Gemini Mac App

Gemini shipped a Mac app yesterday afternoon. And of course I grabbed it to play around with for a few hours, and the first thing I did is probably the most useful thing I'll do with it all week.

I asked Claude to write me a "who I am" summary I could hand to another AI. What I work on, how I like to be talked to, what's been on my mind. One prompt, one paste into Gemini, and a tool that thirty seconds earlier didn't know me from anyone could actually hold a conversation. Not perfect. It missed all the half-baked stuff, the loose ideas I've kicked around with Claude that never turned into anything real. But the working picture was there, and that was wild to watch happen that fast.

Here's the prompt if you want to try it yourself:

"You are helping me import my context from one AI assistant to another. Please generate a comprehensive, structured summary of everything you have learned about me, my preferences, and my style through our conversations. Please include:

• Personal Background: Profession, education, and general residence if known.
• Interests & Preferences: Sustained hobbies, active projects, and topics I frequently discuss.
• Communication Style: Preferred tone, length of responses, and formatting rules I like.
• Relationships & Dated Events: Significant recent activities or key people I've mentioned."

A couple things stuck with me after.

If your company already lives in Google Workspace, and a lot of companies do, agentic AI just got a lot closer to where your people actually work. Option + Space, screen sharing, local files. That's a different thing than "go open a browser tab."

And this app is just the warm-up. Apple and Google announced earlier this year that Siri's next chapter is going to be powered by Gemini, and we'll hear more at WWDC on June 8. Today's Mac app is the first place that partnership actually shows up on your desktop. If you use a Mac, or if you're an admin supporting people who do, this is where things start to get genuinely interesting.

Two assistants at once is its own adjustment, and a bit scary (so much information spread across different second brains). Claude knows my work. Gemini is starting from scratch. I'm honestly excited to see how my use cases sort themselves out between them, which jobs go where, and why.

What I'm watching for at WWDC: how deep Siri actually goes with Gemini under the hood, what that means for the stuff IT teams have to think about, and whether "share your screen with the assistant" stops feeling novel and starts feeling like the default.