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
4dCertified, Not Explained: Exact Token Erasure in Gated Vision State Space ModelsVision models built on Mamba are often explained through their selective step size Delta, on the assumption that a token the scan does…
↗Not briefed
bluesky
bluesky-ai
4dThe asymmetry is real: trusting an LLM when you shouldn't creates harm at the point of use, over and over. Underestimating it just means yo…The asymmetry is real: trusting an LLM when you shouldn't creates harm at the point of use, over and over. Underestimating it just mea…
↗Not briefed
bluesky
bluesky-ai
4dUsing an LLM / agent to summarise large sets of PDFs, and uploading PDFs to a Anthropic or OpenAI, are two different things, with distinct …Using an LLM / agent to summarise large sets of PDFs, and uploading PDFs to a Anthropic or OpenAI, are two different things, with dist…
↗Not briefed
arxiv
arxiv-ai
4dWhich AI Use Cases Deserve a Pilot? A Risk and Value Framework for Funding, Scoping, and Scaling AIA composite score can rank AI use cases while leaving the permitted operating scope and the evidence required for expansion unspecifie…
↗Not briefed
rss
stratechery
4dObituariesThe demise of Jensen Huang was the end of a great many things, including Nvidia's long-term prospects, the AI bubble, and the exceptio…
↗Not briefed
hn
hn-ai
4dMay be data is the bottleneck for AGIIf you have seen pictures of our universe and how small we are, like invisible bugs, it's fair to assume our data is not big enou…
↗Not briefed
arxiv
arxiv-ai
4dFrom Recursion to Convolution: Parallel Self-Improvement for Multi-Agent SystemsMulti-agent self-improvement enables agents to improve through interaction and feedback. However, sequential dependencies across agent…
↗Not briefed
bluesky
bluesky-ai
4dWhen Anthropic goes public on Wall Street, the new pressure to deliver profits will test a commitment it has made since its founding: to pr…When Anthropic goes public on Wall Street, the new pressure to deliver profits will test a commitment it has made since its founding:…
↗Not briefed
bluesky
bluesky-ai
4dAnthropic recently rolled out Claude Mythos 5.2 to selected partners and Fable 5.2 to regular users, with improvements to coding and reason…Anthropic recently rolled out Claude Mythos 5.2 to selected partners and Fable 5.2 to regular users, with improvements to coding and r…
↗Not briefed
rss
wired-ai
4dNvidia Was Already in Trouble Before the Jensen Huang MessAI competitors are catching up to Nvidia—and poised to take advantage of the chaos.
↗Not briefed
arxiv
arxiv-ai
4dComfort as a Constraint or as a Cost in Reinforcement Learning and Model Predictive Building ControlA building controller that keeps a zone within comfort bounds at minimum energy can treat the bounds as a soft constraint, by adding t…
↗Not briefed
arxiv
arxiv-ai
4dAccelerate Bench: Evaluating Agents for Application-Level Program OptimizationApplication optimization requires coding agents to discover useful changes and coordinate them across a complete program while preserv…
↗Not briefed
bluesky
bluesky-ai
4dLe bon IA slop dopé à la décompilation faite par Claude ou Anthropic, hmmmm miam miam on adore ça 🥰🥰🥰 (non)Le bon IA slop dopé à la décompilation faite par Claude ou Anthropic, hmmmm miam miam on adore ça 🥰🥰🥰 (non)
↗Not briefed
hn
hn-ai
4dTwo Dogmas of Agent Communication—
↗Not briefed
arxiv
arxiv-ai
4dMarkov-Equivalent Conditional Independence Graphs on Path SpacesCausal discovery for continuous-time stochastic processes with latent confounding typically faces a trade-off between (i) approaches t…
↗Not briefed
arxiv
arxiv-ai
4dRe-Reading Is Not Free: Token Traces and Cache-Aware Energy Estimation for LLM Agents in Software EngineeringAs large language models (LLMs) evolve from conversational assistants into autonomous agents embedded in software development workflow…
↗Not briefed
hn
hn-ai
4dShow HN: Interpretability-guided weight edits for Gemma 270M, no training—
↗Not briefed
bluesky
bluesky-ai
4d#andrewcampbell #britishartist #art #subvert #meta #agitate #algorithms #provoke #X #attack #openai #resist #ai #disrupt #instagram #artfor…#andrewcampbell #britishartist #art #subvert #meta #agitate #algorithms #provoke #X #attack #openai #resist #ai #disrupt #instagram #a…
