Summary
Paul J Lipsky breaks down five recent updates that he says turn the Gemini desktop app into an AI “super app,” and shows how to use each one. The demos run on the Mac app, though much of it is also available on Windows and mobile.
The headline change is the new switch between Chat and Task. Chat is for questions and brainstorming, while Task runs Gemini Spark, Google’s always-on agent, to actually do work. Tasks can now be pointed at a folder on your computer, where the agent can read, create and edit files. Lipsky demonstrates this with a fictional plumbing company’s financial files, asking Gemini to produce a presentation-style report saved back to the folder. Gemini also creates a backup of connected folders in case something goes wrong. Tasks can additionally be launched from the Gemini phone app.
Other updates include connectors that let Gemini work with third-party tools such as Granola, Canva, Dropbox and Wix, and skills, which are reusable prompts or workflows that can be called with an @ mention. Lipsky shows how to create a skill, such as an inbox manager, simply by describing it. The video is a practical guide for anyone wanting to get more out of Gemini’s agent features.
📺 Source: Paul J Lipsky · Published October 18, 2026
🏷️ Format: Tutorial Demo
