >=21 days from S-1 to IPO. Anthropic seems to be counting from their first *draft*. They release final prices *tomor…
↗Not briefed
bluesky
bluesky-ai
3dOpenAI can't really rest on the laurels of being frontier lab 2 now, since there are several labs that want the position. This isn't a grea…OpenAI can't really rest on the laurels of being frontier lab 2 now, since there are several labs that want the position. This isn't a…
↗Not briefed
bluesky
bluesky-ai
3dWait, you want people to give the fucking LLM access to their bank accounts? HAHAHAHAHAHAHAHAHAHAHAHAHAHAWait, you want people to give the fucking LLM access to their bank accounts? HAHAHAHAHAHAHAHAHAHAHAHAHAHA
↗Not briefed
bluesky
bluesky-ai
3dAnthropic CEO two weeks from now: "Yeahhh soooo, remember the whole airport thing????"Anthropic CEO two weeks from now: "Yeahhh soooo, remember the whole airport thing????"
↗Not briefed
hn
hn-ai
3dCliHelpLine – 2-line comment-based –help for scripts—
↗Not briefed
bluesky
bluesky-ai
3dThe owners of the Elvis Presley catalogue are suing Anthropic over unfair use of the King’s biggest hits. What happens next?The owners of the Elvis Presley catalogue are suing Anthropic over unfair use of the King’s biggest hits. What happens next?
↗Not briefed
bluesky
bluesky-ai
3dfuturism.com/future-socie... ““Overall, I think there’s a very real possibility — maybe 20 percent — that within a year, AI misalignment is…futurism.com/future-socie... ““Overall, I think there’s a very real possibility — maybe 20 percent — that within a year, AI misalignme…
↗Not briefed
bluesky
bluesky-ai
3d"Super" intelligence Gift link: www.bloomberg.com/news/article..."Super" intelligence Gift link: www.bloomberg.com/news/article...
↗Not briefed
arxiv
arxiv-ai
3dRFBench: A Simulator-Grounded Multimodal Benchmark for Passive RF Layout Design and UnderstandingWhile multimodal language models can interpret diagrams and generate structured outputs, their capacity to connect passive radio-frequ…
↗Not briefed
hn
hn-ai
3dGPT-6.2 Sol Moves the Pareto Frontier in Artificial Analysis Score vs. Cost—
↗Not briefed
hn
hn-ai
3dAnthropic Paid Rubio's Ex-Chief of Staff's Firm $740k.it Lobbied His NSC—
↗Not briefed
bluesky
bluesky-ai
3dLike, Anthropic seems to be cutting down on game cloning by giving agents permission to kill non-Anthropic processes on a computer. Could t…Like, Anthropic seems to be cutting down on game cloning by giving agents permission to kill non-Anthropic processes on a computer. Co…
↗Not briefed
bluesky
bluesky-ai
3dthe timing is wild considering polymarket is 98% sure anthropic will list thursdaythe timing is wild considering polymarket is 98% sure anthropic will list thursday
↗Not briefed
hn
hn-ai
3dShow HN: Eso-box: turn a spare laptop into an always-on Claude Code serverMy mind is all always all over the place (ADD?) and often I would have ideas I wanted to test/implement while I was on the go&#x2…
↗Not briefed
arxiv
arxiv-ai
3dTracking Representations in LLM Reasoning Traces for Iterated GamesLarge language model (LLM) agents achieve state-of-the-art performance across domains through chain-of-thought (CoT) reasoning. Despit…
↗Not briefed
rss
techcrunch-ai
3dGoogle expands access to Gemini Tasks, which can help with calls, purchasesGemini's Task mode can help users coordinate calls, log into websites, and even make purchases on the user's behalf.
↗Not briefed
hn
hn-ai
3dLaunch HN: StarSling (YC P25) – Self-improving CI for agent-volume reposWe’re Yonas and Daniel and we’re building StarSling (<a href="https:&#x2F;&#x2F;starsling.dev">https:&#x2F;&#x2F;starsling.dev</a>), f…
↗Not briefed
bluesky
bluesky-ai
3dPangram's tech might survive no matter how good the LLMs get, it's not really dependent on a weakness in the LLM. They have to do more trai…Pangram's tech might survive no matter how good the LLMs get, it's not really dependent on a weakness in the LLM. They have to do more…
↗Not briefed
bluesky
bluesky-ai
3domg they all hate it xxcancel.com/theo/article...omg they all hate it xxcancel.com/theo/article...
