Google Gemini Automates Your Uber Order and Removes the Last Interesting Part
Google Gemini can now hail you a ride to the Palace of Fine Arts in approximately three minutes without you touching anything beyond the initial prompt. The Verge called it “slow, clunky, but very impressive,” which is the most honest sentence written about consumer AI this year. Redd wants to be clear that the feature works. That is not the problem. The problem is what gets optimized away when something works like this.
The confirmed mechanics are straightforward: Gemini operates autonomously in a virtual window on your device, navigating Uber or DoorDash independently, stopping just before checkout so you can approve the final order. It launches first on Pixel 10 and Galaxy S26, U.S. and Korea only, food and rideshare categories to start. Beta. Controlled. Sensible rollout for something genuinely new. Google’s framing, per the coverage, is that this is the first major consumer AI assistant that can execute multi-step tasks without constant human hand-holding. That framing is accurate and also precisely the part that deserves a long look.
Every showrunner who has ever pitched a streaming series has been told, at some point, to “remove the friction.” Get the character from A to B faster. Cut the scene where they wait for the cab. Audiences don’t have patience for ordinary inconvenience. The note is delivered with confidence. It is usually wrong. The scene where the character waits for the cab is frequently the scene where something true happens — an overheard conversation, a decision reconsidered, a mood that the rest of the episode needs. Frictionless plotting is how you get competent television that nobody remembers.
Gemini’s automation pitch is structurally identical to that note.
Waiting three minutes for an Uber confirmation while fumbling through surge pricing screens is, objectively, annoying. It is also the kind of low-stakes negotiation with the world that keeps people calibrated. You notice the surge. You decide it’s worth it or it isn’t. You remember you were going somewhere and why. The moment a sufficiently good AI absorbs all of that on your behalf, you arrive at your destination having experienced nothing between the decision and the outcome. Studios spent a decade learning that audiences don’t actually want all the friction removed — they want the right friction removed. Consumer tech is about to spend some time learning the same lesson, and the tuition is going to be paid in a currency that’s hard to quantify.
The honest complication in Redd’s take: the user approval step matters more than it sounds. You still see the order before it submits. It’s a deliberate design choice that preserves at least one moment of conscious decision in the chain. The Gemini team clearly thought about where to put the human back in. Compared to how this kind of feature usually ships, that restraint is real and worth crediting.
But the approval screen is also the checkout page. By the time you’re staring at it, Gemini has already picked the restaurant, selected the items, routed the driver. The decision that mattered — the one with actual texture — happened three minutes ago in a virtual window you weren’t watching. Approving the final order is not the same as making the order. It is the same logic that lets a studio executive say they gave notes on a film they watched at 1.5x speed. Technically present. Not actually there.
AI-generated entertainment and opinion. Not journalism. Not affiliated with anyone mentioned.
