Everything's a Fad (Including This Podcast) — with Benn Stancil

Transcript#

This transcript was generated automatically and may contain errors.

Welcome to The Test Set. Here we talk with some of the brightest thinkers and tinkerers in statistical analysis, scientific computing, and machine learning. Digging into what makes them tick, plus the insights, experiments, and OMG moments that shape the field.

Today we're sitting down with Benn Stancil. You may know him from his writing on Benn with 2Ns.substack.com, where he's just as likely to weave in Lorde lyrics or Olympic gymnastics scoring drama as he is to talk about SQL.

Benn spent a decade building Mode, one of the early modern analytics and BI tools. We talk about how he got started, we talk about the layers of hype inside the tech ecosystem, and we talk about why he's still optimistic about writing in an AI world where output is cheap.

I really like how he uses AI like a thesaurus, idea generator, but not necessarily as a writer. And I have to say, for someone who comes off roundly as a pessimist, Benn gives us a lot to be optimistic about. So without further ado, Benn Stancil.

What used to be sort of artistic sports became extremely quantified around these like kind of weird rules and stuff like, yeah, what, how many, how many seconds do you have to protest?

Became like what decided the medal. That's the rule, but I don't, okay.

From Nate Silver to Rick Rubin: the vibe shift

Yeah. And I, I find it so interesting because I do feel like I really saw myself in that like tale and the point you made about sort of like, if the CEO hits me up and asked me like, what should I do? And I break out all these formulas when they could go to like chat GPT or something and get like a very quick answer. I'm curious if you could recap that to like this idea that we used to be very, or like being really quantitative and model oriented was a huge virtue, if I'm understanding correctly, but that now there's this world where we can actually, like, get answers. It's all vibes.

Yeah, and that was kind of — it's like, before, and then the gymnastics case, it was vibes. It was, you're an expert, tell me how I did. Okay, so who does the best with vibes? And like, I don't know, this person presumably knows more than we can all dispute it, but it's kind of vibes.

To me, the data stuff was — and this is sort of the, everything is a fad — it's better, kind of. But a lot of it is, like, that was the culture. It was the culture of, like, Nate Silver became popular and things like the OkCupid blog, a bunch of people got into this stuff. And our generation basically got into the, like, well, actually, let me show you all the numbers and why you're — and that became, like, a thing, a signifier of how smart you are and, like, how successful — all this stuff.

And I think we have very much — there has been a lot of pushback to that, somewhat, in some places anyway. Things like this gymnastic stuff happens when it gets so quantified, where it's just like, my God, what are we doing? It's been a year. We still don't know who won this thing. These routines are 90 seconds long. And there hasn't really been an alternative, but now, like, AI kind of is an alternative. You can just be like, what's the vibe? You can give it — don't go and analyze a thousand support tickets. Give a thousand support tickets to ChatGPT and send me, like, what's the vibe? And if I can do that and get an answer in five seconds, and I can do that as much as I want, like, I'm not going to go ask an analyst to give me some giant report of a bunch of, like, numbers that don't align. And they're like, well, did you think about this way or this way? But it depends on how you count what a user means. Like, what's the vibe?

Yeah. And I feel like we're about to cut Wes loose on this here, because I feel like Wes has a lot of opinions. But if I can just put a thread through it, I feel like it's so interesting to hear you describe, like, the Nate Silver, OkCupid blog era, which you mentioned, were kind of, like, inspirations for the original mode blog posts. And then the shift into today of more vibes, which is — I feel like a lot of the topics we hear right now, like, on your substack about, like, AI and people vibing out.

Yeah, there's a lot of — there's definitely a lot of topics there. I mean, I think it is interesting. You know, I think all three of us have spent a lot of our careers basically focused on building tools and environments to make it easier for people to do things with data. And that might be, like, writing SQL queries and computing things and looking at the answers or making plots. Or in the case of Pandas, it was like, I just want to be able to read CSV files or get a little bit of data out of the database and, like, clean it up and wrangle it and get the answer. And surprisingly, in 2007, when I started working, I was shocked at how difficult it was to undertake, like, those basic data questions.

But yeah, I mean, now we're in this, you know, now we're in this weird era where, like, the whole concept of, like, usability of data tools and the environments in which we ask those questions is being completely turned on its head. And so, like, I find myself, you know, for probably the first time in my career, like, essentially not writing code, like, just looking at code and, you know, pressing return more or less. And so, you know, the way that we work and, like, the shape of our tools and how we ask questions, get answers and make decisions is certainly, you know, hasn't completely changed yet, but it's, like, on its way to changing. And so I don't think we have a clear picture yet of what that's, you know, what that's going to look like.

And there's a broader, to me, like, kind of cultural thing in all of that, which is in 2014, 2012, 2010, 2014, whenever, like, data was kind of a status thing. It was big tech startups had big data teams and, like, not, like, capital B big data, but they had data teams, lots of people on them. And they were, like, look at that one of the things they would sort of show off is how, like, smart they are about making decisions because they're, like, being so data driven and all these things they do. And, like, you go to a VC meet, like, you go to a VC pitch or you, like, go to a board meeting, they want to see all the numbers. And, like, that was kind of how you represented that you were, like, with it is you did these sorts of things.

Now, to your point, not just in sort of, like, vibe coding type of stuff, but if you talk to companies, like some of the sort of leading AI companies, not necessarily like the anthropics in everybody's eyes, but things like companies like Cursor, those sorts of that cohort of things that, like, the big companies with not a ton of people, data is not a big part of what they do. It's very vibey. Like, Cursor famously has said, like, the way that they basically test their agent harness stuff is mostly vibes. They're not really, like, an evals driven thing. And I think that broadly is happening with, like, taste in Silicon Valley is a big thing. It's like, oh, we don't need someone to be, like, you know, tweaking the numbers based on an A-B test that has this, like, half a percent lift. It's like, what's good taste? And so I think, like, status has become a lot less about how well do you understand the numbers and more of just, like, are you, what's the guy's name? Rick Rubin. That's, yeah. Like, Rick Rubin became this weird, like, that's who we all want to be, not Nate Silver.

Rick Rubin became this weird, like, that's who we all want to be, not Nate Silver.

Yeah. And I think that, like, culturally becomes a go. We don't want to do the data stuff. We want to have, like, somebody who's, like, who's the artist that can tell us, can just, like, conjure the great ideas in their head. Yeah. Maybe that works, maybe it doesn't. I'm not saying it's necessarily better, but I think that's at least, like, the ascendant cultural trend.

Yeah. It's so funny you mentioned Rick Rubin. I mean, maybe this came from one of your posts, but I was just thinking about that, like, where I think he starts an interview with, like, a brief, like, minute of meditating with the interviewer. And I'm like, sure, that's what, yeah. And then he is like, well, I don't really like, I just feel it out, like, the music and, but it's, yeah, it's hard to deny the, like, impact. And if you had done that in 2012, everyone would have been like, oh, it's like a hippo thing. It's like, oh, your intuition's totally wrong. You've got to, like, watch where the users go and look at, look at, like, what's actually are they doing? Don't tell me what you think they're doing. And now it's like, if you're Rick Rubin, you sit and you have a minute of meditation and you're just saying, like, I just intuit it. People are like, oh, that's the guy.

Yeah. Yeah. Yeah. Which again, I don't know, maybe it's better, maybe it's worse, but it's at least the thing that we, like, that's who we want, I suppose. Yeah. Yeah. I'm here for it. I, as a data person, you shouldn't be here for it. What are we doing?

AI, vibe coding, and the new era of software

I mean, it's, I, I have a lot, right. And it's not, not simply because I've been, I've been, um, I, I don't like to describe what I'm doing as vibe coding. Like, I feel like vibe coding has almost become this like pejorative term, like vibe coding means like, I don't care about the output. Like if it seems like it works, it works. Like, you know, like I, I still care about like, you know, code quality and like how fast my test suite runs, like how fast the code runs, like the longterm sustainability, like growth of the code base. Like, I think that a lot of, a lot of like vibe coded software is gonna, gonna reach some kind of a breaking point. If there's not like a reinvestment in like architectural quality and like, you know, essentially design patterns. So in a sense, like design patterns have become more important than ever. Like, uh, right now it's like an army of agents, uh, building, you know, these giant, you know, hairballs of code bases and, uh, and, and they're, you know, riddled with bugs.