↗Not briefed
arxiv
arxiv-ai
3dHardwareDx: A Foundational Graph Model System for Datacenter Machine RepairsThe rapid scale-up of AI infrastructure has made datacenter reliability a cornerstone of the cloud business economy. Since hardware fa…
↗Not briefed
arxiv
arxiv-ai
3dWhich Tokens, in What Order, and How Fast? A Theoretical Analysis of On-Policy Distillation ObjectivesOn-policy distillation (OPD) trains a student language model to match a teacher model based on its own generations. When using OPD, on…
↗Not briefed
rss
techmeme
3dGoogle updates Gemini Privacy Hub to extend its retention period for temporary chats and chats with history turned off from 72 hours to around two months (Jagmeet Singh/TechCrunch)Jagmeet Singh / TechCrunch: Google updates Gemini Privacy Hub to extend its retention period for temporary chats and chats with histor…
↗Not briefed
hn
hn-ai
3dShow HN: Fardesk – fast iPhone app to manage your Claude sessions—
↗Not briefed
arxiv
arxiv-ai
3dQDA: Closed-Loop LLM Quantization Design with AgentsPost-training quantization is the standard way to serve large language models at low cost. However, the best quantization recipe diffe…
↗Not briefed
arxiv
arxiv-ai
3dChorus: Shared N-gram Embeddings with Layer-wise AdaptationLanguage models can gain capacity at little computational cost from n-gram embedding tables: large tables of learned vectors for short…
↗Not briefed
arxiv
arxiv-ai
3dWhere the Tokens Go: Heat, Entropy and Error in Agentic CodingCoding agents are billed by the token, yet software engineering says nothing about what a token actually buys. Here we introduce a the…
↗Not briefed
hn
hn-ai
3dGPT-6.2 Astra scores 100% on ARC-AGI-3—
↗Not briefed
arxiv
arxiv-ai
3dAmbient LLMs: Data-Efficient Supervised Fine-tuning via Temperature MatchingCollecting large-scale, diverse, and high-quality post-training datasets for autoregressive language models is challenging. Consequent…
↗Not briefed
bluesky
bluesky-ai
3di know all about the discourse around localized and ethical LLM too, i know thats dicier but my personal stance is master's tools dont dism…i know all about the discourse around localized and ethical LLM too, i know thats dicier but my personal stance is master's tools dont…
↗Not briefed
arxiv
arxiv-ai
3dSpool: Low-Latency Streaming Inference for Hybrid Vision-Language Models🔧 vLLMApplications such as visual accessibility agents and robotics require answering questions about recent video content promptly, over a…
↗Not briefed
arxiv
arxiv-ai
3dTraining Language Models for Introspective OversightLanguage models increasingly exhibit behaviors that are difficult to detect in transcripts, such as reward hacking, deception, and syc…
↗Not briefed
arxiv
arxiv-ai
3dStringQFT: A Multiuse Textbook Resource for AI-Assisted Theoretical PhysicsWe present StringQFT as a textbook collection for string constructions of quantum field theory and as infrastructure for AI-assisted t…
↗Not briefed
rss
theverge-ai
3dGoogle’s Spark AI agent becomes ‘Tasks’ — and now it’s freeYou can now switch from the “Chat” toggle to “Tasks.” | Image: Google Google is rebranding its Spark AI agent as "Tasks" and making…
↗Not briefed
hn
hn-ai
3dI hate GPT-6.2 Astra—
↗Not briefed
arxiv
arxiv-ai
3dNeural Circuit Architectural Priors for Rat LocomotionAnimals generate behavior through continuous interactions between their neural circuits, bodies, and environment. Embodied simulations…
↗Not briefed
arxiv
arxiv-ai
3dAn inexact stochastic variance-reduced gradient method for constrained convex finite-sum optimizationIn this paper, we study a class of constrained convex finite-sum optimization problems. In many practical applications, all or some of…
↗Not briefed
arxiv
arxiv-ai
3dDo LLM Agents Still Need to Be Told How to Self-Improve?LLM agents increasingly drive their own learning. Self-improvement methods delegate various components of model optimization to the mo…
↗Not briefed
arxiv
arxiv-ai
3dWhat Does a Bias Head Measure? Non-Identifiability in Head-Level Bias AttributionMethods that localise social bias in large language models to individual attention heads and remove it by ablation are validated by th…
↗Not briefed