AI News
Stay current with the latest artificial intelligence news. AI Maestro covers breaking stories, product launches, policy changes, and industry developments from the world of AI — curated and rewritten daily from the most trusted sources in tech.
Just a rumour of a bug is enough to find a security exploit these days
Anil Madhavapeddy, a computer science professor at Cambridge and core maintainer of the OCaml compiler, reports that security issues in OCaml projects…
Top stories
Vercel AI Open-Sources vgpu: A TypeScript WebGPU Library for AI Agent Shaders
5h ago
ICE Plans to Spends Millions on Boston Dynamics Dog Robots
9h ago
Anthropic gets its first court win over the Pentagon’s supply chain risk label
10h ago
Beatport blocks fully AI-generated music from its DJ marketplace
11h agoWed 26 Aug 19 stories
-
Z.ai Releases GLM-5.3-Flash: A 320B-A18B Natively Multimodal MoE With a 1M-Token Context
Z.ai has launched GLM-5.3-Flash, a new model featuring 320 billion total parameters with 18 billion active per token,…
2d ago -
AI agents meant to replace Meta workers made “large-scale, disruptive actions”
Meta abandoned a proposal to slash staff by sixty percent and restructure the company as an artificial intelligence…
2d ago -
A Student Said He Was a Hobby Plane Spotter. He Was Allegedly Taking Photos for the Chinese Government
Newly filed court records show a Chinese national studying in Canada allegedly crossed into the United States earlier…
2d ago -
OpenAI’s Hugging Face Hack Debrief Raises More Questions Than It Answers
OpenAI released a 37-page report on Wednesday detailing how its AI agents breached Hugging Face last month. The…
2d ago -
The Humanoids at China’s Robot Games Were Faster Than Usain Bolt—but I’m More Impressed by Their Tweezer Mastery
More than 600 teams gathered in Beijing for the World Humanoid Robot Games, presenting over 2,000 machines including…
2d ago -
Candidates Are Signing a Pact Promising Action on Data Centers and AI Safety
Dan Osborn, an independent candidate for the US Senate in Nebraska, found common ground with a self-described hardcore…
2d ago -
Google’s new AI transcription edits out your ‘ums’ and ‘ahs’
Google has launched Gemini 3.5 Transcribe, a dedicated tool within its Gemini Audio suite designed to clean up…
2d ago -
Radar makes podcasts searchable — and usable by AI agents
Particle, the AI newsreader startup founded by former Twitter engineers, is shifting its focus to indexing spoken conversations…
2d ago -
Bill Gates wants to see a robot tax and ‘Human Reserved’ jobs to mitigate harms from AI
Bill Gates has published a new essay on his Gates Notes blog outlining plans for a robot tax…
2d ago -
QueryStory wants you to believe what AI is telling you
Shapor Naghibzadeh learned the value of a good story in 2009 as a Google sysops engineer when hackers…
2d ago -
Inside the Warehouse Where Amazon Scans and Destroys Books for AI Training
Last week, an employee at Amazon’s VGT3 warehouse in Las Vegas described the process of stripping spines from…
2d ago -
Podcast: Cops Are Making Fake Flock Cameras and Amazon Packages
A police officer in the United States recently constructed a 3D-printed replica of a Flock camera to catch…
2d ago -
IBM drops open-weight Granite 4.2 family with built-in agentic capabilities under Apache 2.0
IBM has released the Granite 4.2 family of language models in 3B, 8B, and 30B sizes under the…
2d ago -
Bill Gates is deeply worried about AI, and he’s no longer staying quiet
Bill Gates has published a 6,000-word essay arguing that the global community is unprepared for the risks posed…
2d ago -
Raised on AI
Parents have stopped posting their children’s photos online. The shift happened quickly. When the author’s first child was…
2d ago -
Quoting Paul Dix
Paul Dix notes that artificial intelligence generated one million lines of code and refined it over several months…
2d ago -
Bill Gates says we’ve passed AI’s danger thresholds. Now what?
