The Deepdive

How Google’s Disco And GenTabs Turn Browsing Into Getting Things Done

• Allen & Ida • Season 2 • Episode 29

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 13:58

Send a text

We trace the browser’s leap from passive page viewer to agentic toolmaker, led by Google’s Disco, GenTabs, and Gemini 3. Automation, multi-tab intelligence, and a chat-first UI promise finished outcomes—while raising urgent questions about publisher economics.

• the private chef analogy replacing manual tab synthesis
• gen tabs generating custom interactive tools from plain language
• agentic browsing automating tasks and nudging proactive interfaces
• multi tab intelligence consolidating context across all open pages
• conversational browser memory replacing linear history
• google’s strategy with tensor chips and sandboxed disco
• a chat-first, split-screen interface redefining navigation
• the open web tension between perfect answers and source traffic


Leave your thoughts in the comments and subscribe for more tech updates and reviews.

Ida:

All right, I have a feeling this is gonna be relatable for well probably everyone listening. You know that feeling when you're planning something complex, maybe a multi-stop vacation, or you're just deep diving a new topic for work.

Allan:

Absolutely.

Ida:

And before you know it, you've got, what, dozens of tabs open?

Allan:

At least. It's just a sea of little icons.

Ida:

Exactly. It's this uh this digital clutter fatigue. Your computer screen just becomes this messy, unorganized desk, that exact feeling that need to manually stitch together information from 15 different places. That's really the core problem we're tackling in this deep dive.

Allan:

It is.

Ida:

We're looking at a huge strategic shift in web browsing. And it's being spearheaded by Google's experimental disco browser and this uh powerful new feature called GenTabs, all running on their latest AI, Gemini 3.

Allan:

And the stakes here are, I mean, they couldn't be higher. This isn't just some incremental software update. Right. This is being framed as maybe the biggest change to the core browsing experience in history. We're talking about redefining the browser's whole job.

Ida:

Aaron Powell From what to what?

Allan:

From a passive viewer of web pages to what the sources call a transaction-oriented answer engine.

Ida:

A transaction-oriented answer engine. So its job is to get things done, not just show you links.

Allan:

Aaron Powell Precisely. It's about outcomes.

Ida:

Aaron Powell Okay. So that distinction passive viewer versus active agent, that's the mission for this deep dive. We want to understand what this new world actually looks like and you know why Google is making this big push right now.

Allan:

To really get your head around it, the experts offer this fantastic analogy. It perfectly illustrates the change.

Ida:

Aaron Powell I love a good analogy. Lay it on me.

Allan:

They say the traditional browser is like a trip to the grocery store.

Ida:

Okay. Makes sense. You're gathering ingredients.

Allan:

You got it. You, the user, you walk the aisles of the internet, you gather all the ingredients, all those open tabs and articles, and then you have to go home and manually cook the meal. You synthesize it.

Ida:

Which is exactly where the fatigue comes in. We're doing all the work of the chef.

Allan:

We are. But the AI integrated browser with this gen tabs feature, it flips the whole experience. It becomes your private chef. You might still gather some ingredients, but the AI proactively looks at what you have, synthesizes it all, and instantly prepares this, you know, custom plated, ready-to-eat meal.

Ida:

A finished solution.

Allan:

A finished, custom-coded solution right on the screen. It moves you from being this weary gatherer to a relaxed consumer.

Ida:

Okay, let's unpack that custom solution part because I think that leads right into our first big takeaway. The browser is becoming a bespoke software generator. It's not just a display cabinet anymore.

Allan:

This is the really revolutionary part. And the feature we keep mentioning, gen tabs, which is short for generative tabs, that's the engine. Right. The browser isn't just, you know, rendering code that a human programmer wrote somewhere. It's writing its own functional code based on what you ask it in plain English.

Ida:

So wait, it's not just summarizing text, it's actually building interactive tools.

Allan:

Exactly.

Ida:

What does that look like for a user? I mean, normally if you want a custom web app, you have to go hire a developer. Trevor Burrus, Jr.

Allan:

Not anymore. And that's what challenges this core assumption that building software requires some kind of deep expertise. The sources say Gemini 3 can uh quickly code and build visual tools from just a simple command. Let's use one of the examples from the research. It's pretty vivid. Yeah, let's say you're trying to learn about the solar system. Instead of getting 10 Wikipedia links, you just ask, help me learn about the solar system.

Ida:

Okay.

Allan:

And GenTabs immediately generates an interactive 3D model right there in the browser. You can spin it around, you can zoom in, ask follow-up questions about, say, Mars.

Ida:

Wow. That is a huge distinction. That means the web is changing from a place of static consumption to a dynamic application layer. We're moving way beyond just reading.

