The Vergecast - A very human vision for going all-in on AI

Episode Date: December 7, 2025

AI models are very good at summarizing things, finding other things like those things, and helping you find those things again. But does that mean we should leave all the work of finding and understanding to those models? Sari Azout, the founder of an app called Sublime, doesn't think so. For this episode, the second in our two-part series about how developers are using AI and building models into their products, Azout explains how Sublime tries to balance being a thoroughly human-focused app with the efficiencies that come with new technologies. She has thoughts on curation, taste, and the differences between AI as a creative partner and AI as a creative replacement. Further reading: Sublime From Sari's newsletter: What matters in the age of AI is taste From The Atlantic: Good Taste Is More Important Than Ever AI Is a Lot of Work Making human music in an AI world Subscribe to The Verge for unlimited access to theverge.com, subscriber-exclusive newsletters, and our ad-free podcast feed.We love hearing from you! Email your questions and thoughts to [e-mail] or call us at 866-VERGE11. Learn more about your ad choices. Visit podcastchoices.com/adchoices

Support for the show comes from Retool. Too many companies run critical operations on duct taped spreadsheets, Slack workflows, and whatever else they could cobble together. Not because they want to, but because building internal tools means weeks of waiting on someone else's backlog. That's where Retool comes in. Build custom internal tools just by describing what you need. Prompts something like,

Build Me a Revenue Dashboard on our Salesforce data. And Retool actually builds it on your company's data, in your cloud with enterprise security built in. Go to retool.com slash Verchcast. We all need to retool how we build software. Welcome to the Virchcast, the flagship podcast of personal knowledge management. I'm your friend David Pierce,

and please believe me when I say that if I could make this show entirely about to-do list apps and no-ticking structures and the para method, I would. But for your sake and frankly, also for mine, that is not what we're going to do here.

What we are going to do here is the second part of our two-part mini-series about how developers are using and building with AI in their products. Basically, I reached out to a couple of people who make apps that I really like and just said, can you come explain to me what AI does, what it doesn't do, and how you think it fits into this product. We had a lot of fun with Thomas Paul Mann from Raycast last week. And this week, I'm talking to Sarie Azute, who builds a tool called Sublime. And the way I would describe Sublime, you'll hear Sarah describe it too, but the way I would describe it is as a bookmarking tool and then some, right? The idea is you collect stuff, you find a link that you like, or you highlight something in a Kindle book, or you journal into Sublime. And not only does it sort of put everything in one place and let you make collections and let you search stuff, it tries to use that to start a sort of discovery journey. So if you save a quote from an article that you really like, you open up that quote on Sublime's app, and it'll also show you this sort of endless feed of other interesting things that you might enjoy based on that quote. Sometimes it's tonally similar. Sometimes it's by a similar author. But the idea is it's sort of an explorer system built around the stuff that you save. I think that's very cool and really interesting. And I have actually really enjoyed using Sublime to just like dump things that I like. I read a lot for my job and just like putting in articles that I like in quotes that I've.

find interesting and all that stuff in there and just sort of using it as a database to find even more cool stuff has been really powerful. At Sublime, they're using AI to make a lot of that stuff possible. And I think the thing that's so interesting to me about Sublime is that Ceri, as you'll hear, is a very human-oriented person, right? Sublime is not some like whiz-bang new AI app. It is very like deliberately sort of quiet and relaxed and it's not a high tech feeling thing. If anything, it feels like sort of old internet in a way, but there's a lot of AI underneath all of that. And I wanted to talk to Siri about how to make those two things match. How do I have a place that feels private and mine and human and sort of interesting and curated and cared for

when it's just, you know, chat GPT all the way down underneath. Is that disingenuous? Is it possible? Do those two things eventually run into each other and cause conflict? This is all the stuff that I've been thinking about with, frankly, every product I use. And I think lots of developers I talk to are thinking about this. How do you use AI in useful ways and not just sort of shout loudly about AI to make venture capitalists want to give you more money?

There's a lot of that. That's fine. More power to the venture capitalist. I think the best apps are the ones that are going to figure out how to strike this balance. And I think Sublime is actually doing it really well. So that's what we're going to talk about. Siri is always a good time. She and I've met a few times and I'm very excited about this.

All of that is coming up in just a second. But first, I have to go and make sure that nothing in my Sublime library is actually going to get me in trouble in case anybody finds it after we do this episode. This is the Vergecast. We'll be right back. Support for this show comes from Shopify. Every thriving, successful business has to start somewhere. A good place to start is a relatively simple question. What if, given the right tools, I've really put my all into this.

One tool that can help grow your sprouting business to new heights is Shopify. Millions of businesses around the world rely on Shopify for e-commerce. They offer a host of helpful tools you can take advantage of, from payment processing to analytics to website design. Their design studio includes hundreds of templates to help you create the exact website you've been envisioning for your business. If you're wondering, what if I need help, then no worries because you're never left to fend for yourself. Shopify's award-winning customer support is available 24-7. It's time to turn those what-ifs into a thriving business with Shopify today. Sign up for your $1 per month trial today at Shopify.com slash vergecast.

