According to Computerworld, Google is launching a new experimental AI tool called Disco for the Chrome browser, built on its Gemini 3 model. The tool allows users to describe a task in natural language and then turns their open browser tabs into interactive web applications called “GenTabs,” useful for things like study sessions or travel planning. These generated apps can be further customized with more instructions, requiring no traditional coding knowledge. Disco is directly integrated into Chrome, letting it consider multiple open tabs and previous Gemini chats for context. Right now, the tool is only available on Mac OS through Google’s test platform, Google Labs, with access being limited to a waiting list that interested users can sign up for.
The Bigger Play: Chrome as an OS
Here’s the thing: this isn’t just a cute productivity hack. It’s a strategic move to make the browser, specifically Chrome, the center of your computing experience. By letting you spin up functional, persistent apps from your ephemeral browsing sessions, Google is blurring the line between the web and the desktop. Think about it. Why install a dedicated app when Chrome can dynamically build you one tailored to your immediate needs, pulling from the sources you already have open? It’s a power move to lock users deeper into the Chrome-Gemini ecosystem. The integration with previous chats and tabs is key—it means Disco isn’t working in a vacuum, but with your entire browsing history and intent as context.
Who Actually Benefits Here?
So who wins? Casual users who get intimidated by complex software might love this. Describing what you want in plain English and getting a working app is a compelling demo. But I’m skeptical about power users. Will a “GenTab” for travel planning truly beat a meticulously crafted Notion page or a dedicated project management tool? Probably not for complex workflows. The real beneficiary, at least in the short term, is Google itself. Every GenTab created is another data point on how people want to use information, another reason to stay in Chrome, and another showcase for the Gemini 3 model. It’s an experiment that feeds the core business: advertising and AI model training.
The Limited Rollout Says a Lot
Now, the limited Mac-only rollout via Google Labs is telling. This is a very early, probably buggy experiment. They need to see how people break it, what they try to build, and whether the concept has legs before even thinking about a wider release. It also feels like a direct shot across the bow of other “AI agent” workflows popping up everywhere. Can Disco do what a user might try to cobble together with a custom GPT or an automation tool? Google’s betting that deep browser integration is their killer feature. We’ll see if that’s enough. For now, it’s another sign that the browser is becoming the most important—and most AI-native—piece of software on your computer.
