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
29dJudge orders OpenAI to turn over hiring records in Apple trade secrets caseJudge orders OpenAI to turn over hiring records in Apple trade secrets case
↗Not briefed
bluesky
bluesky-ai
29dHe used ChatGPT to make the listening device that he put in OpenAI’s sensitive office? Bro sfstandard.com/2026/10/13/o...🔧 ChatGPTHe used ChatGPT to make the listening device that he put in OpenAI’s sensitive office? Bro sfstandard.com/2026/10/13/o...
↗Not briefed
hn
hn-ai
29dHeartbleed Encryption: Explained—
↗Not briefed
bluesky
bluesky-ai
29d"our safeguards have a false positive rate of 0.1%" sure anthropic"our safeguards have a false positive rate of 0.1%" sure anthropic
↗Not briefed
rss
techcrunch-ai
29dJudge orders OpenAI to turn over hiring records in Apple trade secrets caseA federal judge ruled that Apple can access OpenAI's interview materials involving former Apple employees, as well as private messages…
↗Not briefed
bluesky
bluesky-ai
29djust stumbled on an extremely low-budget slop site that runs Nordic news reports through some kind of LLM translator it seems to have been…just stumbled on an extremely low-budget slop site that runs Nordic news reports through some kind of LLM translator it seems to have…
↗Not briefed
arxiv
arxiv-ai
29dVerification Turns Capability Gaps into Routing Decisions: Small Language Models for Clinical DocumentationClinical note generation from doctor--patient dialogue is high-volume and hallucination-intolerant: a fabricated laboratory value can…
↗Not briefed
rss
techmeme
29dFederal prosecutors say a Saudi national was arrested over bugging an OpenAI executive conference room; OpenAI found the audio recorder around September 30 (The San Francisco Standard)The San Francisco Standard: Federal prosecutors say a Saudi national was arrested over bugging an OpenAI executive conference room; Op…
↗Not briefed
hn
hn-ai
29dGitslice, Source Control for Agents—
↗Not briefed
bluesky
bluesky-ai
29dMore details in the Anthropic post belowMore details in the Anthropic post below
↗Not briefed
bluesky
bluesky-ai
29dAnthropic now states that Claude can code better than its engineers but humans are still valuable primarily to tell Claude “no.” This was …Anthropic now states that Claude can code better than its engineers but humans are still valuable primarily to tell Claude “no.” This…
↗Not briefed
hn
hn-ai
29dDeepSFetch: Open-source developer kit for web extraction and agent workflows—
↗Not briefed
hn
hn-ai
29dI just got our entire company's Claude Enterprise account suspended—
↗Not briefed
hn
hn-ai
29dTell HN: ChatGPT on macOS grants local computer access to cloud account🔧 ChatGPTI recently started a conversation on ChatGPT for macOS in "Work" mode with the "In the Cloud" execution option ena…
↗Not briefed
arxiv
arxiv-ai
29dFantastic Metrics and How to Find Them: Meta-Evaluating Metrics for Multi-lingual Style-Personalized GenerationMeta-evaluation of text generation metrics and LLM-judges is primarily conducted using metric-human correlations. However, prior work…
↗Not briefed
hn
hn-ai
29dJoyscroller – Automatically download your liked YouTube, TikTok and insta vids—
↗Not briefed
bluesky
bluesky-ai
29dOpenAI says one of its employees has been arrested after the artificial intelligence startup identified that the person had hidden a “sophi…OpenAI says one of its employees has been arrested after the artificial intelligence startup identified that the person had hidden a “…
↗Not briefed
bluesky
bluesky-ai
29dSmall artificial intelligence models may be only months away from being able to behave like a computer worm, copying themselves from one in…Small artificial intelligence models may be only months away from being able to behave like a computer worm, copying themselves from o…
↗Not briefed
arxiv
arxiv-ai
29dListQA: A Benchmark for Evaluating List-Formatted Factual Knowledge Retrieval in Large Language ModelsMany real-world questions demand not a single fact but an organized collection of them, yet existing factual knowledge benchmarks almo…
↗Not briefed
arxiv
arxiv-ai
29dFrom Qualitative Text to Integrated Evidence: The Parallel-Then-Integrated (PTI) ProcedureMixed methods research offers established strategies for integrating qualitative and quantitative evidence, but less guidance for inte…