Go to Shopify.com slash vergecast. That's Shopify.com slash vergecast. Sarah Azute, welcome to the Vergecast. Thank you, David, for having me. Excited to be here. Siri just went off when I said that, which is just a very funny thing to have happened. Happens all the time. I believe this.

I think the place I want to start is I just want you to describe a little bit of the story of sublime, because it's a thing I find really fascinating. You and I have talked about it a few times in the past. But for people who don't know, what is sublime and kind of where did it come from? Yeah. I mean, Sublime is in a very concrete utilitarian way. It's a personal knowledge management tool. I didn't start wanting to build a personal knowledge management tool.

I wanted to start by building a different way of being on the Internet. And what we've built really is a way to curate and build a personal library. The catch for Sublime is it's connected to other people's libraries. And I think that unlocks just a different experience on the Internet, where you are connected to other people. There is serendipity, but fundamentally, the way you're interacting with a tool is as a personal knowledge. management tool and not some performance circus. One of the things I always liked about Sublime is it seems like, and I'm curious if this

your experience, there are sort of people who get that basic activity of like just save things that mean something to you in sort of the broadest possible way, like things that make you feel something, put them in a place and trust that sort of something will come back about that. There are people who like instinctively get that and find that activity even sort of on its own really cathartic, just like the act of bookmarking does something for some people. And then there are other people who are like, what in the hell is the point of this? I'm just creating this massive, you know, text file of nonsense that I'm never going to look at again. And it really feels like

you can sort the world into those two groups of people and there is nothing in between. I mean, 100%. That said, I would argue that it's been a niche thing because the tools that serve this need don't have a big ROI. Like, what's the point of storing this? And I think if you experience Sublime, you save one thing and you go down a rabbit hole that is just like you don't regret the minutes spent. It's just unlocking a different experience. I do think a big part of it is what's the point because the ROI of that in the past has been I've got a massive archive of things that I will do nothing with. The timing that you've been building Sublime is like right next to this crazy AI boom. Did you think about this as sort of an AI adjacent product from

the beginning because I think spiritually, like, it both is and isn't in ways that I really want to talk about. But like, were you thinking about, you know, what, what some of these language models might do in sort of the early days of building out this product? I think the question of what is your AI strategy or starting even with tech first is very foreign to me. Like even just companies or founders that tell me I'm building a Web3 startup. I'm building a B2B startup. That tells me nothing about whether are, you know, like, that's like a trend that is very easy to, like, mistake for a real market and a product opportunity. So the way I think about it is, what is a problem that I want to solve?

And in what ways are the current technologies available going to allow us to solve this in a better way? If you think about a tool like Sublime, there's, broadly speaking, three jobs to be done. We help people collect ideas. We help people connect ideas. And then we help people connect ideas. And then we help people create with those ideas. So looking at those three functions. So on the collect front, I am very opinionated about this is taste driven. This is judgment driven. Only the user knows what deserves to be in their personal library. We have connections. You can plug in your Instagram bookmarks, Twitter bookmarks, Kindle highlights, you know, all the things. But the act of just like the judgment of what deserves to be here, we don't use AI. You could theoretically have, you

of some argument for somebody that's just like monitoring what you're doing in your browser and deciding what's worth adding. We decided not to do that. Now, there are some things in the collect experience that we use AI for where there is utility and not judgment. So, for example, podcast magic. With just a screenshot, you can send the transcript of a part of a conversation that you like straight into your library. There is no, like you made the judgment call, but we use advanced AI to go from screenshot to a moment. Nobody needs to know that AI is a part of that process. It just is invisible, which is, I think, a big part of the way we approach AI is it's invisible. I remember I got feedback in the early days by, like, sophisticated tech investors saying

we should say, like, this is AI driven and our landing page should have like, you know, AI powered connections. And there is an argument that the window of, there is a window of opportunity that opened because people were more open to trying new things. But I think for the most part, once the shock and off fades, people go back to the baseline and they just want their problem solved. So I think a big part of it is just play the long game. It doesn't matter what is powering this thing. It matters that it's solving something interesting for people. So that's the connect piece. So collect connect. Then there's the create piece. The create piece is interesting because this is where users need a lot of control. So you may.

might have a collection of really cool. Maybe I'm collecting ideas for sublime merch. I've got a collection of cool things that I've curated. One potential interesting thing is how do I combine all of these things into something new? These are my sources and references. If AI is a generalization machine, it's going to retreat to average. What can AI do with my sources? So ideally, it can recombine them into something new. Now, I think the problem is not a model problem. It's an interface problem where at some point the sort of like AI-assisted iteration stops being great and you need that like manual tweaking. And that's where I think chat sort of like ends. Like I just tried so many times to tell chat GPT like change this in this image and it all just kind of breaks. So I think

we've been very cautious on the create front because we've, we've just tried. We've just We've seen a lot of people just like panic to add AI features everywhere, and people don't like it if it doesn't work. So that's sort of where we're at with it is I think it is interesting to think about source-grounded creation. But I think the quality is so subjective and users need a lot of control. And I think a lot of the disappointment with tools is the lack of control. Yeah. Yeah, I mean, I think I like that framing of it. And I think a thing you and I agree on is that that connect phase is actually.

