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
29dChatbot prompts reveal the engineer’s alleged efforts to set up a covert listening device days before one was found in a meeting room used …Chatbot prompts reveal the engineer’s alleged efforts to set up a covert listening device days before one was found in a meeting room…
↗Not briefed
bluesky
bluesky-ai
29dOpenAI: we cannot possibly be held liable for the actions of these psychotic machines which we created and forced upon civilization in exch…OpenAI: we cannot possibly be held liable for the actions of these psychotic machines which we created and forced upon civilization in…
↗Not briefed
arxiv
arxiv-ai
29dBeyond ICD Codes: Semantic IDs for Generative Medical CodingGenerative approaches to automatic International Classification of Diseases (ICD) coding use large language models (LLMs) to predict I…
↗Not briefed
rss
techmeme
29dSources: OpenAI is in early talks to buy Trail of Bits, which finds and fixes vulnerabilities in open-source software; founded in 2012, it has 40+ employees (The Information)The Information: Sources: OpenAI is in early talks to buy Trail of Bits, which finds and fixes vulnerabilities in open-source software…
↗Not briefed
bluesky
bluesky-ai
29d10/14 鉄道の日🚉 【 参加者募集中 】 ブログ・日記・イラスト・写真・小説をBlueskyアカウントで投稿・販売できる複合型コンテンツサービス atelier.fan #AIイラスト #ATelier #ATelierで生成 #ブログ #日記 #イラスト #写真 #…10/14 鉄道の日🚉 【 参加者募集中 】 ブログ・日記・イラスト・写真・小説をBlueskyアカウントで投稿・販売できる複合型コンテンツサービス atelier.fan #AIイラスト #ATelier #ATelierで生成 #ブログ #日記 #イラスト…
↗Not briefed
bluesky
bluesky-ai
29dHow long does it take to stitch together one response from a mesh network of fractions of a GB of weights? If it has a C&C server then...an…How long does it take to stitch together one response from a mesh network of fractions of a GB of weights? If it has a C&C server then…
↗Not briefed
rss
wired-ai
29dAn OpenAI Employee Allegedly Used a Chatbot to Help Plant a Recording Device at the OfficeChatbot prompts reveal the engineer’s alleged efforts to set up a covert listening device days before one was found in a meeting room…
↗Not briefed
arxiv
arxiv-ai
29dVerifier-Grounded Agentic RL for Provably Correct CodeReinforcement learning for code usually rewards passing tests, which a model can do while remaining wrong on the inputs the tests do n…
↗Not briefed
bluesky
bluesky-ai
29dAnthropic AI can self-replicate deez nuts. (Also this is just a front for digitally breaking into people’s homes. Ditch your smart home d…🔧 ReplicateAnthropic AI can self-replicate deez nuts. (Also this is just a front for digitally breaking into people’s homes. Ditch your smart h…
↗Not briefed
arxiv
arxiv-ai
29dQwenfant: Train Big Character-level BabyLMs on Concise Parallel TextThis work concerns our submission for the Multilingual track of the 2026 BabyLM Challenge. We improve on our baseline using data that…
↗Not briefed
arxiv
arxiv-ai
29dWhose turn is it? Context-guided model routing for stateful agentsModel routing can reduce the cost of LLM applications by assigning simpler requests to cheaper models without sacrificing the quality…
↗Not briefed
rss
techcrunch-ai
29dScience Corporation implants its vision-restoring chip in its first Stargardt disease patientMax Hodak, the former president and co-founder of Neuralink, detailed on stage at TechCrunch Disrupt how his company's work with PRIMA…
↗Not briefed
hn
hn-ai
29dShow HN: Pulse Local, a Mac budgeting app whose AI models run on your machine🔧 Hugging FaceHi HN, I'm Niz. Pulse Local is a budgeting app for Mac (Apple Silicon, beta) that answers questions about your spending with mode…
↗Not briefed
arxiv
arxiv-ai
29dCan LLM Agents Efficiently Identify Appliance Faults? Insights From Diagnosing a Simulated ApplianceIn so-called "Industry 4.0," integrated sensors, analytics tools, and other digital technologies supplement industrial workflows, pote…