Allan:

Way beyond. Or think about complex planning. Instead of you copying flight times from one tab and hotel prices from another.

Ida:

The classic vacation planning headache.

Allan:

Right. The browser synthesizes all of it into one cohesive dashboard. It generates a unified travel planner with like zoomable maps, calendars, timelines. It's purpose-plan software, coded on the fly, just for you.

Ida:

And the user never sees a single line of code.

Allan:

Never. But they get all the functionality.

Ida:

So if the AI is generating these really sophisticated interactive tools, that capability seems to naturally lead to the next shift. Our second takeaway: the rise of agentic browsing. The browser starts doing things for you.

Allan:

This is the real pivot point. Chrome and this disco concept, they're moving from being tools for information retrieval to being tools for, well, action.

Ida:

Okay.

Allan:

The AI isn't just sitting there waiting for you to type in a search. It's actually monitoring your context to offer or even perform tasks automatically.

Ida:

And the updates planned for Chrome, they specifically detailed this kind of automation, right? Focusing on those repetitive tasks.

Allan:

You do. The sources promise the AI will handle those low-level things we all hate doing, like booking appointments or handling the steps for ordering something online. But the real standout feature is what they're calling a password agent.

Ida:

That detail really jumped out at me, the idea that it can automatically change your compromise passwords with a single click on supported sites.

Allan:

Yeah, sites like Spotify, Duolingo, Coursera. It's a huge functional leap in security. And it means the AI has to actively interact with forms and menus on your behalf.

Ida:

It shows a lot of trust is being built into the system.

Allan:

It does. But the agency isn't just about you telling it what to do. There's also this proactive element, the sources call the nudge factor.

Ida:

How is that different from, you know, a regular search suggestion?

Allan:

It's way more anticipatory. It's deeply contextual. So imagine you spent 15 minutes looking at three different tourist activities: a museum, a boat trip, a cooking class.

Ida:

Okay. I'm comparing my options.

Allan:

Exactly. The AI sees that pattern, and it might nudge you by offering to build a comparison tool for you before you even ask. It sees what you're trying to do and instantly codes the best interface to help you decide.

Ida:

So it's moving the whole internet experience toward delegation.

Allan:

That's the perfect word. The browser becomes your proactive assistant. And crucially, the sources emphasize you're always in control. You can stop or override these actions at any time, which is essential for building user trust.

Ida:

Aaron Powell And that idea of delegation brings us right to our third takeaway, which is the immediate goal here. To completely crush tab fatigue with something they call multi-tab intelligence.

Allan:

Right. When you have dozens of tabs open, the hard part isn't finding the information, it's the mental energy it takes to combine it all.

Ida:

The synthesis.

Allan:

The synthesis, exactly. The whole value proposition is shifting from just access, finding the tab to synthesis, understanding what all the tabs mean when you put them together.

Ida:

Aaron Powell So this new integration is literally designed to take on that mental load for us. How does multi tab intelligence actually work?

Allan:

Well, it means the browser with Gemini 3 is analyzing everything at once, the content of all your open tabs, your chat history with the AI, and it's consolidating all of it into a single unified view. It's reading the whole project, not just the one page you're looking at.

Ida:

We saw some great practical examples of this. I mean, beyond travel planning, it applies to daily life, like turning a bunch of different recipe tabs.

Allan:

Yeah, from different websites.

Ida:

Right. A dessert from one, a main from another, and turning it all into a structured, ingredient-optimized weekly meal plan.

Allan:

Or taking scattered advice about gardening and compiling it into a cohesive plan for your specific location. It's basically project management for your browser. It takes all these random inputs and creates one actionable output.

Ida:

There's another fascinating part to this though, browser memory.

Allan:

Yes. This is a big one. The sources say Chrome's AI will be able to recall sites you visited in a really conversational way. So instead of scrolling through your history, you could just ask something broad like, what was that blog I read about back to school shopping last year?

Ida:

And Gemini could find the specific page, not just the website, from a year ago.

Allan:

That's the promise. It would transform your browsing history from this clumsy chronological list into an indexed searchable database you can query with natural language. That saves so much more time.

Ida:

Okay, let's pivot a bit from the user benefits to the business strategy, which is our fourth takeaway. This is a major competitive move by Google to keep its ecosystem dominant.

Allan:

This is absolutely Google playing both defense and offense. The deep integration of features like gen tabs is uh a critical defensive wall against users leaving for other tools.

Ida:

Like ChatGPT or Claude.

Allan:

Exactly. For their complex research needs. If the browser itself is smart enough to be your private chef, you never have a reason to leave Google's world.