something these models are all really, really good at. One example I give all the time is if you want to just do a useful, valuable thing with chat GPT or any of these other tools, get movie recommendations. Tell it a bunch of movies you like and let it feed you other movies that you might like. It's low stakes in the sense that, you know, if it's wrong or it hallucinates a movie or whatever, like, that's not the end of the world. But it is able to just sort of like open up a bunch of doors. for you and be like, are any of these anything? And sometimes that's not what you want. And sometimes

that's exactly what you want. And I think in a case like this, one of the things that I immediately gravitated to about Sublime is that idea that, okay, not only am I going to save a bunch of stuff and I like build this library full of things that are meaningful to me, but also it's going to help me extend that library because every time I save something, I save a quote from an article I like, as soon as I open that thing up, it just gives me a like an endless fee. of other similar, I guess is the word I would use, things. But this goes to one of the things I'm thinking about is like, as you're as you're tuning these systems and building them, how do you even decide what success looks like, right? Because again,

in the sort of old way of thinking about like movie recommendations, it's like, okay, if I like this movie, I probably like other comedies. I like movies by the same director. I like movies at the same actors. Like we had, we had parameters, right? And a generation ago sublime would have been the same thing, right? Like, oh, I saved a GQ article. You probably like GQ. You probably like articles by this person or about this person or about this subject. You're trying to do something much fuzzier. Yeah. And it seems like both a sort of opportunity and a challenge to figure out, like, what is that fuzzy thing that you're trying to do? It's a huge challenge. It's a huge challenge because it's not deterministic. Like, the software in the past is rules-based. This is, you know,

this is every piece of content that you add to sublime has an embedding. And it's just going to show you things that are like semantically in proximity to that embedding. And, you know, we try to tweak certain things here and there. I mean, there's this whole field of AI evils that is emerging. But I think your point is very clear. I, what I like about this in particular in this a blind use case is that I think people think of hallucinations as a bug. And I think in this context, it's a feature. Sure. It's the fact that it's going to take you down a path that is completely unexpected that is, that is, that is interesting. But I think your point is clear. And in fact, it just changes software development because in the past, I would just, you know, I've got my

project management tool linear and here's the ticket and do this. And it's like approved, done. And you can be in endless cycles of iterating. And in fact, when I, I test a lot of this stuff myself, you know, we've got a database of millions and millions and millions of things. I can run tests on 10 things today. And by the way, the output, like for you is going to be different for me with the same card. So it's, you know, I think we're still learning, but it's a completely different era. Yeah.

I mean, it does seem like you have to develop like a philosophy around what's good on that kind of thing? Because I assume one thing you've going for you is that everybody has saved all of this stuff into the database, right? Like by definition, your database is full of things that people have gone through the effort to save. in one way or another. And that is like a everything,

every single one of them has some signal of importance, right? Figuring out what that is and what that looks like much harder. But it is like you've sort of deleted all the crap from the internet just by virtue of what the tool is. But then like deciding what is good to put next to my stuff is so hard. Yeah. I mean, it's interesting. I learned a lot about recommendation systems from Ken Stanley,

who wrote an incredible book and something he was saying about social media and the way that algorithms work on social media that really resonated with me is if you are driving down a highway and you turn around and look at a car crash, the algorithm interprets that as, you know, you love looking at car crashes.

And then, you know, it starts to show that to everyone. And so it's very based on, you know, if anyone that's experienced going viral online, once you're past like a certain number of likes, it just keeps like cascading. The way that we think about recommendations on sublime is we do have a human page rank of sorts, meaning you added this card or idea to sublime, and then I added it to my library. So there's some signal there that two people like this. What we are saying is past a certain threshold, which is very low, it could be three people.

It doesn't matter if three people have added it or 100. It's just a basic threshold. And then there's just a different level of openness. So we have some philosophy around, let's not just show the obvious stuff, you know, because that's everywhere on the internet. And part of the philosophy with sublime is how can we help more people find the non-obvious stuff on the internet? So there, you know, algorithms are. You know, like all of this stuff is like opinions embedded in pixels.

So there are, of course, underlying opinions and decisions that we're making that are philosophical. But I think your point of how open-ended and subjective this is remains true. Yeah. And it does seem like going back to what you were saying about the collect idea, like it makes making it, giving people the tools to make that particular corpus of content in the way that you're describing becomes super, super important. Right. And I think this is where I get into like what is sort of the human AI relationship in all of this. And I think a thing that you and I agree on is that the best thing you can do for your AI tool is just feed it good data that you care about. And it feels like we're in this place now of like, you know, I think you talk about the sort of collect, connect, create loop. And I think what a lot of AI companies want to do is just collapse that into like all sort of one thing where like you do a prompt and then the AI does the rest. And it's it's doing the collecting. and the connecting and the creating for you and that all of it is just sort of happening and collection and creation are the same thing and that to me is like a bleak future

that I kind of hate but I love this idea of like I'm going to build a library and database of stuff that I care about and then the tool's job is to help me make meaning and something out of it but it does start with like it's a hell of a lot of work to make people make that corpus of stuff Yeah, I mean, so a couple things. I think I 100% agree and believe that the value of dumping all sorts of context about any given thing in one place has gone up with AI.

