Tech
How Token Pricing Works Across Large Language Models
Output tokens in large language model APIs cost 3 to 10 times more than input tokens, while prompt caching cuts repeated prefix charges by up to 95%.
Tech
Nvidia Board Probes CEO Jensen Huang Over Affair With Suspected Spy
Nvidia's board opened an independent probe on Sunday after CEO Jensen Huang told staff he had an affair with Alice Huang, a businesswoman U.S. intelligence agencies suspect is a Chinese spy.
Tech
Apple CEO John Ternus Pushes Craig Federighi to Speed Up AI Roadmap
Apple CEO John Ternus reportedly pressed software chief Craig Federighi in a closed-door meeting to accelerate the company's AI roadmap as iOS 28 adds six extra weeks of coding time.
Tech
Ocarina of Time Switch 2 Remake Runs on PC With DLSS 5 and Hookshot
Modders have the Switch 2 remake of The Legend of Zelda: Ocarina of Time running at ~50 FPS on Windows PCs via Hookshot and 60 FPS on Mac with Nxnative.
Tech
SpaceXAI’s Free Grok Add-On Runs in Google Docs, Sheets, and Slides
SpaceXAI's free Grok add-on puts an AI side panel inside Google Docs, Sheets, and Slides, though users must grant read, edit, and delete access to all Docs and Slides files.
Tech
Gemini 4 Argon Beats GPT-6 Astra and Claude Opus 5.5 in AI Sailing Race
Google's Gemini 4 Argon won a virtual yacht race built with Claude Code at one-fifth the token cost of runner-up GPT-6 Astra, while Claude Opus 5.5 stalled and Qwen 3.5-9B hit a beach.