Ida:

And the sources really emphasize that Google has this massive, almost insurmountable resource advantage that makes all of this possible. Let's talk about their tensor chips.

Allan:

This is the key technical detail under the hood. AI at this scale requires a staggering amount of expensive computing power. Google is leveraging its huge market cap and specifically its proprietary tensor chips.

Ida:

So why is having their own custom silicon so important? Why not just buy chips from, say, NVIDIA like everyone else?

Allan:

It comes down to cost, optimization, and independence. Tensor chips are custom built to run their models, like Gemini 3, much more efficiently than a general purpose chip could.

Ida:

So they get a cost advantage on every single query.

Allan:

A major one. Which is vital for a feature that's so computationally intensive. And maybe more importantly, it makes them independent from the outside chip supply chain, which is a huge bottleneck for their competitors.

Ida:

Aaron Powell So they're not just betting on the idea of the private chef. They've actually built their own custom kitchen to make it economically viable, which is where the experimental browser disco comes in.

Allan:

Right. Disco is explicitly called a discovery vehicle. It's from Google Labs, and it's designed purely to test these big ideas quickly and safely with a small group of users.

Ida:

It's a sandbox.

Allan:

It's a sandbox, exactly. The sources are clear that the most compelling ideas from disco may one day make their way into larger Google products, which is basically code for Chrome. It's where they figure out what actually works before rolling it out to a billion people.

Ida:

Speaking of disco, let's get to the interface, which is our fifth and final takeaway. The whole user experience is changing, moving away from that classic search box that's defined the web for decades.

Allan:

The visual change is uh profound, especially in disco. While Chrome still has the familiar address bar for now, disco just gets rid of it. It's a conversational chatbot experience from the start.

Ida:

So you open the browser and you're just talking to it.

Allan:

You are. You're immediately greeted by the AI chat box, and that effectively is your new address bar. You start everything with a conversation, not a URL.

Ida:

And the layout is different too, to handle this generative process.

Allan:

Correct. It uses a split screen layout. You have your ongoing conversation with the AI in one window, and the custom app or browser experience it generates appears right next to it. It makes it really easy to keep refining what it's building for you.

Ida:

Before we wrap up, we should probably ground this in reality and mention availability. This isn't rolling out to everyone tomorrow.

Allan:

Not at all. Disco is highly experimental. It's only available via a wait list, and interestingly, it's initially exclusive to Mac OS users. So that really underscores how early we are in this process.

Ida:

Aaron Powell Okay, so we have to address the giant tension this all creates, something the sources bring up with publishers. If the AI does all the work, all the synthesis, doesn't that create a kind of AI black box that starves websites of the traffic they need to survive?

Allan:

This is the critical balancing act for the future of the web. I mean, if the private chef makes me a perfect kneel and I never have to visit the grocery store, the grocery store goes out of business. Right. Google is trying to get ahead of this by stressing that, and this is a quote, every generative element ties back to the web and always links to the original sources.

Ida:

So they're trying to thread this incredibly tiny needle.

Allan:

They are. Deliver a perfect custom coded result, but still give credit and hopefully traffic back to the sites that provided the raw data.

Ida:

But the question is, is that sustainable? If the AI gives me that perfect 3D model or a full weekly meal plan, how many people are really going to click through to read the original recipe? They already have the solution.

Allan:

That's the billion-dollar question. The sources imply Google has to get this balance right for the health of the entire ecosystem. If they fail, they could push all the high-quality content behind paywalls, which would, in the long run, make their own AI less valuable.

Ida:

So the browser's success as an agent is totally dependent on the health of the open web that it's drawing from.

Allan:

It's an existential tension. Can you create that perfect plated meal and still get the binder to visit the farms where the ingredients came from? The answer to that really defines the economics of the new web.

Ida:

So to synthesize everything we've uncovered, the browser is going through this radical change. It's moving from a simple viewer to a proactive agent that builds bespoke apps and manages our complex projects. This huge pivot, fueled by Gemini 3 and tested in disco, is this aggressive move to finally end tab fatigue by turning scattered data into ready-to-use solutions. It's less about browsing the web and more about building things from the web.

Allan:

And that integration of eugenic browsing and deep synthesis just fundamentally changes the relationship between a user and the internet. If the AI successfully aggregates everything, codes that perfect custom planner, and delivers a finished outcome, it leaves us with one massive question.

Ida:

That's that.

Allan:

When the consumer is no longer rewarded for browsing static pages but for consuming these custom coded results, how must content creators and publishers evolve their entire strategy, their website design, their value proposition to stay relevant and get traffic in this new agentic internet?

Head Shot

Allan

Host
Head Shot

Ida

Host