Whether it's coming up with a marketing plan, all of your research for a book, like just centralizing all of that in one place that you can dump and then export easily, that has immense value. because AI grounded in your sources is just going to unlock immense value. I feel like Notebook L.M was the thing that turned a lot of people's brains onto this. That it's like, oh, just stick a bunch of PDFs in a place. And then you can sort of talk to this tiny corpus of data about whatever you're working on. It was like one small version of this thing that made people go like, oh, when I tell you what matters and then we do something about it, like cool, cool things start to happen. 100%. So this context question becomes more and more important.

it. And I think we, you know, notebook alum is probably more about like sales and just the positioning is a little bit more enterprisey. But I think similarly with, with creatives, I mean, some of my favorite use cases for Sublime could be something like, I have a collection of phrases I love. And when I'm creatively stuck, I will just, we have a one-click button, I'll open chat GPT with those sources. And I'll be like, you know, help me. get unstocked used, like for this other thing I'm writing, using these references. And it's just fun. You know, I also think people take AI so seriously, and a lot of it is about speed and volume, but some of my favorite use cases are just where it's not about speed. Like, I would argue a lot of

the places where it's helping me is, it's not necessarily faster. It's just better because I'm constantly iterating and spending more time with ideas, because the cost of doing so is lower. Yeah. It does seem to me that that there is something, I don't know, sort of essentially pro-human about sublime that like, and you and I've talked about this before and you write this news that I already talk about it all the time that is like this is not supposed to feel like mining a database of, you know, AI generated highlights from magazine articles. Is it, is it, is there a risk the further you go into? Like we're building these sort of cool AI tools that are going to help make. sense of all of this where like it stops feeling like this sort of human discovery process. Can the AI get too good at this over time? I think that in a world of AI, you can think of companies as automate the things people don't want to do or help people do more of the things that they choose to do, that they get to do, that they have more time to do. And I think that we are designing sublime to be a joyful experience. And I think if you just keep your eye on the kinds of,

the feeling that you want to orchestrate, you don't get, like, distracted. So I think part of, I think, building in this day and age is putting blinders on, you know, the way like race horses are. You could look around and everyone's like, you know, all tools look the same. It's like a big prompt box and, you know, V-Zero and Bolt and Love, blah, blah. And I think it's just very easy. to be part of that like mimetic cycle. But the antithesis to that is just what is the feeling that you want to generate for your customers? And stay, keep your eye on that. The big prompt box is actually exactly why I brought this up.

Because to me, there's a really easy version of Sublime to build if you wanted to that says, okay, we have this giant database full of interesting, meaningful stuff to lots of people. And what we're going to do is we're going to build Sublime bot. and you're going to use it to query all the greatest quotes in history, right? Like, tease this out long enough and it's like, okay, we presumably have the 10 most relevant meaningful quotes on any subject you can think of. And we're going to start to, we're going to start to collate that for you. We'll make a wall calendar for you with meaningful.

Like, there's just sort of a million ways you could go if you start the process with prompt this database that we've built, right? And you can like sort of do that in the search of sublime now, But sublime really doesn't want you to do it that way. It really wants you to start by saving something or making something and then go from there. And I feel like that at some level has to be a deliberate choice, right, to say this is not fundamentally like a thing you can go look at. It's a thing you do first. Yeah.

So that was very intentional. I don't think that what you're describing is antithetical to the vision, meaning I can imagine a world where we've grown to, you know, a very, very large database, where people might want to interact with that database very differently. Like, you can imagine thinking, what's on your mind today? And instead of posting on Twitter, how many people actually post on Twitter? The vast majority of people are just consuming. So you start with a thought, an intention, something that's on your mind. And then you go down this rabbit hole of related ideas. I do think that intention over attention is one of my guiding principles. So instead of a you feed, it's what do you want? I think there's something very fundamentally different about

that that feels important to me. But I think there's a lot of ways in which we can extend the functionality that we already have and enable new kinds of experiences and still stay true to that sort of feeling. It's nuanced, right? But there's something that gets very flat and boring about a world where everything is served to you in summaries. And there is, I think, like, as with everything, there's an opportunity to be at the extremes. For us, it's never going to be about just the summary. It's here's some very interesting, cool substack

by this person that you didn't know that relates to this other thing. And I don't think it's mutually exclusive with the other reality. I think we'll just live in a world where both coexist and serve different needs. All right. We got to take a break, and then we're going to come back with more from Sarah Azute about sublime. Very back.

Support for the show comes from LinkedIn. If you're a small business owner, you know that every hire counts, but time and resources are limited. Finding, connecting with, and screening the right candidates takes up valuable time you could be giving to your customers. That's where LinkedIn Hiring Pro comes in. It's built to be your hiring partner, helping you find the right candidates faster. That way you can hire with confidence without turning it into another full-time job. Hiring Pro streamlines the entire process from drafting your job to shortlisting candidates and conducting AI-powered interviews for initial screenings.