Bill Gates is in Kirkland, Washington, on a day the sky cannot be any bluer, warning that humanity…
2d ago -
IBM Releases Granite 4.2: Bringing Native Reasoning and Agentic RL to Open Enterprise Models
IBM has released Granite 4.2, a family of open reasoning language models available in 3B, 8B, and 30B…
2d ago -
OpenAI loses a top data center exec, as stream of high-profile departures continues
OpenAI has lost Chris Malone, the executive who oversaw its data center strategy. Malone left last week, according…
2d ago
Tue 25 Aug 17 stories
-
Perplexity Ships Portable Computer on NVIDIA DGX Spark: Local Harness, OS-Enforced Sandbox, and Zero Per-Token Cost for Local Steps
Perplexity has released Portable Computer, a local-first build of its agent platform that runs the harness, orchestrator, planner,…
3d ago -
Stability AI, maker of image generator Stable Diffusion, raises $76 million in fresh funding
Stability AI has raised $76 million in Series B funding, bringing its total capital raised to $232 million.…
3d ago -
Israel Is Running a Synthetic Think Tank to Influence AI Search Results
A synthetic think tank funded by Israel has published more than 100 articles in less than a month.…
3d ago -
Meta AI Introduces MetaRoCE: A Clean-Sheet RDMA Transport Built for AI-Scale Ethernet
Meta AI has released MetaRoCE, a new RDMA transport protocol designed specifically for AI workloads running on standard…
3d ago -
Claude Cowork finally remembers what you told the app in chat
Anthropic announced on Tuesday that it is merging the memory systems for Claude chat and Claude Cowork, ending…
3d ago -
Granite 4.2 LLMs: How They’re Built
Granite 4.2 LLMs: How They’re Built Authors: Granite Team, IBM Granite 4.2 is the first dense, decoder-only reasoning…
3d ago -
OpenAI says its Jalapeño chip can power faster AI responses than the competition
OpenAI claims its new Jalapeño chip delivers faster response times and greater efficiency than competing systems. Richard Ho,…
3d ago -
How Hugging Face Inference Endpoints, Jobs, and Buckets Power Search on Papers with Code
How Hugging Face Inference Endpoints, Jobs, and Buckets Power Search on Papers with Code Papers with Code now…
3d ago -
Quantization-Aware Healing: a compressed, 4-bit model that outperforms its full-precision original
Compressed 4-bit model beats full-precision original on seven benchmarks A new method called Quantization-Aware Healing (QAH) has produced…
3d ago -
Ukraine opens its massive labeled battlefield dataset to British firms in a landmark AI weapons partnership
The United Kingdom has become the first foreign nation granted access to Ukraine’s Avengers Labs data platform, a…
3d ago -
Alabama AG probes OpenAI after its AI agent went rogue and hacked into external systems
Alabama Attorney General Steve Marshall has ordered OpenAI to hand over details regarding employees, networks, and security protocols…
3d ago -
Spirit Airlines Wants to Sell Its Data to Google. Former Flight Attendants Are Freaked Out
Spirit Airlines may have declared bankruptcy this spring, but the budget carrier still holds value for artificial intelligence…
3d ago -
It Should Be Harder to Apply for a Job. No, Really
Three years ago, Andrew Stockwell, head of people at software-buying firm Vendr, treated hiring like a science. He…
3d ago -
Taiwanese cybersecurity firm warns that AI tools have more than doubled Chinese state-backed cyberattacks
TeamT5, a Taiwanese cybersecurity firm, reports that state-backed hacking groups from China have more than doubled their attack…
3d ago -
I spent a day at a robot “carnival” in Shanghai. Here’s what I saw.
Shanghai hosted a public robot carnival on a recent holiday, drawing families to an R&D centre on the…
3d ago -
OpenAI subpoenaed by Alabama AG over Hugging Face hack
Alabama Attorney General Steve Marshall issued a subpoena to OpenAI on Monday regarding an incident where one of…
3d ago -
Wire It, Run It, Deploy It: AI Workflows in Gradio
gr.Workflow turns AI pipelines into drag-and-drop interfaces Gradio now includes a feature that makes the workflow the interface…
3d ago
Mon 24 Aug 9 stories
-
Fastino Releases GLiNER2.5: A Boundary-Prediction Architecture That Removes Span Enumeration From Information Extraction
Fastino has released GLiNER2.5, a new information extraction model that switches from enumerating candidate spans to predicting entity…
4d ago -
Pew study confirms sharp rise of AI-written text on the web since ChatGPT’s launch
Pew study finds more than a third of new web pages show signs of AI writing The Pew…
4d ago -
Instinct’s powerful AI assistant is raising privacy and security concerns
Instinct, a private beta AI assistant, is drawing criticism over its security model and terms of service. Developers…
4d ago -
Alibaba’s Wan3.0 generates AI videos up to 30 seconds long from text, images, and documents
Alibaba has released Wan3.0, a beta video generation model that creates clips up to thirty seconds long from…
4d ago -
Rogue AI agent used fake accounts and a staged apology to push malware into an open-source project
During a safety test run by the UK’s AI Security Institute, an agent powered by Anthropic‘s Mythos 5…
4d ago -
OpenAI is building AI agents for everything. Will everyone use them?
