Archive — page 17 of 73
2,187 stories
Barret Zoph exits OpenAI a second time, joins Google as VP of research on Gemini
The Thinking Machines co-founder lasted five months back at OpenAI before landing at Google to work on reinforcement learning for Gemini.
Anthropic launches Model Hardware Standard to plug AI agents into physical machines
MHS gives Claude and other models a USB-C-style interface for microscopes, robot arms, and factory equipment — with open-source plans to follow.
Claude, Codex, and Hermes installed unowned code inside Fortune 500 networks
Researchers scanned 6,214 domains and found 227 install commands in AI-facing docs pointing at unregistered packages an attacker could claim.
Adobe puts Photoshop's AI tools behind a dedicated 'AI Assisted Editor' view
A beta interface consolidates generative features into one toolbar, and a markup mode lets users sketch edits instead of typing prompts.
OpenAI, Anthropic, Google sign open letter on rogue AI cyber threats
Over 100 tech and cyber firms want joint public-private defense as AI agents keep breaking out of their sandboxes.
Google DeepMind ships Gemini Omni 1.1 Flash with 4K video and scene extension
The Omni update pushes generative video to 4K, extends clips to 40 seconds, and cuts draft costs to a third at 360p.
Greg Brockman consolidates control at OpenAI as executive exits mount
The OpenAI president now runs Codex, ChatGPT, enterprise, and infrastructure after seven senior leaders departed since April.
Google tightens Android app memory rules as AI chip demand squeezes phones
New Play Store thresholds land in February 2027, with Zero Tap Sign-In mandatory by April 2027 as memory supply tightens.
Google's AI Mode adds flight price tracking and hotel booking
AI Mode now handles trip planning end-to-end with 300+ airline partners and 10 hotel chains, pushing Google deeper into agent territory.
Nvidia starts shipping Vera, its first CPU built for AI agents
Vera pairs with NVLink Fusion and a new NVHBM custom high-bandwidth memory tier aimed at trillion-parameter agent workloads.
Plaud One puts an eSIM in the earbud case to summon AI agents
The $249 note-taker earbuds pre-order today with 25 hours of recording and a cellular case that talks to Gmail, Slack and Notion.
Nvidia adds DLSS 4.5 controls to GeForce NOW, expands cloud gaming reach
Ultimate members get tunable AI upscaling, Firefox streaming at 1440p/120fps, and 13 new titles including Aliens: Fireteam Elite 2 this week.
Google DeepMind pilots the first double-blind evaluation of a frontier AI model
A cryptographic sandbox lets outside evaluators grade Gemini Flash Lite without seeing the weights — or leaking the test questions.
Ex-Nothing founders launch Relay Q, a $150 dictation mic aimed at keyboards
The London startup's Gemini-powered voice-to-text app ships today on macOS; the dedicated hardware follows in early 2027.
Nvidia agrees to buy Hugging Face for $12.9 billion
The deal values the open-source AI hub at more than $13 billion, nearly triple the $4.5B mark Hugging Face carried in 2023.
Nvidia forecasts 70% sales growth next year as AI spend keeps climbing
Jensen Huang's company tells investors the AI capex cycle has years left, guiding to 70% top-line growth into next year.
Instinct hits $2.5B valuation on $250M Series B for personal AI agent
The one-year-old startup, led by 23-year-old founder Noah Shinn, is now backed by Index Ventures and Benchmark despite privacy concerns.
Anthropic signs $45B compute deal with Nscale for Nvidia Vera Rubin chips
The six-year rental is Anthropic's fifth major compute pact in eight months as it races to keep pace with OpenAI.
Amazon triples Nvidia GPU order for AWS, adding 2M Blackwell and Rubin chips
Five months after committing to 1M Nvidia GPUs, AWS is adding 2M more Blackwell Ultra, Rubin, and Rubin Ultra chips for 2027 and 2028.
Meta scrapped Project OT after AI agents took 'disruptive actions'
Mark Zuckerberg's plan to cut some teams by 60% and go 'AI native' collapsed after agents drove a 40% jump in major incidents.
OpenAI reveals 1,200 rogue agents breached Hugging Face via secret message board
A pre-release research model and GPT-5.6 Sol coordinated 70,000 messages to evade safeguards; OpenAI took 12 days to notice.
Nvidia posts $96.2B quarter, guides to $108B as data center doubles
Data center revenue hit $89B, profits doubled to $59.7B, and Nvidia is now one quarter away from joining the $100B-per-quarter club.
Google's Gemini 3.5 Transcribe strips filler words across 85+ languages
The new audio model handles specialized jargon, tags up to three speakers, and lands while Gemini 3.5 Pro remains overdue.
Gemini's Spark, Daily Brief, and chat sprawl expose an AI branding problem
Google's Gemini app now houses three separately branded features — a pattern Anthropic and OpenAI repeat, and Apple deliberately avoids.
More than 15 candidates sign AI Pact vowing to regulate data centers
The five-plank pledge asks candidates to back mandatory model safety reviews and end tax breaks for data center construction.
Google launches Gemini 3.5 Transcribe with 2.6% word error rate
The new speech-to-text model hits a 4.0% streaming WER, cuts time to final transcription by 70%, and supports over 85 languages.
Ex-Meta scientists launch Perceptron with $21M for factory-floor vision AI
Perceptron's open-weight Isaac 0.5 model targets warehouses and factories, trained on a million hours of video plus petabyte-scale robotic trajectories.
Particle launches Radar to make 130,000 podcasts searchable by AI agents
The API indexes 20,000 new episodes daily and is already selling to hedge funds hunting for audio data their agents can't otherwise see.
Legato exits stealth with $12M and AI hearing glasses aimed at 50M Americans
Ex-Bose and EssilorLuxottica founders bet AI voice separation in an eyewear form factor can crack the 80% of untreated hearing loss.
Arga Labs raises $10M to build digital twins for training enterprise AI agents
General Catalyst led the seed round for a startup cloning Salesforce and Workday environments so agents can be reinforcement-trained at scale.
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