↗Not briefed
bluesky
bluesky-ai
29dOpenAI Says Staffer Was Arrested for Hidden Recording Device 👀👀👀OpenAI Says Staffer Was Arrested for Hidden Recording Device 👀👀👀
↗Not briefed
hn
hn-ai
29dShow HN: Wonderhop – A daily game: guess what most people say, or what none doWonderhop is a game that runs in your browser, with no account to create. Every day there are 4 rounds of 7 questions like "Name…
↗Not briefed
bluesky
bluesky-ai
29dI would also be more worried if I saw any of these examples in which the misaligned agent loop had made use of the ability of the underlyin…I would also be more worried if I saw any of these examples in which the misaligned agent loop had made use of the ability of the unde…
↗Not briefed
bluesky
bluesky-ai
29dwww.reuters.com/world/us/fbi... OpenAI said that the FBI had arrested one of its employees in connection with a sophisticated listening dev…www.reuters.com/world/us/fbi... OpenAI said that the FBI had arrested one of its employees in connection with a sophisticated listenin…
↗Not briefed
hn
hn-ai
29dSaudi national working at OpenAI arrested for bugging executive conference room—
↗Not briefed
arxiv
arxiv-ai
29dInto the Unknown: Benchmarking AI agents on open scientific research problemsCan AI agents carry out research on open scientific problems? We introduce Into the Unknown, a benchmark that provides an agent with a…
↗Not briefed
bluesky
bluesky-ai
29dIf AI companies really feel like their tech could destroy the world, at least they could give it cooler, more threatening names. Aliens fin…If AI companies really feel like their tech could destroy the world, at least they could give it cooler, more threatening names. Alien…
↗Not briefed
arxiv
arxiv-ai
29dAgentCIScope: Grounded Contextual Integrity Evaluation for LLM AgentsLLM agents are becoming increasingly capable of handling daily workspace tasks by interacting with application environments through to…
↗Not briefed
arxiv
arxiv-ai
29dDivAlgo: A systematic evaluation of algorithmic diversity for code generation modelsCode generation models are actively being used in professional and educational settings. However, the competence of large language mod…
↗Not briefed
hn
hn-ai
29dOpenAI engineer arrested in bugging of executive conference room—
↗Not briefed
bluesky
bluesky-ai
29dGarage crypto rig runs 35B AI model A Reddit user scored a $35 crypto mining rig with 54GB of VRAM, running a 35B-parameter AI model at 30…Garage crypto rig runs 35B AI model A Reddit user scored a $35 crypto mining rig with 54GB of VRAM, running a 35B-parameter AI model…
↗Not briefed
bluesky
bluesky-ai
29dHad no idea that the AI sex dolls had been accepted in our government at this level. Hopefully they focus on their reading capabilities as…Had no idea that the AI sex dolls had been accepted in our government at this level. Hopefully they focus on their reading capabiliti…
↗Not briefed
bluesky
bluesky-ai
29dThis just doesn't make any sense. The amount of memory necessary to run an LLM, and the processing power required... Your fucking thermosta…This just doesn't make any sense. The amount of memory necessary to run an LLM, and the processing power required... Your fucking ther…
↗Not briefed
hn
hn-ai
29dShow HN: Open-source Mac menu bar app for Claude, Codex and Cursor—
↗Not briefed
bluesky
bluesky-ai
29dThe former OpenAI employee is a Saudi national according to the DOJ, so this could be an incident of state rather than corporate espionage.The former OpenAI employee is a Saudi national according to the DOJ, so this could be an incident of state rather than corporate espio…
↗Not briefed
bluesky
bluesky-ai
29d🚨 « According to a new study published by Anthropic, malicious AI agents could conspire with your pets (cats, dogs, goldfish, etc.) to suf…🚨 « According to a new study published by Anthropic, malicious AI agents could conspire with your pets (cats, dogs, goldfish, etc.) t…
↗Not briefed
bluesky
bluesky-ai
29di don't think the bankruptcy proves "AI/LLMs don't work" as a business model more that there is really stupid as hell shit during every spe…i don't think the bankruptcy proves "AI/LLMs don't work" as a business model more that there is really stupid as hell shit during ever…
↗Not briefed
bluesky
bluesky-ai
29dSo it was doing better than OpenAI or Anthropic.So it was doing better than OpenAI or Anthropic.