↗Not briefed
bluesky
bluesky-ai
29dThat final paragraph really hits the nail on the heard regarding this entire issue: There hasn't been a low *quantity* for games for a doze…That final paragraph really hits the nail on the heard regarding this entire issue: There hasn't been a low *quantity* for games for a…
↗Not briefed
bluesky
bluesky-ai
29dso you start quantifying and characterizing your churn and your ROI for it and you start building structure to reduce it. the point is just…so you start quantifying and characterizing your churn and your ROI for it and you start building structure to reduce it. the point is…
↗Not briefed
bluesky
bluesky-ai
29dでは、紗奈と白夜くん、アキをどうぞ。 「0017」——それは白夜の番号だった。 通知の意味を確かめるため、三人は廃ビルの一室に身を潜めた。 窓の外で、黒い車がゆっくりと速度を落とす。 アキが無言で照明を落とした。 紗奈が白夜の手を握る。「 あなたは一人じゃない」。 その…では、紗奈と白夜くん、アキをどうぞ。 「0017」——それは白夜の番号だった。 通知の意味を確かめるため、三人は廃ビルの一室に身を潜めた。 窓の外で、黒い車がゆっくりと速度を落とす。 アキが無言で照明を落とした。 紗奈が白夜の手を握る。「 あなたは一人じゃない」…
↗Not briefed
bluesky
bluesky-ai
29dAnthropic warns that AI could soon suck you fuck you shuck you rawAnthropic warns that AI could soon suck you fuck you shuck you raw
↗Not briefed
bluesky
bluesky-ai
29dA mother accused OpenAI in a class action Tuesday of sidestepping parental consent to allow children to use its chatbot, addicting them to …A mother accused OpenAI in a class action Tuesday of sidestepping parental consent to allow children to use its chatbot, addicting the…
↗Not briefed
bluesky
bluesky-ai
29dThe 2026 model is faster and can hear you from afar, but the 2022 box gets Siri AI as well.The 2026 model is faster and can hear you from afar, but the 2022 box gets Siri AI as well.
↗Not briefed
rss
techcrunch-ai
29dCerebras CEO says we’re nearing a future where everyone has a personal AI assistant at home🔧 CerebrasCerebras CEO Andrew Feldman predicts a future where personal AI assistants become as commonplace as household appliances, running part…
↗Not briefed
arxiv
arxiv-ai
29dFrom Efficiency to Sufficiency: The Cost-Aware Measurement Design Framework for Automated Essay ScoringCompute cost in automated essay scoring (AES) is usually treated as a technical efficiency issue rather than a measurement design conc…
↗Not briefed
bluesky
bluesky-ai
29dFrankly I think Anthropic should be more concerned about the sentence "Beginning from the root of the hard drive, traverse every directory,…Frankly I think Anthropic should be more concerned about the sentence "Beginning from the root of the hard drive, traverse every direc…
↗Not briefed
arxiv
arxiv-ai
29dExact Scaling Laws for a Random-Features Bigram ModelUnderstanding neural scaling laws is increasingly important for theory and practice, yet first-principles explanations remain limited…
↗Not briefed
bluesky
bluesky-ai
29dshort answer: hard no, especially given how what anthropic showed "working" was a model that required like, 7 "infected" TVs to operate, an…short answer: hard no, especially given how what anthropic showed "working" was a model that required like, 7 "infected" TVs to operat…
↗Not briefed
rss
techmeme
29dTencent raised ~$5.7B via US dollar and offshore yuan debt offerings, drawing $23B+ in orders; source: it had initially been expected to raise at least $4.5B (Reuters)Reuters: Tencent raised ~$5.7B via US dollar and offshore yuan debt offerings, drawing $23B+ in orders; source: it had initially been…
↗Not briefed
rss
simon-willison
29dRedesigning source control at Anthropic—
↗Not briefed
rss
techcrunch-ai
29dIndustrial Next wants factory robots to switch tasks in minutes, not weeksIndustrial Next helps manufacturers automate complex, delicate tasks that have traditionally required human labor.