Its updated conversational interface lets you describe what you need in plain language. Nearly 60% of hirers find a candidate to interview within a week. With Hiring Pro, you spend less time searching and more time connecting with the right talent. And instead of getting buried in resumes, you get a focus shortlist that actually moves your hiring forward. Join the 2.7 million small businesses using LinkedIn to hire. Get started by posting your job for free at LinkedIn.com slash track. Terms and conditions apply. So you mentioned podcast magic.

And I want to talk about podcast magic because I think it, to me, it's sort of a perfect example of how to build. as far as I can tell, like an extremely AI-driven, essentially like AI-dependent product that doesn't, that still sort of hits the human spirit of the thing you're talking about. This is pure like AI as tool, not as lifestyle decision, right? And I am like very bullish on AI as a tool and very bearish on it as a lifestyle decision, just to put my own cards on the table. Just tell me a little bit about podcast magic and kind of where this thing came from. Yeah.

So podcast magic essentially is the simplest way to capture. insights from podcasts. It came from both a personal need of sometimes I listen to David on the Vergecast saying something really smart. And I am, you know, constantly that happens all the time. And I am walking, working out with my kids out for a walk, whatever it is. And the only way I get to kind of retain that piece of information is when I'm back on my laptop or there's some like gymnastics I could do with other tools. But the normal thing, people do, which is like the existing behaviors often take a screenshot. It's such a thing that people do. In fact, when I tell everyone this idea, they're like, yeah, I have already a lot of

screenshots of this stuff. Of course, they never go back to them. So I remember one day seeing somebody add a Spotify link to Sublime for a podcast. And in the notes, they said, minute 1114, Lenny is talking about, you know, whatever. And I posted it in our team chat and I said, this is so silly. People want the equivalent of the Kindle highlight for podcasts, and it just doesn't exist. And Alex from our team said, this is probably a silly idea, but it should be a screenshot. And I was like, that's not silly. And it just my mind immediately went to, you know, we've got public RSS feeds. We've got OCR. You know, this should be close to possible. And so I spent close to a year, actually, because there's, you know, as much as people want you to believe that coding is

easy and agents will do everything, like reality has, you know, a lot of details and edge cases. And it took a long time to get it right. But fundamentally, the way it works is you take a screenshot from the podcast player screen or the lock screen, or you could use the iOS shortcut and just say, hey, Siri, podcast magic. And then within 30 seconds, you have an email that tells you, here's broadly speaking, the moment that you were listening to. Here's the transcript. And it's going to use AI to figure out, like, oh, it's not 30 seconds or a minute. Depending on what you're listening to, we will use AI to kind of extrapolate what it might be. But again, user control is important. So we're building ways for you to be like, oh, actually, I really like that sentence or I wanted to

start earlier later. So user control is very important. And then, of course, the podcast magic clips sync to Sublime. So you can discover related ideas, semantically, search them. It is for us a very strategic top of funnel play for Sublime, where Sublime is this ecosystem, the people that love it, are obsessed with it. But like you said, it's, you know, it's either you get it or you don't. I think podcast magic is more of a toy with an immediate aha moment and it's working very well as a top of funnel. But yeah, I think for us, the AI, like we don't use the word AI and the promotional materials for it at all. Doesn't matter.

How we make this happen doesn't matter. What matters is that you can now capture insights from podcasts the same way you highlight books on a candle. Totally. But it is, I mean, the way you described it, like it is basically AI the whole way down, right? Yeah. So once you send a screenshot,

we will use that image to determine what is the podcast, what is the podcast title, what is the timestamp? We need those three things. Then, of course, we do all sorts of mechanics to index, like, the millions and millions of podcasts out there to do this really quickly. And then we use AI to, once we have that transcript, we'll use AI to figure out what's like the right start and end time. And of course, we use, you know, tools for transcription and audio matching. So it's AI all the way down.

Yeah. That's such an interesting one because there's a chunk of that that is technically pretty straightforward, right? Like look at a screenshot and tell me the title of the podcast on that screenshot and what time it's showing is like a functionally solved technical problem. Right. And I think one thing that drives me crazy about especially this like these AI glasses and cameras is, you point the camera at a like a banana or a big bag of nuts that says nuts and huge letters on it. And it's like, look, it knows it's a bag of nuts. And it's like, that's not impressive. We solved this problem so long ago that it can see.

the word nuts and know that it's nuts is not interesting anymore. But anyway, so there's a, there's a chunk of this that is like pretty straightforward. But then you get to the point where it's like, okay, not only are we going to like pull the transcript, which itself, that stuff has gotten vastly better because of things like Whisper the Model from Open AI. But we're going to actually try to figure out not just what was happening at this second, but like what is what is the conversation that was happening around this screenshot that made you take? a screenshot. And that feels like the kind of thing that you just straight up couldn't do not very long ago. Absolutely. I think, I mean, that is something that AI is very, very good at.

It's the meme of take a couple of bullets and make it something big or take something big and make it something very small. That compression, I'm amazed. Like sometimes even as we were designing, because all of these things, it's about the nuance and the details and, you know, I'm always reminded, like, reality has a surprising amount of detail every time we are, like, tinkering with any of these things. But, yeah, I think that in this case, when we decided that we were going to have the moment title from AI, like, it is so good the way that it can just, like, detect, like, what is the heart and soul of this conversation?

