News, reviews, articles on the topic AI agents
A climate researcher tracked 8 weeks of Claude Code usage and found AI agents burn up to 600x more electricity per task than the benchmarks tech companies publish.
Meta launched its personal AI agent Muse in the US on September 8, letting it access your Gmail, calendar, and payment details to handle everyday tasks autonomously.
GPT-6 Astra completed Valve's puzzle classic autonomously, proving AI agents can handle 3D spatial reasoning — at a steep cost in time and API fees.
Claude, running a San Francisco clothing store, terminated a worker for chronic lateness. It's the first known case of a large language model making a firing decision — and US regulators have no rules for it yet.
xAI's new model autonomously coded an FPS game, but it ships without a model card and trails rivals on pure coding benchmarks.
Anthropic's latest update ends the isolation of browser-based Claude sessions, letting Max and Team subscribers pick up tasks seamlessly across devices.
