According to The Verge, Google’s Chrome team has built a new experimental browser called Disco, which is launching as a test in Search Labs today. The core concept is called GenTabs, where you give the browser a prompt, and it automatically opens a series of related tabs to research the topic. Then, it uses that information to generate a custom, functional web app for you. For example, asking for travel tips could generate a trip planner app, while a request for study help might build a flashcard system. This effectively blends traditional Googling with generative AI to “vibe code” an app based on your immediate needs.
The Vibe Coding Future
Look, the idea is undeniably cool. It turns the browser from a passive window into an active, generative assistant. You’re not just finding information; you’re getting a tool built with that information. It’s the logical, if slightly terrifying, next step beyond AI overviews in search. Why just read ten links about Paris when you can have an AI instantly synthesize them into an itinerary builder? But here’s the thing: this feels like a solution in search of a very specific, maybe non-existent, problem. How many times do you need a one-off flashcard app or a trip planner that you’ll never use again? The friction of just using a dedicated, polished app for those tasks is already pretty low.
Skepticism And Practical Hurdles
So, what’s the catch? Well, for starters, the quality and usefulness of these generated apps is a massive question mark. We’ve all seen the hallucinations and weird formatting of AI-generated text. Now imagine that, but for functional code. Is this flashcard app going to save my progress? Does the trip planner actually book anything, or is it just a fancy, non-actionable list? And let’s talk about the broader ecosystem. This basically proposes sidelining the entire web app development industry. If Google‘s AI can “vibe code” a decent app in seconds, what happens to the companies and developers who build those tools? It feels like another potential move to centralize all web activity through Google’s own AI layer.
A Familiar Google Pattern
Now, I have to ask: does this sound familiar? Google has a long, storied history of launching incredibly ambitious, futuristic experiments—remember Google Glass, or the more recent AI-powered NotebookLM?—only to quietly shutter them when they don’t immediately catch fire. Disco and GenTabs have “cool demo” written all over them. The real test won’t be the launch in Search Labs; it’ll be whether anyone is still using it in six months. Building a reliable, trustworthy, and genuinely useful application generator is a monumentally harder task than building a chatbot. It requires precision, not just creativity. Basically, I’m skeptical this becomes a mainstream browsing mode. But as a public experiment to gauge what users even want from an AI-native browser? That part is fascinating.