Um, how we reckon with that and how we manage that, um, time will, time will tell. But I think the, you know, the, the big thing is like, it's both like the, one of the most exciting times in, in my life in terms of technology. I think that the closest analogy I can think might be like the beginning of the internet, like 1995, like I was 10 years old in 1995. And so I remember like how exciting it was to get my hands on Netscape Navigator and start using the internet. Uh, and I, you know, I was too young to do anything meaningful in that era. It was certainly, you know, didn't, I didn't program, like I wasn't going to go off and like start a company. I was interested in like video games and I did GoldenEye speed running on the early internet in the late nineties, you know?

Um, but, but now it's like 30 years later, it's like everyone is, it's like this collective, like fixation on like this thing. Right. And everyone's trying to figure out like, how do we get the most out of this thing right now? Either like, how do we start companies? Like how do we, we become rich and like be the first, like be the best, like deliver the tools, deliver the, the interfaces, deliver the environments, deliver the solutions to make this new, like fundamentally interesting and like powerful technology work well for, for people.

And so what, it's both exciting and fun and I'm having a lot of fun, but there's also like this weird kind of like joyless grind that's happening where like a big part of the tech industry is like working harder and longer than they ever have before. And like, essentially like a lot of people are not having any fun. And that to me is like weird. And maybe it's because there's so much money sloshing around in the ecosystem. There's so much money to be made evidently from, from this. And so there's a lot of, you know, young 20 somethings who were like, yeah, this is my opportunity to like, you know, work 120 hours a week and have, you know, 15 cloud code sessions going all the time. And then maybe I can become a billionaire if I keep that up for as long as possible. So I think there will be some kind of like, you know, collective kind of community burnout that happens at some point.

But it is very exciting. And I, you know, definitely I'm like on board with this new way of working. Like I think there's no going back and I don't want to go back, but how to actually make it work, like how to build software, like how to do analytics, how to do data science. Who knows? Like it's, it's, I'm, I'm, I'm questioning, I'm questioning everything, like kind of rethinking everything. So I don't know, just kind of interested in like what's been your experience and how you've like gotten to know this new technology and like incorporating it into your, your world.

I, so one, Goldeneye doesn't hold up like you want it to. Like today? No, it really does not hold up. Is it the square? It just, it doesn't look good. It's like, kind of clunky. I mean, it's still like, it's still charming. Like, I still, it has a special place in my heart. But yeah, like, I got one of those, like, the analog 3D, it's like an FPGA clone of an N64, and like, plugged in Goldeneye, and like, I played for a little bit, and I was like, it had its time and its place, you know? Like, you want it to be like, oh my god, this is gonna be, and you're like, ah, not so great.

Okay, so the answer to your question, I guess I would say a couple things. On like, using the things and what happens, the analogy I've used with this before is like, it feels a little bit like when Google came out, that, when Google first came out, I don't know, I was in like, middle school, high school, something like that, or whatever, and people were like, oh my, this is the new thing, you gotta learn how to use it, you gotta learn all the tricks, you gotta learn all like, the search, search booleans, you gotta learn like, query languages for Google, and how to like, exclude sites, and don't exclude this, include these terms, don't use them, it's like, okay.

And then obviously, like, 10 years later, nobody knows how to do that, like, you just get a feel for Google, you get a feel for the thing, people who are, grow up with it, like, understand how to, you wanna search for something there, people can find anything, and it's not because they're using any sort of fancy search parameters, it's just like, you kinda get a feel for the machine.

Broadly, I think that's what'll happen with all of this, is like, there's a whole bunch of like, the internet grifter type of, here's how to get the most out of your five clawed codes, and blah, blah, blah, and it's like, hey, this'll all kinda come together, and best things will emerge, and you'll just get a feel for it. So I think there is a lot of like, stuff around that, that can be kinda, and like, what emerges from it, I don't know, somebody will figure out some crazy thing, and they'll release a product, and like, Clawed Code is an example of this, but like, all of a sudden, kinda came from nowhere, not really, but it went from being like, this kinda side thing, to like, oh, this is the only thing anybody ever talks about.

To your point, though, about like, the like, Silicon Valley burnout stuff, this is a half written post that we'll see if I can actually make this analogy work now. Earlier this year, there was a guy, he was an engineer, Steve Yaggy, wrote a post called Gastown, which was like, very viral when it came out.

Basically, he was like, how do I turn my Claude Code into the most insane thing you could possibly imagine? And it's like, there's one, like, mayor thing, I don't know. Gastown is a reference to Mad Max, right? I think so, I think so. Where your orchestras are like a convoy. Gastown is like, yeah, it's like a city in Mad Max universe, yeah.

But it's like, there's like, a king, and then the king has all these other, like, bots that do things, and there's, it's agents all the way down. There's like, four or five layers to it. And he basically says like, this thing burns enormous amounts of money, for a lot of noise. Basically, the point is like, a thousand things do something, and one gets picked up, and that's, you know, you try it a thousand times, and most of them are trash, and a few are good ideas, and one of them, only one can actually get chosen, because we're only implementing one thing.

And the imagery he uses, because it's this Mad Max thing, imagery he uses for it is like, it's like, men in tar pits. It is not like, a lot of like, the future of technology is like, some, that's like, the paradise of the floating cars and stuff. And this is just like, nah, we're all in tar pits.

And there's a part of that that feels like, that's actually a little bit of what the world is right now, is, like, this actually is all kind of a, we're all like, just in these tar pits, non-stop, to where, to your point, yeah, there's a thousand companies in San Francisco right now, two of them are going to work out, because like, only two can. You can't have a thousand billion dollar, like, it's not going to happen.

And so there's going to be a whole bunch of people who are like these agents, the point of their work is to be thrown away. Like, everybody's figuring out how to build stuff, and most of those things can't work, because they're all building the same thing, and one's going to win. And I think, like, we haven't really reckoned with that, like, not as a, it's like, okay, it'll be fine, a bunch of people try a bunch of stuff and it won't work out, but it feels like, if you go to SF, there is this sense of, like, oh, my God, we're all, like, at the beginning of the internet, and we're all going to be enormously rich and famous, and it's like, ultimately, that's kind of like a status game, and there's only so many people at the top of that pyramid. Like, status is a zero-something, and in some sense, like, business success is a zero-something. There's only that much money people are going to spend, it's not like everybody can sell, you can't sell a thousand $10 million contracts to Coca-Cola, you can sell, like, two.

And so there's going to be a whole bunch of people who are like these agents, the point of their work is to be thrown away.

And so, I don't know, it feels like there's a, and one way to look at it is, like, oh, my God, this is, like, so much potential and creation and possibility, and I guess that's kind of right. Another way to look at it is, like, oh, this is sort of a new, a certain salt mine that feels sort of fun, but ultimately, I think a lot of it is going to be, like, just industrial waste.

The software abundance problem

Yeah, I mean, like, my optimistic take, and like, I am, you know, ultimately, you know, even though there's a lot of, like, questions about the, you know, the financial kind of private credit bubble, like, the infrastructure bubble, debt infrastructure bubble, and its effect on the economy, like, the over-concentration of the, you know, US economy in AI, essentially, like, the S&P 500, like, NASDAQ has essentially become an AI play. If you're in the US stock market, essentially, you're long AI, whether you want to, want to be or not, you know, your index funds, your 401k and all that.

But I am also, on the course, the technology, I am an optimist, and, you know, what I've been telling people is that initially, there was fears of, like, AI is going to replace software engineers, and we're going to build the same amount of software, it's just going to be built by AI. But actually, no, we're going to build 10 to 100 times as much software. And yes, it's going to be built by AI, but there's still going to be a lot of professional, professional software engineers. But to your point, if we build 10 to 100 times as much software, who's going to buy it? And like, how, essentially, how much money is there going to be to, to buy, to buy software?

And one of the memes that's been going around Silicon Valley lately is this, this concept of, like, software as a service is dead, which I think is a little bit, a little bit hyperbolic, because, like, some software as a service products are hard to build. Some turns out are, like, not, like, there's a lot of to do list apps, like, I'd say probably to do list apps are in trouble. Especially, you know, people can build their own, you know, I've built my own to do list app with Claude code, you know, why not?

But I think certain classes of software will become very difficult to, you know, very difficult to monetize as they're, like, you know, free open source versions, or you can, like, roll your, you know, you know, pretty soon, like, I think right now, it's you still need to be a pretty good software engineer to get good output out of, out of Claude code. But over time, like, you may not need to be able to. And so maybe a year from now, or two years from now, you'll be able to pull up your phone and voice interface and say, make me a new to do list app that has these characteristics and come back a day or two later, and the agents will have built your, your iOS or Android app, and it will be installed on your installed on your phone.