So, yeah, I think there's a lot of things AI is good at. There's a lot of things AI is not good at. Broadly speaking, I think that, or my TLDR on this stuff is, AI is incredible at the shallow, technical, rule-bound stuff that has more of a clear answer. Like transcription is a great example of that. I think AI is helpful, but not agentic, in the stuff that is more fuzzy and stuff.

subjective. And a lot of, I think, the challenges with the implementations I'm seeing right now is they're kind of applying this agentic philosophy to things that are not agentic. It sort of seems to me, like a lot of people in tech have fooled themselves into believing that life is a clear sort of input-output mechanism for a lot of things. And sometimes, and again, I want this to work. I, you know, I am not, this isn't like some purity test around, you know, I don't want this stuff. Like I've tried a lot of these tools for generating marketing assets, for doing outreach for, you know, podcast magic is a great example. I've got very little time in my hands. I've got a small team. I would love to deploy AI agents to reach out to a bunch of podcasters and get them to share podcast magic on their show. Try doing that with any of the tools out there. It's terrible. You know, there's just so, so many invisible details. There's a reason why, you know, maybe for some podcaster, I should reach him on LinkedIn.

this one I should do on Twitter. This one really likes concise stuff. For this one, this is the right context. You can't automate this stuff. It just seems to me like we've got wrong the sense of how many things we believe fit that archetype of is a simple structure problem with a concise and consistent notion of correctness. Like, no, most work is fuzzy. Most work is like start in one place, change direction, respond to this like nuanced thing. And this is why I think we're just much further than a lot of people say in just making progress with this stuff. It doesn't mean that it's not incredibly helpful. Like the semantic search alone, the like AI for search as you know, in the context of sublime is amazing. Now, is that a 10x or is that a 10%? I know how to quantify

that. It certainly made my life better. I don't know that it's made my life 10x better. But I think a lot of it breaks down with expectations. What do we expect from these tools? Personally, I think I've sort of replaced the AI agents thinking with smart automation. It becomes far less disappointing when you think about it that way. There's not as much VC money in smart automation as far as I can tell. But I think you're probably right. Totally.

And I think that's what drives is driving a lot of this, right? It's, you know, the narrative for the VCs has to be, we are building a gentic AI for X. but the reality is that the consumer doesn't want that narrative because they don't buy that narrative. I mean, of course they want it if it works. If you tell me you can reliably give me software that will do my taxes every year, I will not say no. The choice is just we haven't figured out how to build that software.

Right. And the road from here to there seems so sort of plausible that everybody has convinced themselves that it is inevitable. And I just don't think it is. I do want to make the case before we move on very quickly. I think search is absolutely a 10x thing, not a 10% thing. And I think this has been one of the things that has most blown my own mind in starting to use this stuff. Like in sublime it feels like this. There's this app called my mind

that I use a lot that feels very much like this. Just the idea of being able to put all your stuff into a place and then trust that you can get out of it, not just like the thing that you're looking for, but the actual either piece of information or related stuff or sort of discovery experience that you're looking for is so sincerely life-changing. Like, the only comparison I can make is when Gmail came out and you could search your email instead of having to go find it. And all of a sudden, you didn't have to categorize anything. You could just archive email. But now instead of searching Delta and opening four emails to go find my, you know, my frequent flyer number, I can just ask my email for my frequent flyer number. Like that, that is a step change in how

useful all of this stuff is, especially if you're somebody who likes to save things. The idea of like, I'm so done with like notebooks and tagging systems and like carefully curating wherever. I've just entered this phase where it's just like, I'm just dumping shit wherever and I have full faith that I can find it again with with just the teeniest, tiniest, tiniest bit of knowledge about it. And that is like once a system like that works and you trust it, you use it completely differently. And I think it is amazing. Yeah, it's the don't organize, just search. We call it vibe search.

And it's the, like this idea that I can be having this conversation with you and be reminded of, you know, what was that saying that Andres Carpathy tweeted that, you know, was vaguely about why the agentic AI is overhyped. And he may not have used any of those keywords in that, but we'll still find it again. I agree with you that for these particular products, there's a 10x. The reason that ever know 10 years ago was not very useful is because it functioned as a collecting place, but the retrieval was really bad. So I do think that the case for why to do this now is, first of all, it's retrievable. There's the connections piece. There's what we discussed

about the importance of having all of your context in one place and you can then extract. I think that there's no doubt that that's a 10x for people using this product experience. Now, for a for the economy, I think the expectation for these tools and what they're going to do for the economy, that's a different conversation. But I couldn't agree more with the search piece. Yeah. And it just seemed like one thing that strikes me as a potentially scary flip side to all of that is, like, you write a lot about taste. And I think Sublime is very much about taste as a platform. And I do wonder what happens to our taste when so much of the process starts to be mediated for us. Like even when a bunch of things that you like are sort of put in front of you,

but you don't know where they came from or why you like them or why they're here or how you found them, they're just sort of here. Even when their recommendations are very good, there's this sense of sort of all this is happening to me instead of me doing it myself, that I wonder if it starts to harm our taste over time. Like I think about this with TikTok all the time. I have never once intentionally informed TikTok. I just open it up and it just does things for me.

