According to Digital Trends, Google Labs has launched a new experimental browser called Disco, built to test ideas for the future of the web. Its standout feature is GenTabs, powered by Gemini 3, which creates interactive web applications by combining information from a user’s open tabs and Gemini chat history. In demos, it built a dynamic trip planner pulling in itineraries, maps, and tips, and created other tools like meal planners and study aids. The browser includes a chat column for refining these apps with natural language prompts. Google has opened a waitlist for users on macOS to try Disco, stating that the best ideas from this experiment may eventually make their way into the main Chrome browser.
This is way bigger than a sidebar chatbot
Here’s the thing: everyone’s been slapping a chatbot sidebar into their browser and calling it “AI-powered.” Google‘s doing that with Gemini in Chrome right now. But Disco and GenTabs show they’re thinking about something fundamentally different. This isn’t about answering questions alongside your browsing. It’s about replacing the browsing session itself for complex tasks.
Think about it. How many tabs do you open to plan a trip? Flights, hotels, a dozen restaurant reviews, a map, a Google Doc for the itinerary. It’s a mess. Disco’s vision is you just start, and the browser itself assembles a single, live dashboard from all that scattered info. It’s not just fetching links—it’s building a custom application on the fly. That’s a completely different paradigm. And honestly, it’s the first browser concept in a while that actually feels new, not just incremental.
Can it actually work without being weird?
Now, the big question is whether this can move from a slick demo to something people actually use. The potential for friction is huge. Will these generated apps feel intuitive, or will they be clunky and confusing? How much control do you have to tweak them? The chat column for refinement is crucial, but getting AI to understand “make the map bigger and move the hotel list over here” is a non-trivial challenge.
Also, it’s entirely dependent on the quality of Gemini 3. If the model misinterprets your tabs or pulls in wrong information, the whole generated app is built on a shaky foundation. Google’s smart to test this in a separate, experimental browser called Disco. This isn’t ready for prime time Chrome. It’s a lab experiment, a “discovery vehicle” as they call it. But it gives us a clear look at the trajectory: the browser is shifting from a window to the web to an orchestrator of the web.
So what happens next?
If this vision sticks, it puts a lot of pressure on other browser makers. Microsoft with Edge/Copilot, or even niche players like Arc, have to think beyond the chat interface. The battleground becomes which browser can best synthesize and act on the information you have open, not just retrieve it. It also hints at a future where the line between a website and a browser-generated app gets really, really blurry.
Basically, Google is prototyping a world where you don’t go to a travel planning website. You tell your browser your goal, and it becomes the travel planning website, built just for you. That’s ambitious. Maybe even a little scary. But it’s definitely not boring. You can read more about the tech behind GenTabs and Gemini 3 if you’re into the nitty-gritty. I’ll be watching this one closely.