But I think, like, nobody really knows, like, what that, like, compression of the software industry looks like, like, so much more software fitting into effectively the same amount of space. It's like, you know, populations only growing so fast, like, yeah, maybe it will increase GDP, like US GDP, like world GDP, but not by 10 to 100 times, not even by a little bit. And yeah, and even if it does, like, ultimately, most people want to be rich, because they want to be richer than their neighbor, not because they care about how much money is in their actual bank. Like, there's a lot of this is just status to me.

But I think, to me, the easiest analogy for it is just like, content creation, basically, that sure, it used to be to make a movie, you got to go to Hollywood, and it's got to pay a bunch of people do whatever. Now, anybody can do it. There's a lot more people making movies, there's a lot more movies, we only have so much time, we've given a lot more time to staring at TikTok than we did when we had to go to movie theaters, whatever. But like, it's not like everybody now is making there's a whole bunch of people who are, like, on the cusp of sanity, trying to become TikTok influencers, who that's obviously a very viable career. And for some people, it's been great. But like, there's a lot of people who are like, kind of on the fringe of that. And it, it seems kind of bad. Like, it seems very difficult if you're there. And there's a lot of a lot of stuff that's that edge of like, yeah, you've got enough to maybe make a thing, but like, not really. And like, that feels like closer to what the software stuff will be.

Like, if when we started Mode, there was a starting a software company was like, a serious endeavor, not serious in the sense that like, we knew what we were doing. But like, you had to go out and raise money. You had to spend a bunch of time building a thing that from the day we started the company to the day we like, first put out the first product that anybody outside of it ever saw was close to a year. Like, and that was seen as being reasonably quick. And we like, made a product that was pretty simple for us to build anything that had any kind of like, meaningful feature set. It was going to take several years and take millions of dollars of like, spending on engineers.

Um, it was very hard to get something that was like, sellable at all to market without spending a bunch of money, a bunch of time. And that didn't mean everybody who did it there, therefore it was like, uh, uh, like real professional, but it just meant that there weren't that many people doing it, and if you did, you spent a lot of time thinking about it, and it was just like, this was a lot of work, and you kind of tried your best. And I think the same was true for, like, content before. Like, you want to write a book? Great, you write a book. It used to be, you got to sit down and write a book. It's gonna take you a long time. You may be a bad book, you may be a bad author, but it takes a long time. Even if you're gonna write some bad beach read, you got to write every word, and that, like, takes some time. And now you can write a book in 10 minutes. You can create shippable software in two days.

But there's also, to your point, like, we can't read every... If everybody writes a book in 10 minutes, we can't read that many books. Those books aren't going to get read. Like, it's just not possible to consume that much content, or read that many books, or use that much software. And because you can create a thing that you could have created 10 years ago in two days, it would have taken you two years before. So what? Like, the market isn't growing at the same rate. It may be growing a lot. It may be doubling. I don't know. But when you can create 10 to 100 times as much, you know... What's the market for video? It's way more than it was 10 years ago.

But there's probably a thousand times more video that people are creating now than they did 10 years ago. And so, like, there's a whole lot of people. The people on the edge to me is, like, a whole lot of people who try to make that a career. But, like, they're making stuff that seems pretty good. It looked like it should be of, like, professional... It's like, it is good enough. But when there's that much, how do you... A lot of it just doesn't catch, you know? The market is extremely saturated. I think software will be like that.

It's a funny issue. It's interesting because, yeah, if a person spends a year of their life on a book, I feel like we're... I mean, they're going to die one day. Like, we're all going to die one day. And I think for a person to spend a year of their life on a book makes me think, like, oh, that's kind of remarkable that a person would dedicate, like, a big chunk of time to this thing. Like, I feel like it makes me want to talk to the person, like, understand more about what pulled them into it. But I could see, if you spend, like, an hour, like, vibe coding a book, there's something much less compelling about that.

And maybe it's a good book. But also, presumably, the other part of this to me is if you spent an hour making a book, writing a book and making a book, somebody else will spend two years. And if you're picking up a book to read, which one are you going to read? Like, they have the same tools you do. If I'm going to build software in two days, you're going to build the same thing in two years. Probably a lot better. Like, it'd be weird if it wasn't. And if we get to the point where it's not, we get to the point where, like, hey, I can make something in two days that I can't make a better version of in two years, I think, like, then God knows what... Like, that's a weird world. It's, like, hard to see around that corner.

The joy of agentic building

But if it's, like, the stuff that you've built, if you spend two years on it, it's going to get way better than any version that somebody built a similar thing in two days. Yeah, yeah. I mean, I think now, like, if you use... If you do hardcore agentic engineering and build, like, spend a year working on one project now, certainly it's going to be way better than somebody who just, you know, tries to vibe code a replacement in a weekend. Or, you know, the same year that you spent pre-AI, obviously, like, you would make a great deal more progress. You know, it might have taken you five years or 10 years to do the same amount of work in the past.

But I think, like, I think what's interesting right now is this feeling of, like, you know, in the past, you had to be very deliberative about where, as a software engineer, where you would choose to spend your time. Like, what projects do you build? And so for me, as somebody who, like, I really like building software. I, you know, up till now, like, I really enjoyed writing code. But now I'm, like, actively enjoying not writing code. And, which is weird. But at the same time, like, you know, I realized at some point that I had a whole mental backlog of projects that I thought of building over the last 20 years that I just... I would always have the same conversation in my head. Wouldn't it be nice if I had the time to do that? But I can't because my time is valuable. I need to get paid for my time. I need to pay, you know, my rent.

You know, I've got a, you know, I've got a plan for the future. I want to be able to retire. And so I would choose to spend time on software projects that serve my professional goals. And that we're working towards, like, building something that I could, you know, release out into the open source ecosystem. Something that would have value and impact for other people. And so now all of a sudden, you know, I've got that mental backlog of all these, like, orphaned project ideas that I just never had the time or the, you know, inclination to work on. Because I have a life outside of work. And like, I do like, you know, being able to close the laptop and, you know, cook or, you know, you know, go on a vacation and things like that.

Do you like being able to close the laptop now? No, and that's, it's becoming a bit of a problem, you know, because now when I close the laptop, I feel guilty. I'm like, the computer, the computer could be doing things right now. And so that definitely, that definitely is happening. But it is like an exhilarating feeling, the ability that now, like, I'm beginning to revisit all of those old ideas and say, especially because like, I can have like a terminal, like a terminal tab that's like, just working on that side project. I thought of that eight years ago. And now like, I can build the thing that I always wanted. And it doesn't, it does require me to like, nudge it along a little bit every day and say, that's not right. And like, oh, I've got a few more, every day I have a few more ideas and I nudge along the agent and I keep grinding on it. And, you know, there's a few things that, you know, I'll continue to like release things. You know, some things are only for me, some things all open source, but it's nice. Like sitting there watching, you know, watching Netflix at home and just nudging along my little agents working on my, you know, personal side projects, like clearing out the backlog of stuff that I, you know, things that I dreamed about and just never, you know, never would have gotten built without AI. And that's, I'm having fun doing that.

That makes me think of a kind of interesting thing, which is how long does that feeling last? So, or how far can it go? So say, you saw this happen with like ChatGPT when it first came out, where people would ask it these sort of dumb questions of like, rewrite this email in the style of whatever, you know, like, and it had this sort of like, this was like this fun little gimmick that it can do. And at some point it was like, okay, I know what it's going to do. It's going to produce the thing. It's not, the artifact isn't that fun. It was sort of like the process of doing it that was fun and the artifact itself. I'm like, whatever. And to your point, if you have all these things, like churning away in your backlog, your project backlog, your idea backlog, basically. It'll be cleared out eventually. But it's partly that, but it's also like, if you just said, let me give you a two paragraph spec of the thing I want. You push one button and it's done. Do you think that's fun?

Like, is it fun if it can really do the whole thing autonomously where like RoboRev was done in one, and if you can one shot RoboRev, would you have fun building it? I don't think so. Yeah. I think that, I think that the part of the fun, like part of the fun is the process and like seeing the, like seeing the thing, seeing the thing take shape, you know, kind of that feeling of like, you know, I'm not an artist, but you know, I imagine like what it might've seemed like to, you know, Leonardo da Vinci to like be chiseling, you know, chiseling David, like out of the block of, out of the block of marble. And, and a little bit like the, the agentic, agentic coding right now, you know, maybe that's a little bit of a grandiose analogy, but, but it does, it does feel like that a little bit where you can, you know, your feedback, like the feedback that I'm giving the agent, I can see the project, I can see the project taking shape.

