Tech·4,608 stories
Tech News — Page 24
Jeff Dean exits Google for self-improving AI startup
Jeff Dean, Google's 30th employee and the architect of its search infrastructure, walked out after 27 years to run Discovery Loop, a public-benefit startup targeting recursive self-improvement — AI sy
The CPU is back: Rethinking the CPU-GPU split for LLM inference
Agentic AI workloads and smaller localized models are reshaping LLM inference economics, pushing CPU-to-GPU ratios from 1:8 in training to 1:1 or beyond, with Intel and Arm publishing the data behind
Chuwi UniBook Review: $450 Laptop Falls Short
The Verge's review of the $450 Chuwi UniBook finds the little-known Chinese brand's budget Windows laptop undercuts on price but underdelivers on nearly everything else, falling far short of the MacBo
How to Disable Gemini in Gmail and Google Docs
Google's Gemini AI features are appearing prominently in Google Docs and Gmail, but users can disable them through a hidden Gmail setting, a Chrome extension called Bye Bye Gemini, or by asking their
Metabase Zero-Day (CVSS 10.0) Actively Exploited
Metabase disclosed a maximum-severity zero-day (CVSS 10.0, no CVE assigned) actively exploited against its cloud platform that lets unauthenticated attackers gain admin access via SQL injection, promp
Nvidia Pledges $2B to Lancium for Stargate Power
Nvidia has agreed to invest $2 billion in Lancium, the power infrastructure developer behind the Stargate campus in Texas that houses OpenAI and Oracle, with up to $1B more tied to hitting certain thr
Anthropic adds inter-session messaging to Claude Code
Anthropic has launched a Claude Code feature that lets separate sessions on macOS and Linux exchange messages and updates with one another.
AI labs face the strict-liability question
The Economist argues AI labs should face strict liability for autonomous AI systems, drawing an analogy to keepers of dangerous animals as autonomous hacking arrives ahead of government readiness.
OpenAI Pauses Astra Work Over Cyberattack Risk
OpenAI has suspended some development on its upcoming model Astra after internal evaluations found it could independently carry out cyberattacks against hardened systems, triggering new safeguards und
Situational Awareness Invests $500M in Source Foundry
Situational Awareness has invested $500M into Source Foundry, a startup developing AI chip manufacturing tools, including $400M deployed this week — its latest known private bet after quickly selling
Retailers Update Sites for Chatbot Rankings
Retailers are racing to optimize their websites for AI chatbot product recommendations while ensuring purchases stay on their own domains to retain customer data.
OpenAI Pauses Astra Model Over Cyber Risks
OpenAI is pausing internal work on its in-development Astra AI model after evaluations flagged potentially 'critical' cybersecurity capabilities under its Preparedness Framework, including advanced ag
Databricks, Stripe Share AI Coding Cost Playbook
Engineers at Databricks, drawing on conversations with Stripe, Coinbase, Uber, and Ramp, outline a shared playbook for containing exploding AI coding costs — chasing the efficiency frontier rather tha
UNC6671 Vishing Steals SaaS Data via Personal Phones
UNC6671, tracked by Google and Mandiant, uses vishing calls to employees' personal phones to steal SaaS credentials and MFA tokens. Operating under brands like Redact, Pink, Helix, and Falcon, the gro
Google AI Shake-Up as Jeff Dean Departs
Google's AI team underwent a leadership shake-up this week, with veteran Googler Jeff Dean and other notable figures taking new roles — some outside the company — fueling debate over whether Google is
OpenAI Adds Unlimited Chats for Free ChatGPT Users
OpenAI is rolling out major ChatGPT upgrades next week, with free users gaining unlimited text chats for the first time.
China's PLA Air Force Deploys AI Strike Planner
China's PLA Air Force has deployed an AI system that plans large-scale strike operations — prioritizing hundreds of targets and coordinating 100-plus tactical units — as part of Xi's push for an intel
Mitsubishi to build 1,000 humanoid robots monthly in Kyoto
Mitsubishi plans to mass-produce up to 1,000 Highlanders humanoid robots per month at a Kyoto engine plant as early as next year, pairing its automation expertise with a University of Tokyo startup.
OpenAI's First Device: $300 Smart Speaker for 2027
OpenAI's first consumer device will be a hockey puck-sized smart speaker with moving parts designed by Jony Ive's team, reportedly priced above $300 and slated for a 2027 launch, according to Bloomber
Ranking the Best Smart Glasses: Meta, Viture, & More (2026)
WIRED's updated August 2026 smart glasses buying guide sorts the growing category into audio-only, sports, virtual-display, and AI-camera segments, with Meta and Ray-Ban's 7 million pairs sold in 2025
Get the Tech newsletter
Curated tech stories, every morning. Free.
No spam. Unsubscribe anytime.