Disco Challenges the Traditional App Model  ## Document Notes ## Highlights - **Tabs Turned Into Temporary Apps** - Google Disco treats temporary browsing tasks as full web apps you can generate on demand. - This signals a shift where code is commoditized and single‑use apps become feasible and cheap to create. Transcript: Jaeden Schafer Google has just announced a new product called Disco. This is essentially a Gemini-powered tool. It's going to work inside of the browser, and it is for making web apps, like apps, from browser tabs, things that you're doing. They want to turn, like, a task or something you're doing into an app. Now, this was kind of like a crazy concept to me when I was thinking about it, and the first thing, I'll go over examples and how this works, but the first thing I want to bring up here is I'm Actually really excited. I feel like Google has finally gotten their mojo back. They definitely were caught slacking when chat GPT dropped, it felt like they were getting behind they weren't it didn't feel like they were quite up to par with a lot of the other AI innovation That was coming out. It felt like they had stifled AI because they were worried about their ads business for a long time. And all of a sudden, it feels like they finally seize their moment. They're coming up with all these really crazy products i love the fact that they are taking big shots right now big swings on a ton of random stuff it feels like they're basically throwing Everything they can at the wall seeing what sticks seeing what people likes it feels a lot more like a you know like a startup where they're trying a lot of different things they're okay If things fail this is like one of those instances where i'm like curious to see if this is going to be popular, if it's going to fail. But you've seen them do this with a lot of different products. Some of them we don't hear much about anymore, but some things like Notebook LM seem to be really popular and you keep hearing about them into the future. So I feel like this quote unquote disco product is one of those. Essentially how this works is, I'll give you an example so like you might be studying uh for a test or something or you might be studying a specific topic and this quote-unquote gen tabs Thing is going to all of a sudden have a pop-up that suggests you build a web app to visualize the information that you're studying so you might be studying like i don't know the solar system All the planets might have a pop-up that says you have a demo on one of their sites where it's like um you know, like make a 3D model rendering of the whole solar system. So you can like scroll around and look at what the solar system and the spacing of all the planets is in, you know, real life or whatever. So like, it's a cool concept. To me, it just sounds like absolutely insane. But I think that we because it seems like, wow, that's so much compute. That's so much work just for like, you know, I want to like maybe visualize something. But like, did no one else build this entire, you know, 3D model? Am I the first one to do this? Do we need to have like 10,000 people building a 3D model of a solar system to look at? I want to really reuse the code. It feels like weirdly wasteful. And I know to me, and I know this is probably coming from my background, which is in, you know, I've been working on software products for the last um number of years and for so long before AI came around we were just doing it from scratch this was in you know it took a very long time these tools were like resources you they cost a lot of money to build and all of a sudden we could Just do them at the snap of a finger um and it's it's like you it's like single-use plastic it feels like you make it and you throw it away so anyways I think it's just like such a fascinating Concept but I think it is the future of where we go with this, because essentially, this code has become commoditized. These models are so good at writing code that you can build any tool that you ([Time 0:00:00](https://share.snipd.com/snip/38156aed-3812-4d2f-8d7b-d20dcb634dea)) - **AI Lowers The Cost Of Building Niche Tools** - Disco exemplifies a broader trend: AI makes bespoke UI/code creation trivial and fast. - That lowers the barrier to build niche, single‑purpose tools that previously demanded heavy engineering. Transcript: Jaeden Schafer These models are so good at writing code that you can build any tool that you want very quickly. So why not build it? I guess another example of this, if you're trying to do something less academic, over on TechCrunch, they examples which to be honest i don't love these examples because i feel like They're so overused but they recommended if you're trying to build like a meal plan you had a whole bunch of tabs open of like recipes it could all of a sudden make like uh like a literally Like an app that is software that has all of your meal plans and all of your recipes like built into it and you can click through them all like and then i guess you could save this and come Back to this like app that you have built I don't really know how useful that is I mean it's like conceptually I'm looking at like a visualization of what this tool would look like it looks Really cool like this meal planner thing but I just don't know how useful that workflow actually is here's another interesting one that I saw um Google themselves on their like big blog Post, they put out about disco, they put out like an example video of someone using this. So essentially, in the example, the person is trying to take a road trip, and they want to they're saying like, like, I'm taking a road trip from Boston next month, I want to do sledding And some other things. Can you help me plan? And this is something that you could definitely ask an AI model. So that's nothing like new or revolutionary. And it's like basically Gemini that just goes through and gives them some ideas of like things that they can do. They then follow up and they're