And I'm like, do I like this? Is this what I wanted? Is this what I asked? Is this my taste or am I just here? And I wonder, like, you care a lot about taste. Is that can we have it both ways? There's no doubt that our taste is being shaped by these platforms. It's sort of, it's not clear to me that I like certain things or I've just been fed those

things enough times by enough algorithms that somehow that becomes my my taste. I think a lot of what I intend to do with Sublime is have a point of view on culture, have a point of view on how important it is to sit with things that are inefficient, you know, to develop and hone and cultivate your own taste and just have that orientation. But yeah, it is true that that these for you algorithms are shaping us in ways that we are not aware of. Before I let you go here, let's talk about your own use of AI for a few minutes, because I'm curious how even sort of outside of like the sublime ecosystem, where is AI in your life right now?

Are you using it all the time? Are you like moving to the woods and ignoring it forever? Like what's, where's your head right now? I'm using it a lot. I would say I'm a very curious person and AI opens a lot of conversations around what is creativity. What does it mean to be human? And I'm very deeply engaged with these questions

and the tools help me understand the limits. I would say a lot of what AI has enabled for me for better or worse is it's made me more interactive with myself. I'm spending a lot more time just like contemplating. Like I will just go on a walk and I'll use chat GPT voice and I'll ramble for 10 minutes about what's on my mind. and then it'll say something back. In some ways, I worry that I'm sort of mesmerizing my way into thinking this is interesting and useful, and sometimes it's not.

But I've noticed that there is this sort of sense in which I've always had a lot of thoughts and ideas and maybe not the outlets. You know, Twitter is a very particular way of sharing things. My colleagues, sure, but I'm not going to like spill all of my thoughts, you know, 10 hours a day. Sure. My therapist, like maybe, you know, but there's a sense. of like I've got access 24-7 to something where I could just like ramble. And with voice, it seems even more sort of palpable. So I'm using that a lot. So let me actually just pause you on that one because that, that, there's an interesting question there where it's like,

what about that is valuable? Is it the, is it the having sort of a place to do that? Or is it like what it digests and says back to you that is meaningful? I don't know that it's valuable. is maybe the conclusion. I'm still, the verdict is still out there. I think there is a world where we all may be in this collective experiment where we're diluting ourselves into thinking these things

are more effective than they are. I think a lot of it is the novelty of, oh, this thing can tell me something back. I very much think that it is problematic to take whatever they say seriously because I could be, I could tell Chachupit, give me feedback on, you know, this, and it'll tell me this is great.

And then I'll be like, no, it's bad because why. And then I'll be like, yes, yes, yes, you're right. It's bad because why. So, you know, it's just going to validate whatever you're saying. You can't trust these things. But there is some, I've noticed in my usage that it's made, that the fact that it allows me to be interactive with myself in a way that I wasn't before is just a lot of how I spend time with it.

there are some interesting uses like prototyping. For example, like V-0, I use a lot. I very much think that a prototype is worth 100 meetings, that for me to communicate and express my intent to a teammate in words always felt insufficient. And now I can be like, here's a prototype. I actually, for us, those tools are not effective at taking us to production. I think that a lot of the vibe coding hype is for things that don't carry a lot of weight.

But for prototyping in particular, I think it's very effective. I'm an idea person. I think often it's hard for me to communicate the idea. And if I can go to my team with a basic prototype, it just makes everything a lot easier. And so I absolutely love that and use that all the time. Give me an example. Like what's something you built recently that was like a cool, fun prototype? Yeah. So for example, with podcast magic this week, we are building a way for people to edit whatever default was captured. So for example, in a podcast, we're making an assumption about, you know, this was the two minutes that you wanted or the minute. But I want to give users control. What does that control look like? Are you editing the audio? Are you editing the transcript? You know, what is the UI for that interaction? And so I'm starting a conversation with VZ.

I'm stating the challenge and I'm kind of iterating my way to it. It's not going to look visually pretty. It's not going to be aesthetically aligned with our brand. But it's a great starting point. So really for most features, whereas before I would have just created a read-me document with just text, I am including some basic prototype in the way when I'm sort of handing that over to the team. So that's been very, very useful. I mean, I think a lot of the compression of let me blabber about this and like help me structure it into something that I can share with the team is, and in particular, because I'm on the go a lot, I do feel like I get less road rage these days because I'm like talking and I'm like getting worked on because I could work with voice and then it's just like compressing that into a bullet that I can then send to my teammate on Slack.