↗Not briefed
bluesky
bluesky-ai
4dJust as in the case of OpenAI and Sam Altman’s return as CEO: when everyone, including employees, have an equity stake there will be a maj…Just as in the case of OpenAI and Sam Altman’s return as CEO: when everyone, including employees, have an equity stake there will be…
↗Not briefed
arxiv
arxiv-ai
4dIt Depends on the Task: Rethinking Rhetorical Segmentation of Dutch Judicial DecisionsClassifying the material facts, submissions, judicial reasoning, and the court's decision (rhetorical roles) in case law is a prerequi…
↗Not briefed
bluesky
bluesky-ai
4dMicrosoft's present market cap is ~$4T, so on this story (which seems plausible) the markets are pricing in a 1/3 chance that Anthropic pul…Microsoft's present market cap is ~$4T, so on this story (which seems plausible) the markets are pricing in a 1/3 chance that Anthropi…
↗Not briefed
bluesky
bluesky-ai
4dAnthropic and OpenAI are eagerly pushing to expand the use of agents—but OpenAI is also betting on a number of other means of monetising it…Anthropic and OpenAI are eagerly pushing to expand the use of agents—but OpenAI is also betting on a number of other means of monetisi…
↗Not briefed
rss
techmeme
4dBaidu reports Q3 revenue down 1% YOY to ~$4.6B, its sixth straight quarterly sales decline, and net income of ~$210M; AI businesses are now 50%+ of its revenue (Bloomberg)Bloomberg: Baidu reports Q3 revenue down 1% YOY to ~$4.6B, its sixth straight quarterly sales decline, and net income of ~$210M; AI bu…
↗Not briefed
rss
marktechpost
4dMeet Meta’s JiTTests: LLM-Generated Tests Designed to Fail Before Risky Code LandsMeta has moved Catching JiTTests into production. The LLM system writes tests designed to fail on risky diffs, and caught 74 privacy S…
↗Not briefed
arxiv
arxiv-ai
4dPractical Semantic Communication for General-Purpose Networks: Compression of High-Dimensional Embeddings Under QuantizationThis paper introduces a practical framework for semantic communication over wireless channels, focusing on the compression of high-dim…
↗Not briefed
arxiv
arxiv-ai
4dScenario-Based Risk Certificates for Multi-Property Vulnerability Discovery in Parametric MDPsSafety-critical systems such as network protocols and distributed algorithms depend on environmental parameters whose exact values are…
↗Not briefed
bluesky
bluesky-ai
4d‼️ Japan's cyber response team has found about 1,000 servers it suspects are bases for AI-run attacks on websites worldwide. Most run AI p…‼️ Japan's cyber response team has found about 1,000 servers it suspects are bases for AI-run attacks on websites worldwide. Most run…
↗Not briefed
arxiv
arxiv-ai
4dMotor Imagery Classification Using Hybrid Cross-Subject Class Mixup and Selective Attention🔧 ModalMotor imagery (MI) enables a brain-computer interface (BCI) to decode movement intentions without overt movement. Hybrid electroenceph…
↗Not briefed
arxiv
arxiv-ai
4dartist: A Python package for differentiable ray tracing and gradient-based optimization of solar tower power plant operation🔧 FLUXSolar tower power plants provide dispatchable, renewable energy by concentrating sunlight with thousands of mirrors, called heliostats…
↗Not briefed
arxiv
arxiv-ai
4dMolReps: Taking SELFIES, Making SMILES, Teaching LLMsThe same molecule can be expressed in multiple, syntactically distinct representations, yet it remains unclear whether large language…
↗Not briefed
bluesky
bluesky-ai
4dDPro、広告運用自動化AIエージェント「AIパクくん」をユーザーに無償配布──リサーチからCR生成・入稿・PDCAまでを完全自動化 当初50万円で販売予定だったAIエージェントを 「AIパクくん」は、DProの膨大な広告データを活用し、広告運用の全行程を自動化するAIエージェ…DPro、広告運用自動化AIエージェント「AIパクくん」をユーザーに無償配布──リサーチからCR生成・入稿・PDCAまでを完全自動化 当初50万円で販売予定だったAIエージェントを 「AIパクくん」は、DProの膨大な広告データを活用し、広告運用の全行程を自動化するA…
↗Not briefed
arxiv
arxiv-ai
4dEDITORS Know Your Style! Editing LoRA Subspaces for Stylistic Attribution and ImitationDisentangling writing style from semantic content is a fundamental challenge in literary text modeling. Style-content entanglement cau…
↗Not briefed
arxiv
arxiv-ai
4dDetourBench: Trajectory-Level Evaluation of Interactive Agents with Exact Minimum-Action BoundsBenchmarks judge language-model agents mainly by whether they reach the correct final state, not by how they got there. Endpoint scori…