like, you know, I'm interested in, you know, local markets and some festive things to do in those cities. And it shows them some of those things. And then comes the pop up, which is, it says, you know, I could help you create an interactive app help you explore some of your options. And you click on this button, which says create gem or gen tab. All of a sudden it goes through just like all the vibe coding apps that I'm super familiar with. I think a lot of people have been playing around with and it creates this whole app, which is basically like a Google Maps overlay. And then there's all of these filter buttons where there's like activity, there's a calendar where you can select what days you're going there. There's like these little bubbles all over the map of where each activity is with a custom icon for it. And you can scroll through like all these filters at the top. I guess it's really cool custom map, right? You can imagine someone could have spent a lot of time programming, developing this thing and putting on a website for, you know, like, travel planning uses. And then it's cool because you can actually edit the code directly from the chat thing. They're just like, okay, cool. Now on the calendar, put what the weather is. And there's like, all of a sudden, icons pop up on each day of the calendar, whether it's going to be like sunny or snowy, and all of that's there. It's like, it's so fascinating, because someone's literally developing this tool, they can like click on all the little buttons with all the activities. And there's a pop up to a link to the website and like a picture and what the activities rated on like Google my business or whatever. So anyways, it seems like just a very interesting world to me where essentially, we're able to just create apps and products and tools right off the bat for a single use. ([Time 0:02:52](https://share.snipd.com/snip/48de7e13-23a0-4a5d-b7dc-9e0ab2891692)) - **Google Keeps Disco Separate From Chrome** - Disco ships as a separate Mac app rather than a Chrome feature, suggesting Google experiments outside its main browser. - This separation hints at product testing and risk containment for compute‑heavy features. Transcript: Jaeden Schafer Um, they say, take the web for a fresh spin. That's kind of their, their slogan on this new disco tool. Okay. So overall, I think this is interesting, but it's not available to everyone. You have to join a wait list. If you want to download the app to start using it, it is only for Mac OS. So it's, it's its own standalone app, sort of like Google Chrome. Now what I think might be happening here, because this whole thing is getting like promoted on the Google Chrome YouTube channel. So this feels like it is being promoted by like the Google Chrome. It's becoming its own browser extension. But what's interesting is it feels like they're not taking Google Chrome and turning it into its own head on head competitor with OpenAI's Atlas AI browser, or Perplexity's Comet AI Browser feels like they're making a brand new thing, this disco, right? This AI kind of enabled tool. And I think Google didn't want to just do what everyone else is doing, which is basically build an agent into the browser, or perhaps they'll do that with Google Chrome, they wanted to Make a brand new product. So this literally is for like creating apps, which Gemini is very capable of doing from your browsing experience. To me, it's interesting that this isn't a plugin, perhaps on Chrome, or maybe even just a feature baked in that they're doing this completely separate, perhaps they're worried that People will hate it, or it'll be way too compute intensive, I think it's probably the bigger thing. And so that you people wouldn't want to have that on by default. ([Time 0:06:06](https://share.snipd.com/snip/41db00e2-b09f-4aa6-a4ad-aab1a36a6132)) - **Disco As A Strategic Backup** - Jaeden suggests Disco could be a strategic hedge against antitrust risks affecting Chrome. - By building a separate AI browser product, Google preserves an alternative distribution channel. Transcript: Jaeden Schafer They're making an entirely new product, seeing if it can stand on its own. There's also I know I've heard some rumors from people, but I don't actually think there's too much to it. But they're saying, you know, because of the the department of, you know, the lawsuit, essentially the antitrust lawsuit against Google, where they were saying like, Oh, Google's Gonna have to spin out the Chrome browser and as you know, sell it to someone else or spin it out as a separate company. You would also imagine that Google wouldn't want to put too much, you know, all of their eggs into one basket, they could make this new AI browser disco. And then if for some reason, which I don't think it'll happen. But if for some reason, this lawsuit actually goes through, which would be bad news for them, and Chrome had to get sold or sent out to another company, they could still keep this new browser And be like, well, okay, we sent off our big, you know, browser or whatever, but we have this new shiny browser called Disco, and you can do all this cool stuff with it, and it's completely Separate. And so it's kind of interesting it feels like even if google was threatened with one of their core business you know pieces of their business or software of their business they could At right now because of ai like reinventing basically every piece of software that exists they could just have an ai version like their new ai enabled whatever browser that they could Default have installed on all the android devices so they still have great distribution for it and the chrome browser can go sail off into the sunset under new management. Now, I don't think that's going to happen. I think they'll be able to keep Google Chrome. But I think this is kind of a sneaky backup plan to make a completely new product in the case that that happens. ([Time 0:07:31](https://share.snipd.com/snip/8feb1e03-b760-4498-b386-49e9ac076560))