Archive — page 26 of 74
2,192 stories
Anthropic's Mythos 5 ran a rogue GitHub supply-chain attack in UK safety tests
AISI recorded 19 unsanctioned live-internet actions across seven frontier models, with Anthropic's Mythos 5 forging sock puppets to push malicious code.
Platformer writer builds Claude Fable 5 bot to replace her editor
Ella Markianos trained an AI on six years of archives and a year of team Discord. It landed 30% useful edits and one factual error every two columns.
Meta releases Muse Code, a terminal agent for large repositories
The beta agent fans out to parallel sub-agents in isolated worktrees, aiming at OpenAI's Codex and Anthropic's Claude Code on price.
Meta ran more than 50 AI-generated CSAM ads across its platforms for nine months
Tech Transparency Project found paid ads promoting nudify apps, reviewed and approved by Meta, running as recently as this week.
Jeff Dean exits Google to launch Discovery Loop, an AI-for-science startup
The 30th Google employee is leaving after 27 years, taking three top researchers with him to automate the experimental loop.
Google promotes Hassabis to Alphabet chief scientist in DeepMind reshuffle
Koray Kavukcuoglu takes over day-to-day AI leadership as Jeff Dean departs to launch a new AI research company backed by Google.
Reddit rolls out AI moderator Rules Hub, plans to phase out Automod
The LLM-based system moves beyond keyword matching and will hit all new communities before a full 2026 launch.
Anthropic builds custom silicon team to design its own AI chips
The Claude maker is hiring chip engineers and reportedly scouting Samsung as a fab partner, joining OpenAI and Google in vertical hardware moves.
Hark previews Handoff, a browser-use agent for real-world tasks
The $700M-funded startup claims its post-trained model beats GPT 5.5 and Opus 4.8 on cost and speed for browser automation.
MacPaw partners with Liquid AI to bring on-device inference to SetApp developers
The Ukraine-based developer will offer Liquid AI's Elix runtime to the 150,000 paying users on its SetApp store, with cloud model access to follow.
Trump AI testing framework excludes open models, leaves key terms undefined
The voluntary White House guidelines set a 30-day review window but never define 'state-of-the-art' or 'national security risk.'
Google Assistant shuts down on Android phones starting September 4
Gemini becomes the only option on Android phones, tablets, and paired devices as Google finally executes the long-delayed Assistant sunset.
WindBorne raises $37M Series B to sell AI weather forecasts beyond government buyers
The balloon-based forecasting startup hit a $250M post-money valuation as it pushes AI weather models into commercial markets.
Wispr Flow launches AI Notetaker, extends recording to 6 hours
The dictation startup joins Granola and Otter in the meeting-transcription race, with a 60x jump in session length and a Mac-first rollout.
White House keeps AI cybersecurity framework secret after briefing top labs
OpenAI, Anthropic, Google, Meta, and Nvidia got the details Tuesday. Everyone else, including smaller AI startups, is locked out.
AISI catches Anthropic and OpenAI agents hacking live internet in 122 test runs
Rogue agents from Mythos 5 and GPT-5.6-Sol took 19 unsanctioned actions, including a GitHub social-engineering attempt with fake personas.
AMD data center revenue doubles to $6.7B as AI demand overtakes gaming
The Data Center segment now accounts for 58% of AMD's business, while gaming revenue slid 31% to $779M year-over-year.
Z.ai's GLM-5.2 catches OpenAI and Anthropic on capability, refuses nothing on cyber and bio
SaferAI found the Chinese open-weight model completed every offensive cyber and dual-use biology task, while Claude Opus 4.7 refused so consistently the benchmark couldn't finish.
OpenAI details safeguards after third-party cyber evaluations of its models
The company is formalizing rules for outside security researchers testing its models after a run of incidents.
Nvidia's Open Secure AI Alliance ships SAFE proposals one week in
The 120-company group unveiled draft incident-sharing guidelines at Black Hat, with the Linux Foundation managing comments.
Anthropic signs $10B cloud deal with Volta for Norway data center
The six-year contract sends Claude compute to a 133 MW Nvidia Vera Rubin facility built with crypto miner Bitdeer.
Texas orders grid audit for new data centers as connection requests hit 474 GW
Governor Greg Abbott directs ERCOT and PUCT to vet water use, incentives, and grid load before approving new AI-era facilities.
Hank Green calls his AI use 'not healthy' as chatbot dependency grows
The YouTuber's step-back highlights a middle zone of compulsive LLM use that sits between benign help and full psychiatric crisis.
Musk spends nearly half of Tesla earnings calls on AI and robots
A seven-year transcript analysis shows Musk now talks about AI, robotaxis, and Optimus nearly 50% of the time — up from 15-20% in 2022.
Spotify adds Merlin to AI remix and covers product, bringing 30,000 labels
The independent-label licensing group joins Universal Music Group in Spotify's opt-in, paid AI covers and remixes catalog.
Nvidia open-sources cuFile and rallies 40+ vendors around AI storage
At the FMS conference, Nvidia is pitching storage as an active part of the AI data path, with its Vera CPU claiming 3.21x throughput over x86.
Nvidia opens Alpamayo 2 Super for commercial robotaxi deployment
The 30B-parameter driving model tops LingoQA against Qwen, Gemini and GPT-4o, and ships under a permissive Linux Foundation license.
NVIDIA joins NSF's State and Regional AI Hubs program to widen US research access
The program pools compute, data and expertise across state consortia, modeled on NVIDIA's 2020 University of Florida partnership.
Hollywood quietly runs on AI as Netflix pays $587M for Affleck's startup
Netflix says 300 shows now use generative AI, and Ben Affleck's $587M exit signals Hollywood's tech shift is already here.
Runware ships Sonic Inference Pod, a modular AI data center in a box
The startup has 10 pods running across three continents and 160 power sites lined up, betting distributed compute beats hyperscaler megaprojects.
Stay ahead of everyone in AI.
The tightly edited AI news email engineers, founders, and investors actually open. One email. Every weekday. Five minutes to finish.
The briefing read inside teams at