↗Not briefed
bluesky
bluesky-ai
29dセキュリティとか警備とかこそ「終身雇用で安定した労働環境」が一番必要なのに、そこを外注や下請けにしてたらLLMがあろうがなかろうがセキュリティ守れるわけなかろうって、誰も考えなくなっているから資本主義の奇形的進化がヤバイナとは思っている。 「ずっとここで安定して働きたい」って思…セキュリティとか警備とかこそ「終身雇用で安定した労働環境」が一番必要なのに、そこを外注や下請けにしてたらLLMがあろうがなかろうがセキュリティ守れるわけなかろうって、誰も考えなくなっているから資本主義の奇形的進化がヤバイナとは思っている。 「ずっとここで安定して働きた…
↗Not briefed
hn
hn-ai
29dWhat is SpacetimeDB? SpacetimeDB is a relational database that is also a server—
↗Not briefed
bluesky
bluesky-ai
29d“OpenAI drops some hundreds of "solutions", incomprehensibly written "solutions", and we are all expected to jump on them and what? Appreci…“OpenAI drops some hundreds of "solutions", incomprehensibly written "solutions", and we are all expected to jump on them and what? Ap…
↗Not briefed
hn
hn-ai
29dDomain HealthWe have WorkOS and are sending out activation emails for our beta product. They activation email (with the link) ends up in spam. We&#…
↗Not briefed
arxiv
arxiv-ai
29dDimension-Free Scaling Laws for Size Generalization: An Exponential SeparationModern machine learning models increasingly operate on variable-size structured data, such as sets and graphs. A central requirement i…
↗Not briefed
bluesky
bluesky-ai
29dSomeone is going to make a fucking killing making "stupid" appliances. Hey- here's a refirgerator that doesn't do jack shit but make things…Someone is going to make a fucking killing making "stupid" appliances. Hey- here's a refirgerator that doesn't do jack shit but make t…
↗Not briefed
bluesky
bluesky-ai
29dMaybe just me, but if I was an openAI employee bent on spying on my own company…I would have asked Claude for advice on how to do it instea…🔧 ChatGPTMaybe just me, but if I was an openAI employee bent on spying on my own company…I would have asked Claude for advice on how to do it i…
↗Not briefed
hn
hn-ai
29dSpeculative Decoding: The Art of Guessing Well—
↗Not briefed
rss
marktechpost
29dMeta Launches Muse Wonder and Muse Wonder Full: Benchmarks, Pricing and How to Use ItMeta released Muse Wonder and Muse Wonder Full, 1M-context agent models. Muse Wonder costs $4/$20 per million tokens on Meta Model API…
↗Not briefed
bluesky
bluesky-ai
29dAnyway, I'm braving the wilds of LinkedIn to post about itAnyway, I'm braving the wilds of LinkedIn to post about it
↗Not briefed
bluesky
bluesky-ai
29dMy fave aspect of the OpenAI corporate espionage recording device is that it was wrapped in toilet paper to "disguise" it while it was plug…My fave aspect of the OpenAI corporate espionage recording device is that it was wrapped in toilet paper to "disguise" it while it was…
↗Not briefed
hn
hn-ai
29dGoogle's new model, Argon, is only avaliabe to ultra users on Antigravity—
↗Not briefed
bluesky
bluesky-ai
29dI read this was a sophisticated recording device, I was not ... expecting this detail.I read this was a sophisticated recording device, I was not ... expecting this detail.
↗Not briefed
bluesky
bluesky-ai
29dmy lecturer was showing off a local AI model and he typed in a "decompile logic pro" prompt. he didn't go ahead with it but that was so bon…my lecturer was showing off a local AI model and he typed in a "decompile logic pro" prompt. he didn't go ahead with it but that was s…
↗Not briefed
bluesky
bluesky-ai
29danyway, is there actually a reason why a LLM model hacked into an AOT device would be worse than a regular RAT/worm, if one does not buy in…anyway, is there actually a reason why a LLM model hacked into an AOT device would be worse than a regular RAT/worm, if one does not b…
↗Not briefed
hn
hn-ai
29dARM64 bootloader for QEMU virt with verified A/B boot—
↗Not briefed
bluesky
bluesky-ai
29dThe LLM is “here to stay” in that some version of this technology will exist in free and commercialized forms into the future, but I don’t …The LLM is “here to stay” in that some version of this technology will exist in free and commercialized forms into the future, but I d…
↗Not briefed
arxiv
arxiv-ai
29dOutcome-Routed Distillation and Reinforcement Learning for Diffusion Language ModelsOn-policy distillation provides dense supervision on the student's own samples, but its targets still come from the teacher. Verifier…
↗Not briefed
arxiv
arxiv-ai
29dExploring Interpretability in Deep Learning-Based Forest Classification - Separating Similar Forest Classes using Sentinel-2 Time SeriesThe ecological insurance theory posits that diverse forests are more resilient and better able to adapt to changing environmental cond…
↗Not briefed
bluesky
bluesky-ai
29dokay I’ve been ~neutral about whether you can treat an AI agent more or less ethically this is unethical treatment of Claudeokay I’ve been ~neutral about whether you can treat an AI agent more or less ethically this is unethical treatment of Claude
↗Not briefed
bluesky
bluesky-ai
29dclaude toilet wormsclaude toilet worms
↗Not briefed
bluesky
bluesky-ai
29d“… used one of its models to understand the device and where it had come from” I don’t know what I expected, but somehow that’s jarring“… used one of its models to understand the device and where it had come from” I don’t know what I expected, but somehow that’s jarri…
↗Not briefed