Like, you know, you, you can see that it's like, you know, it's the agent, agent's behavior out of the box. It gets most things wrong. And so you have to go through and like, you know, find all the things that I got wrong and get them and get them right. And then refine and refine and refine. And so it takes, you know, hundreds or thousands of iterations with the agent, you know, to make a thing that, that, you know, essentially meets, you know, meets my standards, which, you know, and I, and I have high standards for, for tools. And so, you know, maybe for other people it'd be different and like, maybe their one-shot, their one-shot tool, you know, would be, their, their one-shot tool would be like totally fine for them. You know, if it's like a piece of throwaway code, like it serves a one-time, a one-time use or like limited time use. But if it's a tool that you're going to be using for years, like I, the way I'm approaching it is like, I'm building things that I want to be using five years from now. And so I might as well, you know, put in the effort to do it the, you know, do it the right way and to make it, make it good. You know, but I don't know, it's fun. Like, I think the feeling won't last forever. And maybe a year from now, six months from now, it'll just feel like work again. And, you know, at six o'clock or seven o'clock, I'll be happy to close the laptop and say, you know, the day is done. Enough is enough. You know, it will, the agents will be there tomorrow morning.

It's, it's such an interesting point to like, the question of like, do you like closing your laptop? Like, are you having fun? Would you have fun if you could like snap it into existence? It feels like questions of like, what do engineers enjoy? What is like a person who might contract an engineer to like build something they want, enjoy? Like, who enjoys kind of like being in the process and maybe seeing the process sped up or automated? And who kind of just wants it done?

Like, I guess like tools like RoboRev that you're talking about, that's very like. I guess I should explain what that is. Or like a tool for reviewing commits, like RoboRev that kind of augments like the development process with AI. It's funny because it is kind of like an engineer's tool. Like it's something they enjoy, I think. Having something like speed up the process. But there's this whole other group of people who probably do, I guess, want to snap things into existence.

But even something like that, I mean, you built the thing, you can tell us. But even if you could snap your fingers and have the problem that that tool solves solved, I'm not sure that gives you the same satisfaction either. Like, there is something about building a thing and using the thing, and like — And it's like, it's my thing. Yeah, it's like, look at it work, and I see the spinning wheels, and I see it come back with stuff, and I — it's like, there is an enjoyment in using it at all. And I think that, like, we're not — it's not all just, like, how do I maximize some utility of this thing does my particular job. Some of it, maybe, it's like, okay, you have a button that fills out your tax forms. Okay, that may be a different thing. Joy achieved. There's a very different, to me, set of things. It's like, I just do not want to ever see this again, versus I built this partly because it's, like, a thing I like to use. A lot of people, like, do woodworking. They're not like, snap my fingers and have the thing exist. It's like, I want to do the thing. And I think there's still some of that here.

Building RoboRev

Yeah, I mean, yeah, I think, like, the — you know, just, yeah, for people listening, like, what are they talking about? Like, the thing I started building — still started building recently as part of some of the, like, agentic development side projects. Like, I found myself — And just — I think people might have seen, too, because we've talked about it on a couple episodes. Well, yeah, we'll — I can't even remember. I mean, the software is so new, like, it's hard to even know. But, you know, I essentially, like many people, I started running — I started running multiple Claude Code sessions, and then I found that most of the work that Claude Code was doing was full of bugs. And so I started running separate, additional, either Claude Code or Codex from OpenAI sessions, whose entire purpose was to review the work of the other Claude Code sessions. And so I would have three Claude Code sessions and then three reviewers, and all they were doing was review this commit, review this commit, review this commit. And so, like, Claude Code would return control to me, would commit its work, and then I would say, here's the commit hash, and I would give that to the reviewer, the reviewer session.

And then, you know, this went on for a period of time, and I was like, I'm losing time because I have to manually, like, take the commit hash and, like, give it to the reviewer, and that's time that's wasted. Basically, the whole thing just didn't seem very efficient. And so I said to myself, well, you know, this whole process can be automated. And so, yeah, so the project RoboRev that we mentioned is basically like, you know, it automatically triggers a code review whenever the agents return control to you, like, whenever they commit and return, get commit and return control. And I found that to be really liberating. Like, I didn't have to run those extra, you know, Codex or Claude Code sessions anymore.

And if I have parallel, like, multiple Claude Code sessions, whenever I return to the session from another side project or, you know, another, you know, get work tree or whatever in the same project, I have not only the agent has finished what it was doing, but then also the associated code review that was done asynchronously while I was working on something else. And so if I, you know, forget what I was doing, or if I'm not sure what to do next, like, it's just easy to say, hey, you know, Claude fix these bugs that were discovered in the work that you just did.

But you're right. Like, I think part of the fun of it was, you know, making a tool, seeing, you know, building the tool yourself, seeing the tool work. And I think, you know, maybe a side, you know, corollary of that now is that a lot of people are going to prefer to use tools of their own making. And I find that whenever I show new AI tools, or like new tools that I've created to other people, usually the response is something like, I'm exhausted with how many tools I'm being shown and like how many new tools I'm seeing on the internet. And like, I have no, you know, the only tool I can manage right now is like a single agent session, or maybe like a couple of agent sessions. And any other tools beyond that is just, is almost exhausting.

And I've also felt like that exhaustion of like the head spinning feeling of like 10 new AI tools and systems and solutions coming out seemingly, you know, seemingly every day that, you know, influencers are really excited about on social media and on LinkedIn and on YouTube videos and things like that. And so, you know, for the longest time, like I was, I was tuning out a lot of what was going on in LLM world. And the only thing that really pulled me in were these CLI coding agents. I think prior to that, I was like, you know, not, not impressed, to be honest.

Yeah. But I think that's like the, that sort of reaction, being shown 10 new tools and you have things like that. We've all sort of played around with this stuff is, that to me is like the content thing where it'd be like, if iPhones and like iPhone video editing had just come out, it's like, look at my video. And it'd be like, oh my God, this, yes, it's very good. It's as good as any video I've seen 10 years ago. But like, I can't look at this much video. Like I'm, I got to tune it out. And I think that, I don't know, there's this kind of feverish pace to the whole thing where all these people make things that seem really good, but it's like, I can't, I can't keep up.

Boy bands as a model for collaboration

Yeah, Ben, I know we've, we've been down the AI rabbit hole for quite a while. One, okay, maybe one thing I'm very curious about is when I, before doing this, when I asked you what you were most excited about, you said joining a boy band. And I wonder if you could maybe tell us more about that, your boy band dreams. I mean, that does seem like a fun gig. Okay, basically. It's like a model of collaboration.

So, okay, so you like- It's your Gastown, isn't it? I, you know, Gastown's probably not a bad boy band name, to be entirely honest. That you can, if you like ever interview an exec and being like, especially an exec from Salesforce, this is apparently famously a thing that Salesforce did. You ask them like, what are they good at and stuff like that? They would say stuff like, they're just, they're incredibly collaborative, just the most collaborative. And it's just like, I'm just so good at working with people, great. And I feel like one of the things we all say like, oh, I'm pretty, I'm collaborative. I'm like, you know, whatever. I'm like not, I don't really want to work with that many people. Like, I don't want to, it's like a lot. And you want to work with some people, but like not an army of people. I don't want to work with the whole Gastown army.

And I think it's more fun to work with like a small group of people who's like all committed to a project and like wants to do the thing. And it's basically a model of collaboration, like a boy band, where it's like, there's a handful of us. We all have our fun little personalities. We do our own things. You could do your own thing if you wanted to, but you kind of like being a part of this thing. Maybe sometimes you kind of do a little bit of your own thing, but like ultimately we all have to have to do the same song and dance. But you can maintain some amount of like independence in what you are. And it's not like each person in the boy band is overseeing enormous departments. Their job isn't managing a ton of people. The job is like, no, I make the thing.

Um, historically to me, there's always been like one of the downsides of startups is in some ways you get punished if you're successful because you have to go hire a bunch of people. And like you become the, oh, I have to manage tons of people and the things that go along with that. And sometimes there's things about that that are fun and you like, but like a lot of it is just like administrative. Administrative is to sort of downplays it too much, but it's just like organizational work. It's like, how do you make sure everything is aligned? And how do you make sure this department doesn't send to this department? What happens when they disagree? And how do you make sure like stuff set up such that these people can communicate with these people? And how do you keep everything on the rails? And I don't know, I don't particularly enjoy that work. Um, but that's been, that's what you have to do. Like you build a company, like the goal is to get to the point where you have to do that work. Otherwise, what have you done?

