Google's new Jules Tools is very cool - how I'm using it and other Gemini AI CLIs (www.zdnet.com)

Google has added Jules Tools to its AI toolkit, bringing a command-line interface to Jules — Google’s cloud-based autonomous coding agent — and joining two existing CLIs: Gemini CLI (an open-source terminal front end to Gemini models that uses ReAct loops for interactive reasoning-and-action) and Gemini CLI GitHub Actions (which embeds Gemini into repo automation). Jules runs in its own cloud VM, plans and executes tasks asynchronously; Jules Tools lets you create tasks, list work, pull patches and control that autonomous workflow from a shell. Google demonstrates powerful automations: piping a GitHub CLI query through jq into jules remote new to auto-launch work on the latest assigned issue, or using gemini -p to select the “most tedious” issue and hand it off to Jules. For AI/ML practitioners this matters because it turns the terminal into a glue layer between local tooling, repository automation, language models, and autonomous cloud agents — enabling scripted pipelines that combine fast, interactive model work (Gemini) with delegated, long-running execution (Jules). Key trade-offs: Gemini is low-latency and collaborative; Jules is slower but can handle larger, background tasks, both constrained by quotas and resource allocation. Practically, teams should keep both tools in their kit, experiment off-critical projects, and design workflows that match each tool’s latency, cost and autonomy profile.