↗Not briefed
arxiv
arxiv-ai
4dVision-Language Model-Based Phase-OTDR Event ClassificationPhase-sensitive optical time-domain reflectometry (Phase-OTDR) enables distributed vibration sensing, but reliable event classificatio…
↗Not briefed
bluesky
bluesky-ai
4d그냥 수능날 전국민 휴무일로 정하고 LLM 서비스 중단하면 안되나 1일 AI프리데이ㅇㅂㅇ)9그냥 수능날 전국민 휴무일로 정하고 LLM 서비스 중단하면 안되나 1일 AI프리데이ㅇㅂㅇ)9
↗Not briefed
bluesky
bluesky-ai
4dAnyone, and I do mean anyone who, unprompted during a face-to-face conversation, says "I asked ChatGPT/some other LLM" is not a person I wa…🔧 ChatGPTAnyone, and I do mean anyone who, unprompted during a face-to-face conversation, says "I asked ChatGPT/some other LLM" is not a person…
↗Not briefed
arxiv
arxiv-ai
4dEvidence Diet: When Graph Evidence Helps LLMs Rank Competing CandidatesWhen several candidates have graph support, does it matter which candidate receives which supporter and path counts? We study how LLM…
↗Not briefed
arxiv
arxiv-ai
4dHierarchical Regime-Conditioned Dynamics for Spatiotemporal GraphsForecasting the behavior of spatio-temporal systems often requires more than accurate predictions: models should uncover the dynamical…
↗Not briefed
hn
hn-ai
4dShow HN: Simplex – Self-hosting AArch64 ELF compiler built on a GrapheneOS phoneSimplex is a compiler that strikes directly at the architectural heresy of modern computing, spitting out pure, self-contained ELF64 A…
↗Not briefed
rss
techmeme
4dApple is launching a video podcasts section on Apple TV and its app, initially for developer testing, and plans to increase video podcasts on its Podcasts app (New York Times)New York Times: Apple is launching a video podcasts section on Apple TV and its app, initially for developer testing, and plans to inc…
↗Not briefed
arxiv
arxiv-ai
4dProcReverser: Multi-Agent Black-Box Microarchitectural Reverse Engineering for Processor Vulnerability AnalysisWhen processor implementation details are unavailable, black-box microarchitectural reverse engineering helps security researchers ide…
↗Not briefed
arxiv
arxiv-ai
4dConversing with a Conversational Agent: A Real Conversation, Not a Simulated OneThe claim that interactions with artificial conversational agents merely simulate human conversation rests on a philosophical distinct…
↗Not briefed
arxiv
arxiv-ai
4dMedical question answering collapses in Hausa: a pre-registered pilot on three small open modelsA system that reads nothing and always answers the most frequent letter scores \BaselineP{} on our Hausa medical multiple-choice set.…
↗Not briefed
arxiv
arxiv-ai
4dCuRSE-Bench: A Benchmark for Probing Cultural Blind Spots in Text-to-Image Safety EvaluationText-to-image (T2I) safety evaluation typically assigns each prompt or image a single culture-blind harm label, so content that is har…
↗Not briefed
hn
hn-ai
4dBest Agent Gateways—
↗Not briefed
arxiv
arxiv-ai
4dSurprise-Gated Memory for Live Speaker DiarizationWe transpose the Surprise-Gated Memory recipe, a frozen encoder plus a calibrated novelty signal that gates a one-shot episodic store…
↗Not briefed
rss
bing-news-ai
4dAI could boost economic growth but disrupt jobs: Fed Chair Warsh flags impact of super intelligenceFederal Reserve Chairman Kevin Warsh on Monday said the emergence of artificial intelligence and what he described as "super intellige…
↗Not briefed
rss
bing-news-ai
4dAI could boost economic growth but disrupt jobs: Fed Chair Warsh flags impact of super intelligenceFederal Reserve Chairman Kevin Warsh on Monday said the emergence of artificial intelligence and what he described as "super intellige…
↗Not briefed
rss
bing-news-ai
4dAI could boost economic growth but disrupt jobs: Fed Chair Warsh flags impact of super intelligenceFederal Reserve Chairman Kevin Warsh on Monday said the emergence of artificial intelligence and what he described as "super intellige…
↗Not briefed
rss
bing-news-ai
4dAI could boost economic growth but disrupt jobs: Fed Chair Warsh flags impact of super intelligenceFederal Reserve Chairman Kevin Warsh on Monday said the emergence of artificial intelligence and what he described as "super intellige…
↗Not briefed