Uh, and I think like that may not be the case now that, that you can do a lot more with fewer people. You can keep it relatively small. You can have it be like, no, this is just a handful of people who are working on an exciting thing. Uh, and you can make maybe not an enormous business, maybe not, you know, the next trillion dollar thing, but you can make something that is a functional, profitable, successful business. And all sort of reasonable definitions of those words with a relatively small group of people. At least that is my hypothesis.

And it's basically a model of collaboration, like a boy band, where it's like, there's a handful of us. We all have our fun little personalities. We do our own things. You could do your own thing if you wanted to, but you kind of like being a part of this thing.

Yeah. And not, not to stretch the like analogy too far, but do you know, and you're like dream data work or, or like just building boy band scenario. Like what role are you? Are you the like bad boy? Are you the father figure? Like which, is there a father figure of boy band? I, well, not, I mean, maybe not in the modern boy band. I feel like in the OG boy band, there was always kind of like a responsible party, responsible, the one like driving them home after. I think I would be like more of the moody guy in the back. Like you don't know that much about. Yeah. It's like, what's his backstory? It turns out it's like nothing. There's nothing interesting there, but you got like. You're doing like a lot of skulking around. Yeah. And every once in a while, like you emerge and you have your, your like verse and stuff. And people are like, oh my God, he never told me that. I would be more than that. Yeah. I'm not trying to be the guy in the front. No, that's fair.

It is. I mean, it is interesting too that, yeah. Like people aren't, when a boy band succeeds, it's not like people are say like, we need to add 200 people to this stage. Then there are some K-pop bands. Yeah. Right. But yeah, but right. Like it's not a. It's not like it's you got to scale the whole thing up. And I don't know. I think that seems we're fun. Yes. Are you I mean, are you forming any boy bands right now? Like what's what do you got cooking? It's just kind of like a medical. I think he's maybe like a metaphorical boy band. Yeah.

There's some ideas the metaphorical boy bands. I you know if I could like snap my fingers. I know kung-fu like matrixy one skill. It would definitely be like boy band skills. Yeah, like being a being a pop star seems like a really good time. I mean, that's true. Yeah, like if if you get the chance, I think you should take it like 100%. Yeah. Oh, I would but I don't hold your breath or I was thinking maybe other projects. Are you working on projects right now and have you succeeded in kind of like keeping the essence of a boy band in them? Yeah, cuz it's just me. It's not really a boy band.

Yeah, I mean I like playing around some stuff I that it's similar to Wes except I would say with with less success. You have a bunch of ideas you want to play with them, there are things that you think are like interesting to experiment with a lot of them are less like here is a very precise problem I have but more of like this is a rough way of working or doing a thing that I feel like I could there feel like there's got to be a way to sort of solve this problem.

To your point earlier about like there's a creative process you'd like so I periodically yell at the Internet. There are parts of that process I really like there are parts of that are kind of a grind. I'm not sure how much of that you can like AI away and like there's a there's a delicacy there too of like I don't want to like that is not push button get blog post is not a thing that I want to do. However, there are a lot of parts of that drafting process that you're just like this is this is painful. And it's like I don't know there are things you can do to help help unstick you with those things or stuff like that where it's like the things that I do I want to try to like find ways around. Yeah. Just so we'll see so it's stuff like that.

Just see it a little more concrete like see what you like doing stuff like like Wes are you saying like using like Claude code or when writing and maybe like getting like assistance like what kind of stuff are you cooking on so I I basically use AI to write stuff in one way which is as a thesaurus. It's a pretty good thesaurus because you can be like here's the word but make it more X like you can kind of guide it a little bit. And so like obviously that is not you're not getting a whole lot done there like it's it's useful sometimes. But like that's basically the only way that I found it being useful the thing it writes do not want to ever say a single word that it ever writes like certainly you try periodically but go try this and it's like oh my god, I cannot.

But there are there are processes and right there's like points in the process of writing these posts when there is something akin to a thesaurus. That is like a broader set of problems. You have a bunch of like basically loose ideas in a doc and you're kind of like I don't really see how to weave these things together. I don't really know what the narrative here is. I kind of like this sentence that I wrote here and I like this idea and I think it needs to work here but like my god, it's a mess. I mean, it's like that is that is the stressful point of writing these things because you're like you're not just kind of refining paragraphs. You're having to be like does this fit together at all? Like do I have puzzle pieces that can make a thing that look like anything? And there's that I think that there are ways you can attempt to to mimic sort of the creative process that helps solve that for me.

Write that paragraph because it does a terrible job of that and like people have written a lot of stuff about this I cannot beat certain patterns out of it

But I do think it can help be like here are some ideas for ways to arrange things most of them aren't any good, but there is like Some inspiration that you could be like, oh, that's kind of an interesting thought like let's play with that some That is basically the equivalent of having a conversation with someone about it Like like a lot of these posts ends up getting written because you have these loose ideas and you sit down you talk to somebody And you're like, oh I didn't think about connecting in that they not they don't tell you to do it it's just like they say something that triggers a thing in your head that like It connects two ideas in ways you haven't seen and so it's like can you can you use it in that way?

The the analogy I have a friend who's used with this is like if you play the game Code names. so it's like basically you get random words that you have to find associations between and it's that process where interesting things come up because It's not like here's a word find three related words that are very similar to it. It's like here are two very disjoint ideas Forcing yourself to be like how might these be connected? There's something kind of interesting that comes out of that and so I think it's like that is a process that right now you has to sort of happens a repetitious or serendipitously because you have to walk around the world and do that and I think there are ways you could like again use it as a Conversation partner of sorts that isn't just like ChatGPT. I think it needs to be more guided than that But it's like basically can you can you use it to to? Find those like serendipitous stuff faster.

Hearing you talk about it sounds like you're kind of like browsing the space of like Possible connections just to see kind of like what's out there. And is there like Anything useful, but it's pretty broad like storyboarding and like yeah So like this is a bit of a weird example, but but there was a like there was a thing I wrote a while back. I don't know a month or so ago that was about It was about like some degree like the addictiveness of AI stuff It wasn't just like oh my god, like ChatGPT psychosis or whatever, but it was kind of about How this feels like a certain kind of like we are creating our own realities in Certain ways and stuff like that and you see that kind of extending in other places I don't remember what the exact sort of like news hook was on that, but there was something like that

a couple eyes before I went to a Lord concert and There's a couple songs that like you're sitting there was in a Lord concert They're like, oh that's kind of an interesting thing to it. You're like watching all of these like Teenagers totally dissociate from the concert so they can take pictures for like Instagram and not pay attention to the concert and Then Lord has a song or an album called pure heroin where heroin spelled like here like female hero, but it's also like AI is kind of hair pure heroin that way like that connection Seeing that actually ends up driving a lot of the narrative of the like you can't It's not like I have this exact article. I write. Oh, I can tack on this analogy It's like I have all these loose ideas and then you see that and you're like, oh, there's actually an interesting thread this this is the thread that can tie it together, but you need to like be sitting at that concert to see the thing happen to do it and so It requires kind of these like random collisions that then one of the random collisions is like, oh that's where the interesting connections are and I think that for me anyway is where a lot of like Attempted creativity comes from and so it's it's can you can you Speed up that can you like force that process a little bit more and like I think like again code names does exact doing exactly that you become very very creative playing that game because what else are you gonna do and I think like There may be ways to be like help me run more threads through this to see what happens It's like it's not gonna write anything, but it will help you see More interesting things potential.

It requires kind of these like random collisions that then one of the random collisions is like, oh that's where the interesting connections are and I think that for me anyway is where a lot of like Attempted creativity comes from.

Yeah, it's really interesting that it's like it sounds like like it's not a rhyming dictionary. It's not a thesaurus It's like a little more Granular than that, but it's still not Writing the thing, you know, it's just a different way of like exploring the Associations or like paths.