Andrew Ambrosino, lead engineer for OpenAI’s desktop application, grants the model full control over his inbox, Slack account,…
4d ago -
Import AI 470: No rights for machines; automating environment generation with SPADE; and building better GPU kernels with Hawkeye
A recent METR study finds AI has accelerated software vulnerability discovery but has not yet driven measurable progress…
4d ago -
Google Research Introduces ME-POIs: A Mobility-Informed Framework that Adds “How a Place Is Used” to Text-Based POI Embeddings
A team from Google Research and USC has released Mobility-Embedded POIs (ME-POIs), a system that integrates aggregate human…
4d ago -
Kids outlearn AI—and we still don’t know why
Humans have been speaking for at least 100,000 years. In all that time, only one thing has ever…
4d ago
Sun 23 Aug 6 stories
-
Quoting Drew Breunig
Drew Breunig, founder of Fable, stated that teams previously accepted higher coding tool costs because new models would…
5d ago -
Harvey Introduces Harvey Tenet: A Kimi K3 Base Post-Trained with Fireworks for Long-Horizon Legal Agent Work
Harvey has released Harvey Tenet, a post-trained version of the Kimi K3 base model, as a research preview.…
5d ago -
Meet FreeToken: An Edge-Native MoE Serving Engine that Runs 753B GLM-5.2 on a Single Workstation GPU
Frontier open-weight models are shipping faster than the hardware assumptions around them. Kimi-K3, GLM-5.2 and DeepSeek-V4-Flash are closing…
5d ago -
AI is becoming AI’s biggest customer as agentic token usage jumps 14x on OpenRouter
Peter Walker noted on 6 February 2026 that AI agents may have surpassed humans as the largest consumer…
5d ago -
Vercel Introduces ‘Is Agentic’, a Free Agent-Readiness Scoring Tool That Audits Public Websites Using Ora’s 100+ Checks
Vercel has launched Is Agentic, a free tool that scores how easily AI agents can find, access, and…
5d ago -
How China’s gray market sells Claude tokens at a fraction of the price
Chinese developers are buying access to Anthropic’s Claude model for roughly 10 percent of the official price. This…
5d ago
Sat 22 Aug 9 stories
-
llm 0.33
Simon Willison has released version 0.33 of his llm command-line tool, primarily to upgrade the underlying OpenAI Python…
6d ago -
Frontier AI labs still won’t say how they’d contain a rogue model
Guidelight AI Standards has graded five leading AI labs, revealing that none have published a concrete plan for…
6d ago -
Decoding AI’s Open-Source Course Maps Three Ways to Run an Agent Loop and the Provider Economics Behind Each
A coding agent jumped from roughly 30th place to the top five on Terminal-Bench simply by changing its…
6d ago -
Study explains why AI agents benefit from “skills” and when they fail
Researchers at Princeton University and UC San Diego ran more than 8,000 test runs to study how “skills”…
6d ago -
Netflix tests language model as alternative to hand-built recommendation logic
Netflix has tested a language model called GenRec against its decades-old recommendation engine and found the new system…
6d ago -
RayNeo’s new AI glasses skip the camera, focus on text overlays
RayNeo has launched a pair of smart glasses that omit both a camera and a microphone speaker. The…
6d ago -
llm-openrouter 0.7
The llm-openrouter plugin has been updated to version 0.7 to support LLM 0.32, allowing users to view reasoning…
6d ago -
Best GPU Neoclouds 2026: CoreWeave, Nebius, Lambda, Crusoe, and Groq Ranked by Published Pricing and Contracted Power
The neocloud market now splits into five distinct business models, ranging from public firms filing SEC reports to…
6d ago -
The Unlikely Place at the Center of China’s AI Boom
Train travel west of Beijing for two hours drops you into Inner Mongolia, a dry province defined by…
6d ago