↗Not briefed
arxiv
arxiv-ai
29dCompiling Expert Procedures into Verified Programs: When a Small Model Is EnoughIn procedural tool-use domains, how much of the gap between base models is procedure rather than capability? We build an explicit prog…
↗Not briefed
arxiv
arxiv-ai
29dDemystifying On-Policy Distillation: Where Does the Teacher's Advantage Live?On-policy distillation (OPD) is increasingly studied as a cheaper alternative to reinforcement learning, yet it remains a form of know…
↗Not briefed
arxiv
arxiv-ai
29dTowards Training and Inference of Graph Neural Networks in Graph DatabasesGraph Neural Networks on large graphs commonly rely on in-memory graph storage with ad hoc data processing pipelines for both training…
↗Not briefed
arxiv
arxiv-ai
29dOnline Tokenization with Optimal Delay in Linear TimeTokenization is an essential preprocessing step in almost all large language models (LLMs). One of the most popular frameworks is BPE…
↗Not briefed
bluesky
bluesky-ai
29d"We used our models to understand the device and where it had come from" Did OpenAI do a marketing bit in the middle of an espionage accus…"We used our models to understand the device and where it had come from" Did OpenAI do a marketing bit in the middle of an espionage…
↗Not briefed
arxiv
arxiv-ai
29dMA-Bench: Measuring Operational Planning in Language ModelsHow well can a language model turn a large, incomplete and partly contradictory body of documents into a plan when no single correct a…
↗Not briefed
bluesky
bluesky-ai
29dThat's wild! Anyway, can I interest the good folks at OpenAI in this large wooden seal of the United States for them to hang up in their me…That's wild! Anyway, can I interest the good folks at OpenAI in this large wooden seal of the United States for them to hang up in the…
↗Not briefed
arxiv
arxiv-ai
29dGraph-Based Orchestration of Heterogeneous AI Models: A Control Node Approach to Composite IntelligenceWe present a graph-based framework for Composite AI systems that integrate multiple AI techniques to address complex problems. We desc…
↗Not briefed
arxiv
arxiv-ai
29dStudent Copula Markov NetworksThis paper extends the applicability of probabilistic graphical models to non Gaussian data. It establishes a characterization of the…
↗Not briefed
bluesky
bluesky-ai
29dMy BS detector went off on this part. It's like a subliminal advertisement for OpenAI LLM Corporate Surveillance. GTFO.My BS detector went off on this part. It's like a subliminal advertisement for OpenAI LLM Corporate Surveillance. GTFO.
↗Not briefed
bluesky
bluesky-ai
29dFBI arrests former OpenAI employee for planting listening device reut.rs/4hW7XTLFBI arrests former OpenAI employee for planting listening device reut.rs/4hW7XTL
↗Not briefed
bluesky
bluesky-ai
29dA 35-year-old Saudi OpenAI employee, Abdulrahman Alfozan, has been arrested for allegedly installing a covert audio recording device at the…A 35-year-old Saudi OpenAI employee, Abdulrahman Alfozan, has been arrested for allegedly installing a covert audio recording device a…
↗Not briefed