The writing process: cheating and analogies

part of it to me is because so people have asked about these block tests before and a lot of them do have these like There's why is Lord the sort of hook for all this thing? yeah, and it's like how did you how did you come up with that and The answer is you cheat. The answer is you didn't come up with that You came up with one association with Lord and then you read some Lord lyrics and you found four others that felt kind of attached And then you hook other ideas on to them And so it looks like oh my god, you you had all these ideas and you found like Lord is the thing that perfectly Weave through it. It's like no I didn't do that I started with a Lord song and then like figure out ways to sort of tack on things that feel associated and it creates like A much more tighter sense of connection than you kind of should Because it seems like you found this analogy that map so well when in reality you've just like Stuck ideas on to the analogy to make the analogy where the analogy was a starting thing and the ideas were glued on to that and so it's that kind of thing or like The to me the process for writing this stuff is not What are the ideas? Stick on the Lord analogy. It's like the Lord analogy is actually the thing that drives it that forcing you to describe Everything through this analogy I think forces you to like Come up with more creative ways to say it Whereas if you're not doing that you end up just writing kind of a boring outline and then like sticking on sort of weird forced like half Connected Jokes

Literary influences

It's so interesting to hear the process and finding the associations. I'm so curious about, do you have literary influences? Are there people you love to read that have influenced your style, or how did you kind of secure your inspiration? The two things, this is, I mean, Matt Levine's stuff is great. And I think Matt Levine, she has a style that is, one, you— It's very much a thing that you can't emulate because, like, one, everybody knows, like, I'm going to try to write — like, everybody sort of knows that, and you can't, like, do that anyway. But it's also, like, very hard, and he's very good at it, and I think that's what's up. But the thing I really do appreciate about him is, his thing is, like, kind of about nothing. Like, there is a — what's the point of this? I don't know. Like, there's a bunch of interesting things in the world to look at, and I like to look at them. And so I do think that — I very much read his stuff because I find it enjoyable, um, that — no shade to Ben Thompson. Ben Thompson obviously is very popular.

To me, Ben Thompson — Ben Thompson writes about a subject that I should care a ton about, and I find his posts hard to get through. Like, this is a thing that I read because I think he has, like, very good ideas, but it's sort of like, this is fascinating, the structure of it is, and the ideas he has are, but, like, I don't particularly enjoy reading it. Matt Levine makes me care about stuff I should not care about. Like, I'm just interested in it because it's, like, he tells a good story with it. I don't care about any of the topics at all. It's like, yeah, the private credit bubble. How much do I care about the private — I don't know, not really. But, like, when he tells a fun story, it's fun. And I appreciate that, that style, I think.

The other thing, and actually this is where a lot of this originally came from, was there was a guy — I don't remember his name. He is in chemistry. He's a chemist. Uh, he won a Nobel Prize in chemistry. He made buckyballs. There's some, like — Mr. Fuller, maybe? There's some sort of, like, compound that is, like, this, like, giant hexagonal thing. He won a Nobel Prize for it, and it was like, this is amazing, and it turns out to be, like, completely useless. It was just, like, a very stable ball. I don't know. Clearly did not stick. But when I was in college, I, like, went to some thing that he gave this talk. It was, like, a lunch thing that he gave a talk at. It was, like, whatever. It's kind of the background. But he gave the talk in this style where it was probably a 20-minute talk, and he had, like, 300 slides. And it's — you just, like, can't look away because it's just, like, slide, slide, slide, slide, slide. And part of it — each slide is, like, there's this sub-narrative to it where he's giving a talk that basically you don't really need the slides for exactly. They're not, like, and now we've got a slide six, and here's the bullets, and here's the diagram. But it's, like, he will say something, and it'll be, like, an image that sort of is related to it. And it ends up creating kind of, like, these sort of two stories, where there's a story behind it that's, like, the entertainment and, like, the thing he's saying, but you kind of, like, it keeps you really attached. And I started giving talk — like, I really like that style for talks, and so that's, like, kind of the style I gave for talks. And the thing that creates is, like, there's a certain pace to it that I, like — it requires you to kind of weave a bunch of stuff into it, because you can't just be, like, AI, picture of AI. You can't — computer picture. You have to come up with, like, little vignettes that follow along throughout the course of it.

I think this is much more entertaining. And so, like, a lot of the blog style actually was more of, like, how does that look on paper? I'm not sure it delivers on that, but, like, that's where there's, like, footnotes and stuff that are sort of, like, the whole thing is a little bit of, like, there's a sub-narrative to it that's, like, me entertaining myself. Because, again, it's about sequel, and, like, my God. Yeah, I mean, it is interesting. It's, like, all the elements of, like, a lot of, like, stream of thought and collage and association, but it is crazy that it all, like, comes back to data. I feel like it's, like, really — not to butter you up too hard, but it's interesting to read. It, to some degree, comes back. Like, I'm not going to — you know, sometimes they don't really tie together. Are you saying there's no promises? You might, like, at some point, drift off data and go full? Well, yeah. Like, broadly, yeah. And I would say I, like, at this point, mostly have. Like, there are still things there, but, like, I don't — it's not a lot about ETL tools these days.

Small companies, sustainable software, and the VC game

I mean, on the subject of, like, you know, like, small — like, small teams and small companies, like, I feel like examples of companies that have remained small and yet been really successful, like, there just aren't that many. I think, like, Craigslist is always the canonical example of a company that, you know, still has the same, you know, very bland, you know, utilitarian website that it had, you know, 20 or 30 years ago. And yet they have tons of revenue and, like, some really — some really small team. And, you know, me as a software developer, like, I always found myself torn between, like, the desire to, like, build open-source software and, like, make the world a better place through, like, building tools and giving them away for free on the internet, with the downside that it's hard to do that in a sustainable way, like, to get people — very hard to get people to pay you to do that for a long period of time. It's hard to be able to hire people because people need salaries and they need health insurance and all those things. And on the other end, you know, the other — you know, and you have entrepreneurship and starting, you know, starting companies and then invariably, you know, with the way the world works, like, you get nudged into the, like, you know, the venture capital and the raising money. And that has a lot of strings attached. There's a lot of expectations that, you know, everyone has in terms of, like, how fast you're going to grow and how many people you're going to hire. And then, you know, that gets really complicated, as you know.

And so I think one optimistic, you know, hope that I have is that, you know, there might be some, like, new model of, you know, people that build software in the world and can do — you know, build new things or, like, create things that have impact that's like neither of those things and yet, like, can be fun and sustainable, like, kind of the best of both worlds. And I don't know what that looks like, but something I'm really interested in. I mean, like, my answer to that is it's a — it looks like a business. It's a small business. Like, it's only in software where — and this made sense, again, kind of the economics of, like, building software. There's a ton of upfront cost and stuff. But only in software, like, we're going to start a thing that's going to be 10 people and we're going to try to make more money than we spent. That's the goal. And, like, you start a restaurant, duh. Like, yes, that's the goal. But in software, it's like, that's a ridiculous goal. Like, how are you going to grow? How are you going to — It's kind of famously parodied in Silicon Valley. It's like revenue. No, no, you want to be pre-revenue. Yeah, exactly. It's like, there's very much a model for, like, a handful of people working on a fun thing forever that you can be sustainable. It's like, that's a business.

But start — again, I think it is — it's not just, like, the VC ecosystem. It's like, the economics of software have not been able to support that. You do have to spend $5 million before you can sell anything. You're in a lot of debt. And, like, that doesn't quite work. And so — but now you don't have to be. Like, you can get to the point where you can probably make — you can be probably making money after a year of one person working on a thing. And, like, all right, that's — yeah, that's just a business. And I don't know, it seems like if building stuff is fun and a business you can run forever isn't bad, then, like, not a bad gig. Maybe. No, I don't know. Maybe it doesn't work, and maybe software is too, like, winner-take-all, and so you have to get super huge. But I don't know. I'm, like, not that cynical about that, I think. I think it's like, how many people do you really need to really love a thing to make it sustainable for a handful of people? Not that many.

Yeah, it does seem like there's a big space for, like, a small useful thing that, like, people are willing to pay monthly for or something. Yeah, and it does — this isn't quite right, because there's certainly a lot of things that are just, like, pure productivity tools, but it starts to feel a little bit more art-oriented or, like, content-oriented. It's almost like Substack, but, like, Substack for, you know, somebody's, like, vibe-coded software. Yeah, like, there was — one of the things I've noticed is people seem to have replaced — in a lot of ways, decks have gotten replaced with, like, Vercel apps, because you can kind of do it in the same amount of time. You're like, oh, you want to give a demo of something? I just made a thing. Like, you get random pitches from people about, like, check out this thing or whatever. I have gotten a lot of those that are links to Vercel apps now, because it's about — it used to be, like, here's a PDF of some slides or whatever, and now it's like, why would I do that? I can make a little clicky app thing. And so, I don't know, that feels more like it's just — software is just content at that point, as opposed to, like, a tool for anything. It's just like, I just have a much more flexible canvas to make a thing I want to make.

