Archive — page 22 of 62
1,853 stories
Google launches Gemini 3.5 Flash Cyber to undercut Anthropic's Mythos
The new security model runs at a fraction of Mythos 5's cost and found 55 confirmed bugs in V8, beating Claude Opus 4.6's 36.
Google ships three new Gemini models but delays Pro update
Gemini 3.6 Flash cuts token usage by up to 17%, but the flagship Pro refresh is still stuck as OpenAI and Anthropic pull ahead.
Nvidia's Vera Rubin NVL72 hits production with 10x tokens per megawatt
CoreWeave's DeepSeek-R1 benchmark shows a 10x throughput-per-watt gain over Grace Blackwell, with racks now live at four major clouds.
Treasury's Bessent threatens sanctions on Chinese AI models over IP theft
Bessent says Washington will examine open-source models from China for stolen IP, days after reports of a possible wholesale ban.
Nvidia ships Spectrum-6 Ethernet at 102.4 Tb/s for Vera Rubin AI factories
The next-gen switch doubles capacity and lands first at CoreWeave, Microsoft, Nebius, SpaceXAI and Tesla to wire hundreds of thousands of GPUs.
Google ships Gemini 3.6 Flash and 3.5 Flash-Lite to cut agent costs
The Flash refresh trims output tokens 17%, hits 350 tokens/second on Flash-Lite, and adds a cyber-tuned variant for CodeMender.
Deezer says AI-generated tracks now exceed 50% of daily uploads
The French streamer will delist AI tracks unstreamed for six months as uploads hit 90,000 per day in June 2026.
Chinese AI models Kimi K3 and Qwen3.8 close the gap on OpenAI and Anthropic
Six of OpenRouter's top 10 tools are already Chinese. The latest releases from Moonshot and Alibaba shouldn't be shocking anyone.
Gritt exits stealth with $34M to put AI-driven robots on solar construction sites
The Carnegie Mellon-founded startup says its systems help crews install 3,000–4,000 panels a day, up from 800 without them.
US Army burns through a year of AI tokens in weeks on Ask Sage
DEVCOM staff told to throttle usage after the Army CIO's 100M-token annual pool ran dry weeks after unlimited access was announced.
Judge approves Anthropic's $1.5B copyright settlement with authors
The largest copyright payout in U.S. history closes one case but leaves the fair-use question open for OpenAI, Google, Meta, and Midjourney.
Sony sues Udio over 30,000 songs in expanded AI training copyright case
Sony seeks up to $150,000 per work after a judge blocked it from adding the tracks to the original 2024 case against the AI music generator.
Firefighting drones tested in California and Alaska as fire season stretches year-round
CAL FIRE, Seneca, and Dryad Networks put autonomous suppression drones through live trials as the $11M XPRIZE Wildfire competition wraps up.
Model Context Protocol drops stateful sessions in next week's update
The plumbing behind AI agent integrations shifts to a stateless design, addressing a scaling headache that has slowed first-party MCP rollouts.
Google's Frozen v2 chip targets 6-10x efficiency gain for Gemini
Alphabet is designing a 2028 server chip to run Gemini cheaper as it spends $190B on AI infrastructure this year.
Kimi's release splits Trump's AI advisors into open-source war
Moonshot's free Kimi model rivals OpenAI and Anthropic, and Trump's current and former AI advisors are trading insults over what to do about it.
Chris Fall resigns as head of US AI standards agency after three months
The director of Commerce's Center for AI Standards and Innovation is out just three months after taking the role in April.
Adobe wires Google's Nano Banana into Project Indigo camera app
The version 1.1 update adds generative filters, object removal, and AI photo critique — running on Google's model, not Adobe's Firefly.
YouTube tightens monetization rules on AI slop and distressing videos
Three categories of 'inauthentic content' now bar creators from Partner Program payouts, with AI personas on health and finance singled out.
Nvidia opens creative apps to AI agents at SIGGRAPH with MCP push
Adobe, Blender, Unreal and Affinity connect AI agents into production pipelines as Nvidia ships a Synthetic Video Detector hitting 92% accuracy.
CuspAI raises $450M from UK government and Jeff Bezos for materials discovery
The Cambridge-based startup wants to use generative AI to design new materials for carbon capture, batteries, and semiconductors.
Bristol Myers buys Nvidia's newest AI computing system for drug research
The pharma giant becomes an early customer for Nvidia's latest AI hardware, deploying it to accelerate small-molecule and biologic discovery.
UK AISI: open-weight models now trail closed AI on cyber by 4-7 months
GLM-5.2 and DeepSeek V4-Pro closed the cyber capability gap from 6-10 months to 4-7 months. Kimi K3 is next in line.
Augment Code bets on semantic retrieval to beat Claude Code on efficiency
Augment says its context engine runs Terminal-Bench 33% more efficiently than Claude Code using the same underlying model.
Moonshot and Alibaba unveil open-weight rivals to GPT-5.6 and Claude Fable 5
Kimi K3 lands at 2.8 trillion parameters, Qwen3.8 at 2.4 trillion, and both will release weights that US labs keep locked down.
LLMs stereotype job candidates 65% more than humans in Princeton hiring study
OpenAI's o3 scored 1.83 on a segregation scale where 2 is total sorting — and reasoning models were the worst offenders.
Nvidia anchors Japan's $6.2B sovereign AI push with Vera Rubin factory
Jensen Huang's two-day Tokyo swing yielded a national AI factory, a robotics coalition of 44 firms, and deeper Toyota integration.
1010Benja's Suno-made track 'Semiramis' Dream' rewrites the AI music debate
The rapper ran vocals and beats through Suno a few hundred times to build the track — and refuses to hide the AI credit.
Christopher Nolan calls AI a Trojan horse the public already sees through
The Odyssey director says young viewers' 'AI slop' backlash is healthy skepticism, not technophobia.
Current AI raises $400M to build a public AI alternative to OpenAI and Google
The French-seeded nonprofit has deployed $3.2M in grants and shipped an open chatbot built in seven weeks by ten organizations.
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