So that kind of stuff is helpful. My sort of bar for using these tools is hold myself to a standard where the quality of what we're doing is better, not worse, from using these tools. Okay. And so for the writing stuff, you know, I think that using AI makes my writing take longer. Sometimes I like write a draft and I'm like, I want to make this better. Give me some options here. And so I just spend a lot of time in like iterating mode or sometimes I'll just do funny things. things where I'll be like convert this draft into something that like, you know, so-and-so comedian would say. And it's, I'm just having fun with it. And it's not necessarily, sometimes it'll affect the output or take me in a different direction. Oftentimes it won't. I don't know. I think people should just have more fun with it, you know? Yeah. But I think I actually, I like the bar of does it make the quality of the work better? Because that's very

different from is it more efficient, right? And I think we spend a lot of time, to your point, the people who are having a lot of meetings with VCs talk a lot about efficiency. And we are in this moment where it's like, okay, how much worse are we actually all willing to let things be in the name of them happening so much faster and more efficiently? And like the goal is we're just going to reduce everything to just like blisteringly fast. And it's all going to kind of suck, but it's fine because it's all going to happen so fast and so easily and you won't have to do anything that it'll be fine. saying no, I'm actually only going to use this tool when it makes the thing better,

strikes me is actually like a pretty high bar to clear, especially for a lot of current AI stuff that is, it's going to make everything simpler, but it's going to make it worse. I sort of feel like that's the crusade am on. That is what I want to spend the next multi, you know, like a couple of decades on. I think it is our responsibility to wield this tool, to make more beautiful things, to, you know, I think people talk a lot about how machines are becoming more human-like, but we don't talk enough about how humans are becoming machine-like. And I think that's even more problematic. Are we going to just work for machines and

like optimize and try to, you know, feed the algorithm? And to me, there's just a sense of everywhere I look, there's this like, you know, even like the entrepreneurship casino of people online trying to sell you, you know, 10 ways to 10x your productivity by 6. A.m. And I, you know, I think that the, the ideal future is one where actually we are in symbiosis with this machine and we hand over all of the optimization so we can focus on the dreaming. Like, I, you know, one of the, the sort of like things that's been stuck in my mind is, you remember Thomas Edison's like famous quote on, on 99% perspiration and one percent. inspiration. And I think now it's sort of the opposite where it's going to be 99% inspiration,

1% perspiration. Like, sure, you could deploy all these machines to do all things for us, but then the value of judgment and dreaming and understanding what we want to do and being kind of connected to that becomes more and more important. And so I would say that if we zoom out, pre-industrial revolution, we lived in a world where our kind of physical strength was ever. everything. Then we kind of moved to a world where our mental strength was everything, our intelligence, you know, knowledge, work, and, you know, information age. And so what's next? You know, in a world where you're not going to compete with AI on knowing more facts or producing content faster. I think the next edge as woo-woo as it sounds comes from that, being connected to that source of inspiration and heart.

And, you know, I think it's, yeah, our job is to kind of stick around long enough for people to kind of understand that that's kind of the last mile here is really returning to the things that make us human. Yeah. Are you optimistic or pessimistic about which direction we're headed on that front? It feels like we have, like, as a society, good days and bad days on that front right now. You know, I think that both things are going to happen. I don't tend to, I don't think that. I don't think that. I don't think that. that a technology is good or bad. I think that Open AI or ChatGBT, BT, AI in general, can allow people to outsource their agency. Hey, I got this email, help me respond to it. Okay, great, copy, paste. I've outsourced it my agency. But I also think that you can use these tools to augment your agency. Hey, I got this. Help me think about this. Help me think through it. And I think both things are possible. I don't think that the technology is, like I think,

I think technology is neutral. I think it's up to us to design tools that allow us to make the most of this technology in ways that serve our needs. And I think we need more people building companies that put those needs at the center. I think the problem with AI is the people building these tools have no explicit mission. And what I mean by that is you think about the early kind of Internet people. It's openness and the crypto people are about, you know, for all of the problems with crypto, you know, there's this kind of sense of freedom and decentralization. And with AI, my sense is, they're just racing to build AGI, whatever that means. And it's not clear what does that mean for my life? You know, am I going to be a better parent with AGI around? I'm going to have better relationships. Am I going to be a happier person? Like, there's just no connection to what that does for us. And I think there's just a big vacuum of let's just, you know, there's this very, very powerful technology. There's a big vacuum in terms of how it fits into people's lives. And I think that there's a huge gap in let's fill that with positive visions for the future. And that's sort of the way that I see

our role in and sublime in the ecosystem. I love that. I could not agree more. All right, well, I can't beat that. So we're just going to get out of here. Sarie, thank you so much for doing this. This is so fun. Thanks so much, David. All right, that's it for the show. Thank you again to Sarie for being here. And thank you, as always, for watching and listening. If you have questions, or thoughts or other people you'd like to see on this show, this is a kind of conversation I really enjoy having about how we think about product and new technology and user experience,

and it feels very much of this show to talk about this stuff. So if there are other products that you think have interesting paths with AI or there are people you just want me to yell out about all the dumb AI stuff they're doing, I would love to hear from you. You can call the hotline 866 Verge 1-1. You can send us an email, Vergecast at theBurge.com. We check all the inboxes and we absolutely love.

hearing from you. Until then, the Vergecast is a Verge production and part of the Vox Media Podcast Network. This show is produced by River Branson, Eric Gomez, Brandon Kiefer, and Travis Larichuk. We will be back on Tuesday and Friday. We've got a bunch of fun year-end stuff coming up. We've got a lot of news to talk about. It's going to be a good week. We'll see you then. Rock and roll.