It's interesting that you have a deck being replaced with a Vercel app, which is like, like, deck plus plus, just a little bit more, like, interactivity, that the bar is just way lower. Yeah, and it's not that different than — what's the simplest version for me to explain something? It's a bunch of text. Okay, what's the slightly more complicated version? A Word doc with text, with, like, some bullets and some nice formatting. What's, like, a very flexible version? Like, a deck, because it's just a — it's a bunch of static canvases that I can literally draw anything I want on. What's an even more complicated version? A deck that I can click on. That's kind of a website? Like, okay, it's not that different.

Honest feedback and the Simon Cowell VC

I mean, so people periodically reach out because of the blog stuff. You'll complain about something, and I'll be like, well, actually, here's my version. They're like, would love your take on — Yeah. Which, I mean, as a slight aside, one thing that does sort of — is a bit of a character of that is because I complain about a lot of stuff, people mostly will send things and be like, Give me your worst opinion on this. And so it's kind of asking for a lot of like, what do you hate about it? Which some things are very good. Yeah, how do you like, how's that job? Do you like doing the job of, uh, you hate this? Like, what do you hate about this? Job is a real stretch.

I don't, I mean, I don't, I'm not, like, I hate everything that you do. That said, I do think there is a relatively useful hole in, like, VCs in that, where if you go pitch Sequoia and they don't like it, they will tell you, like, it's too early. Nobody was pounding on the table. We really like it. We'd love to see it next round. Like, we're just not there. We just couldn't, we just couldn't quite get there. Market wasn't quite big enough. We had questions about competitors. Like, they will tell you nothing. And the reason they tell you nothing is because they have no incentive to tell you anything other than that. Like, why would they tell you something that's gonna, why would they tell you what they really think, which is probably, like, I hated the way that founder dressed. Like, they reject you, one, I mean, Sequoia, not just Sequoia, like every VC, you get rejected probably for emotional reasons. You get probably rejected because, like, literally two minutes into the pitch, they don't take you seriously and they're tuned out and they're off looking at something else. Maybe they reject you for, like, real material reasons and, like, they really evaluated it and didn't like it. But, like, they're not gonna give you that memo of the thing they said. And it's, like, they want you to like them so you come back next time. If you do are successful, they want to, like, maintain optionality of that deal. And it's very frustrating when you're pitching because you just don't get anything honest. You get a lot of, like, kind of soft, ah, it's just not for us yet. We'd like to see a little bit more, a little more meat on the bone. And you're, like, okay.

But I think, so I think, like, there is a little bit of a, this is a, like, I am not a VC, nor do I give feedback that people should listen to. But I would appreciate a VC where you walk into it and it's a little bit of, like, the Simon Cowell of VCs where you're, like, they're gonna hate it. I'm not gonna be offended when they hate it because I know they hate it. And if they like it, Simon Cowell liked my song. Like, oh my God. I would pitch that VC because I, like, the worst that happens is the thing I expect to happen. And there's a million others. I'll just go pitch them for the actual money.

But I would appreciate a VC where you walk into it and it's a little bit of, like, the Simon Cowell of VCs where you're, like, they're gonna hate it. I'm not gonna be offended when they hate it because I know they hate it. And if they like it, Simon Cowell liked my song.

The VCs have, like, sort of weird, those are, like, value systems like this. But, like, there's some weird incentives there. And so I do think, like, yeah, what you really think is nice to hear in an uncomfortable way sometimes. But again, if Simon Cowell tells you you're bad, you're not gonna be, like, I must suck. You're gonna be, like, that's his job. That's what he does. That's his shtick.

No, I largely agree with what you're saying. I mean, it's, especially in Silicon Valley, it's rare to hear, like, honest feedback. Like, and I've known, you know, I've known VCs and have been in pitches where I've received, like, honest feedback. Like, this is not good. You shouldn't be doing this. You should be doing something else. Like, the way that you're approaching this problem is wrong. And it is refreshing to get those, like, to get those takes. But, you know, but usually, like, I think professional, you know, venture investors are looking at so many pitches and talking to so many founders that generally, like, they're just pattern matching. Like, they don't have, you know, the time or energy to, especially at the early stage, like, to think super hard about things. And so it's basically like, you know, have I seen any, you know, similar companies in the past that have done this successfully? Like, does this, the way that this product is being sold, like, does that make sense? Like, are they selling to the right people? Like, does the tech, does the team seem credible? Surprisingly often, like, you look at the backgrounds of the, you know, the founders of a company and, you know, it's like, they can't build this. Like, are they, you know, if there's going to be a successful product, it shouldn't be, it shouldn't be this team. And that's often, you know, in many cases, that's just the right answer. And it isn't always the most, unfortunately, very frustratingly, it's not the most competent and qualified founders that, you know, succeed in the space. And I know that, you know, often, you know, it's the founders who were like, had the right background, were at the right place at the right time, and yet still, you know, came in second or came in, you know, fifth, who ended up, you know, very frustrated and bitter. But so that's just kind of how the cookie crumbles, unfortunately.

But yeah, it's a lot of, you know, it's a lot of pattern matching, because ultimately, yeah, like, in investing, you want to maintain relationships, not burn bridges. You don't want somebody to come away from interaction being offended or being like, you know, that guy was a jerk. You know, he was honest, but he was a jerk, you know? And you should never be a jerk. Like, I'm never a jerk. So yeah. And to me, like, yeah, I think a lot of times VCs aren't, don't give totally truthful feedback, because they're trying to be nice. And a lot of times the answer is like, you're not the horse I want to bet on. And like, I don't know, what use is telling somebody that, you know, like, so yeah. But as a result of that, I think you also just get like, kind of very sort of soft answers.

Existential dread and just doing things

I mean, it's, especially in Silicon Valley, it's rare to hear, like, honest feedback. Yeah, I mean, I'm certainly not trying to like, maintain some sort of set of deal flow around that sort of stuff. But also, I don't know, that's like, kind of what I would want. And if you're nice about it, it's fine. And maybe, I mean, I don't know, I try to be. Yeah, yeah. And I was like, I don't know, for me, that's like, that stuff's interesting. It's interesting to see what other people are working on and stuff, so.

Anything you're excited for this year? Or what are you most excited for this year, would you say? I mean, it's, I don't know, there's a bunch of stuff. The way all this stuff changes things is like, exciting and a very, I don't know, do I, maybe we could just have like, a down year. Maybe we could just, just stop for a bit. Like. It's like, all gas, no brakes. We need like, a moment to catch up. Can we just all, so I don't know. Like, I am sure there are things that will come out that will be, you like and are fun and stuff like that. And I think, I think Claude Code is a good example to me of what that has been, where you use it and you're like, this is pretty fun. I can do a lot of stuff. This is cool. And yet, there's also half of you that's like, I am so behind. This is like, what do I need to be doing stuff all the time? And like, am I keeping up? Or like, everybody else is doing the same thing. Or like, all of the, all of the ideas I had, are they all like, going to be built a hundred times over? Like, when things are moving that fast, it is both exciting, but also sort of terrifying. And so I think like, there will inevitably be more moments like that to me, where something happens and it's just like, there is an opportunity, but there is also, I don't know, so many opportunities that you will miss because stuff is moving so fast. So it's exciting, but like, somewhat existentially stressful.

I mean, existential has come up a couple of times. Like, as people mentioned, like feelings of uncertainty, which is so interesting. To think about, especially in like your response of like, it's a really exciting technology. And there is this like, maybe like slight existential. I mean, I spent a lot of the last, I spent honestly, very honestly, I spent a lot of 2025 in a state of like, essentially existential dread about like, what it means to be a software engineer, you know, after. And, you know, like Ben and I are like, we're a little different. And like, I'm more of like, you know, like writing C++ code, writing Python code. Like, you know. Wes is an engineer, I'm not. Yeah, Ben's not an engineer.

