UCLA’s Light-Powered AI Detects Deepfakes With Nearly 98% Accuracy
UCLA researchers built an optical AI system that screens 15 videos at once and spots deepfakes with 97.79% accuracy using far less energy. Here’s how it works.
wartono
Always-On AI Gadgets: Why Apple and Google Say They’re ‘Not Recording’
Apple and Google are building always-on AI devices that summarize what they see and hear but claim not to record. Here’s why privacy experts are pushing back.
Personal Agent Protocol: Meta and Sierra’s New Standard for AI Shopping Agents
Meta and Sierra unveiled the Personal Agent Protocol, an open standard that lets AI agents sign in to businesses. Here’s how it works and who’s missing.
OpenAI Releases 722 AI-Generated Math Papers: What It Means for Research
OpenAI published 722 math manuscripts produced by an unreleased AI model. Here’s what was released, how it was made, and why mathematicians urge caution.
GPT-6 Sol and Luna Explained: OpenAI Cuts API Prices in Half
OpenAI’s GPT-6 Sol and GPT-6 Luna cost half as much as GPT-5.6. Here’s what each model does, pricing, availability, and what the benchmarks really show.