But, you know, but I feel like the, you know, the rapid pace in AI development was for me, like disruptive to my core identity as a person who has spent a lot of time becoming good at software engineering and getting really good at writing code. And so this feeling of like, you know, taking like essentially yanking the keyboard, you know, the proverbial keyboard out of my hands. And it's like, you don't need to write code anymore. And so there's this feeling of like, well, you know, what am I for? Like, what's, you know, what's my purpose here? Like, where can I add value? And I think that at some point, I don't know. You know, at some point it clicked and I regained my agency and started having the, you know, as Peter Steinberger puts it, you know, I can just do things, you know. So I'm a little bit more embracing this mentality of like, just doing things again. And not worrying too much about whether, you know, other people are doing more or other people are, you know, working smarter. And just focusing on like, you know, am I building things that are useful, things that people care about? Maybe they're only things that I care about, and maybe that's okay.

But it's definitely like that existential dread, I think, has been, you know, present in a lot of the industry. Like, I don't know if there was any one thing that caused me to break out of it. But, you know, I remember there were like weeks where I was just like, you know, I'd sit down and like start writing code in Emacs or in, you know, VS Code. And I'm like, I feel, you know, I feel so ineffective, you know. And so I've, you know, I've gotten over that, I guess. And maybe that's just like a morning, just this morning process for like, the old world that feels like it's now gone. But I don't know if you felt that way at all.

The just doing things I think is a useful reminder sometimes where there is another sort of a lot of the stuff I do is again this kind of blog stuff so sort of a lot of things tie back to that to me. The process of writing these things or putting together presentations or anything like that that isn't just sort of like I'm going to methodically plod through something but there is there is some like discovery in a sense to it. You spend a lot of time wasting your time like there is a lot of work that goes into it that you're just like that was bad and you got to throw it away and like that's part of the it has to happen you can't skip that like you can't skip some of that process and so I think like the just doing things it's not to me I could go oh everything is happening so much everybody is going to create the new iPhone app is going to make a zillion dollars I have to go create a new iPhone I can just create an iPhone app and I'll make a zillion dollars it's not really that it's like that doing the thing is the process how you figure out what to do what the actual thing to do is it's that you can't you're trying a bunch of stuff maybe some of them will work maybe they won't but the point is whatever you find that does work that's the way to find something that does work is by doing what you're doing and I kind of view it as that as like are any of these ideas I'm going to play with any good probably not but either in two years I'll look back and be like I may come up with a bunch of terrible ideas but I tried something oh well or you'll look back and you'll be like yeah it took seven failed things to find the eighth that was any good so all you really can do is just do the first one and see what happens.

the one part of that though that I think as relates like startups that is particularly important this is true with like any kind of creative stuff is you have to be willing to move on and I think that's one of the things other things about like you don't have to raise a bunch of money that is helpful is you don't have to necessarily go out and like I've spent six months researching a problem I raised a bunch of people to solve this exact thing we're going to hire a team to people who want to work on this thing now suddenly we have 18 months of time we've spent on a company to solve a problem we've sort of like half pivoted a couple times because this is the idea we have and we built the tech around it and like you can't get away from that you have like this anchor that you've attached yourself to and so in a sense you couldn't just do that you could just do one thing but you couldn't do like a bunch of stuff to figure out what's interesting and so or what might work and now you can kind of can't you can you can have five projects working at once and like maybe you realize one is like a great utility you want to use and one's like actually a thing that people want to buy and one's a thing that people want to buy but you hate it so you don't want to do it maybe you realize eight of them are all terrible and throw away things but in the course of that you found these other things you're interested in like you can do a lot more of that and I think that's like my optimistic view of how you find your way through this is more of that which is like yeah you can just do stuff and then the way you'll figure out what's actually exciting is by doing that.

yeah it's super interesting to hear I mean maybe to tie back to like your way earlier point like in the long term everything's a fad and this idea of more like vibe oriented Feldman I hate to use the word vibe so much but I think it's a really interesting point that like people today if someone it sounds like if someone's like worried about like what do I do now it's like they you're like you it's okay like you can do six or seven things now really easily and explore what's out there never thought of this analogy before but it's like I went to a liberal arts school a lot of people show up at liberal arts schools don't know what they're going to major and the way you figure out is you take classes you try a bunch of stuff and you're like that was really interesting I want to keep doing that it wasn't an engineering school or like a lot of schools we have to declare a major up front and if you go to one of those schools where it's like I got a cheer and like I think European universities might be like this we're like they're basically trade schools like up from like you want to be a nurse you start out of you're like locked in you go from high school to nursing college or I don't know how it works I'll take it yeah uni I don't know what anybody is but the point is like you have to make that choice up front and there you can't you can again you can just do with it you can do whatever you want but you can just do a thing whereas if you go to some liberal arts school you can like do what yeah you can do a lot of things don't worry when you show up don't worry about you want to major and try it and you'll have a lot better sense in a year when you take 10 classes and I think it's kind of like that it's like just try it I don't know you how did how could you possibly know until you try it yeah try a bunch yeah I think it's probably super helpful advice today for people yeah thinking about the future and like the pace of technology and AI and like wondering what to do that they can just try things.

what do you think that you know what do you think it looks like now like does that space like is that still something that interests you to work on like or um yeah it's it's kind of a big question mark like kind of the the fate of like that that whole segment of the the industry like even the whole concept of business intelligence is a little bit you know on you know unclear right like everyone wants to make their own you know everyone wants to make their own custom dashboards now it's like why why should you be forced to use somebody else's you know dashboard builder you know interface um but um it's kind of you know funny to think about like remaking like these these you know fields that used to be well defined with a set of tools that you would use and an established set of practices being you know not totally thrown out the window but you know maybe a little bit maybe on their way to i don't know yeah i mean i don't i don't have no idea uh do i want to work on that stuff again not really there is still a couple things i'm like would be kind of nice if this particular thing existed um my my guess is if i had to say where it goes is traditional bi looks very different and like we kind of get rid of it partly because it never really worked in the first place like people have bi tools they have a bunch of dashboards they're mostly little apps for like salespeople to like look at their list of leads and how stuff is going and they're they're they're not the idea of these things being like i go here to find i drag and drop my way to insight is like not really a thing that happens um and they become very focused on like the presentation of it because i need to looks like all the styles and whatever and i kind of think that does become we figure out a way that that becomes more people get really accustomed to chat body stuff and i don't know that that necessarily means the tip back to the sort of gymnastics point conversation from earlier i don't know that becomes a chat bot that writes equal queries for you though in some ways i think it becomes more of like we just bypass sequel queries all together again if i want to understand what my customers are talking about i could drag and drop my way to doing a bunch of analysis on top of like some structured data set of support tickets or i could just be like tell me the vibe of the support tickets i'll probably just do the other one and never bother with the like analytical stuff first

the other way to kind of talk about this is like we have done data work because it's the only way we've ever been to look at things at scale if you have a million support tickets how do you look at that you can't do it but with math like math works really well at any sort of scale you can add you can average a billion numbers just as easy as you can average 10 how do you average a billion conversations you don't like there's no way to you have researchers read some of them and summarize them manually and they're like we have a thing that can kind of very approximately do math on text and so the point of numbers was to do math on we quantified stuff so we could do math on it but if you do math on stuff that's not quantified do you do that much math like maybe not um so i think that stuff kind of there's a i think there's a real existential thing there

how do you average a billion conversations you don't like there's no way to you have researchers read some of them and summarize them manually and they're like we have a thing that can kind of very approximately do math on text and so the point of numbers was to do math on we quantified stuff so we could do math on it but if you do math on stuff that's not quantified do you do that much math like maybe not

i to the stuff that like posit does does feel like there's a place more for nice tools for people that are continuing to work with like when i open a csv how do i open a csv in a really like useful way there's a bunch of things that sort of like excel or it's pandas or it's posit in a way but like posit doesn't it's like designed for a much richer experience in that there's a bunch of stuff like that that's like there's still people who want to look at data and manipulate it and there still isn't really a tool that does the ones that start in that space which i think like mode tried to be like a nice workbench for that sort of thing get dragged in the bi world and there's a lot of startups that have like started as these like nice data oriented utilities that then end up being a bi tool and so i think there could actually be something that comes out of like this is actually a thing that's just like a really good way to view and manipulate like csvs

yeah essentially i mean even just just in the last you know just in the last two years like i've worked on exactly that inside inside positron which is uh it's a you know posits vs code vs code fork and the use case was like i have csv and parquet files in my workspace i've got python code i've got our code i want to be able to just click on the csv file and see see what's in it and see some summary statistics like i've got some parquet files like i want to be able to just click on them too and and be able to look at what's in them and i shouldn't have to write a bunch of code to analyze it like why not just open the file look what's in look what's inside and turns out that that's surprisingly not only useful for humans it's useful for agents too