Why distribution and attention are the real bottleneck in B2B SaaS | Pete Hunt @Dagster

Speaker A: Foreign.

Speaker B: Hey there. Welcome to another episode of Sauce Unbound. And I think for a lot of technical folks, I don't actually need to introduce you, but, uh, but still, welcome to the show with me, Speed Hunt from Dijkstra Lab.

Speaker A: Thanks. I'm excited to be here.

Speaker B: Sure. Awesome. Like I said before this episode, I was kind of bragging a little bit here and there that I'm going to interview you and I'm godlike Creek welcome requests from developers from the team and just folks I know to ask you a bunch of questions because they're like, yeah, we're excited AI and all, but we're also super pissed at it. So, like, ask him about developer experience. Like, okay, so, so, so the things are coming.

Speaker A: Well, it's an exciting year, isn't it?

Speaker B: Yeah. But before we get there, for those who are listening, who don't know what Dexter Labs are doing and where you're coming from, because you've got a hell of a story. Maybe in a nutshell.

Speaker A: Yeah, I can give a quick address. So I'm. I'm the CEO at Dagster Labs. We build an open source framework for building and operating data pipelines called Dagster and we build a commercial product called Dagster plus, which is everything you need to run Dagster in the cloud or on your own in your own environment. I come from a technical background. Originally kind of kicked off my career, got a master's in distributed systems, but somehow found myself working on full stack product instead of a large scale backend infrastructure. Started my career at Facebook pre ipo. Found my way to the Instagram team early on and when we were building out their first web app there, we ended up becoming kind of like the first users of React js. And um, I ended up becoming a core team member pretty quickly and kind of the, uh, one of the main people on the team for the first couple years of that project's life, which really haven't gotten open source. After that I started a company, sold it to Twitter in 2018, stayed there for a while working on a bunch of stuff, and now I'm here at Daxter.

Speaker B: Awesome. So you've seen Code Facebook. What was so bad that you turned CEO?

Speaker A: Oh, um, man.

Speaker B: What have you seen?

Speaker A: It's almost the opposite. The way that every time I've been CEO, it's been because nobody else wanted to do it. It's funny. No, in this most recent gig, I joined actually as the head of engineering. And the, uh, founder is my good friend Nick Schrock. And he's now the cto. We worked together for like a year, and by the end of the year, he was like, listen, I really want to lean in with customers. I really want to do, like, fun, creative work. I don't want to do any of the CEO stuff, working on comp plans and buying D, you know, insurance and stuff like that. So how about, Pete, you be the CEO and then I'll be the cto? Yeah. Thought about it, and in many ways it was kind of picking up where I left off with my last company. And so I decided to take the leap. I actually, like, I wasn't running away from engineering at all. It was much more like the job needed to be done by someone. And why not me?

Speaker B: Yeah, right. I mean, jokes aside, obviously I want to come back to the fact that he said he wanted to do the fun stuff and then became a cto, but I actually meant more of a how was it? What kind of a, uh, mindset shift was it for you? Because don't get me wrong, I believe that for some folks, developing and being a CTO is fun. But as we're acquiring companies and we see a lot of instances where a technical founder will do absolutely everything in their power to not become a CEO. They will sell a company, they will start something else, but they will not basically become a CEO and run a company. What it takes, do you think, for an engineer to become a CEO?

Speaker A: I think that this in many ways actually ties into the AI stuff that you were talking about. Because the way that I've always seen engineers, when I was like, hiring a lot of engineers over my career, I would always kind of interview people and I would always think that there were like, two types of engineers. And I used to say that there were like, user motivated engineers and technology motivated engineers. So, like, engineers that really get excited about solving technical problems and be like, technology motivated and engineers that don't really care about the technical problems, but they really want to, like, have built something that, like, people use and. Or, like, sell a lot of copies or whatever, sell a lot of seats or whatever, you know? And I recently saw it put on Twitter in a different way that I think is a lot more elegant than how I was putting it, which is like, there are people that care about the journey and that there are people that care about the destination. And. And I think that the engineers that are more excited about the destination tend to do a little better in the CEO seat because the journey is so much like, in many ways, the destination is the same destination. Like you're building products for people and you're trying to build a successful company, but the journey is like totally different. Whereas the engineers that really focus on the journey and that's where they get a lot of their energy and their pleasure from, they hate it because it's like not solving technical problems. It's not a lot of creative work. And I've also found that you, ah, see this divergence with AI too, where it's like, you see some people that are like so happy and excited and they're motivated and they're pumped up about what's happening. And then there's others that have this deep existential dread and they're so mad and angry about it. And again, it's like the people that are focused on the destination are like, oh, the AI gets me to the destination faster. Great. And the people that care about the journey are like, we just like, we didn't do anything creative or fun or interesting. It just sat there and let the AI do all the work. It's like, interesting that this, like, you asked that question because it's like, pulls together a couple of different threads I've seen in my career and that have really come to a head over the last year.

Speaker B: So coming back to what now the CTO said about the fact that he wants to do something fun, because that's exactly what I started thinking about because I live will the developer and every time he comes to like a difficult problem problem, he's like, okay, it used to be so much more fun. Like he loves to do this and like, just code and stuff. And it's like the cognitive load and the fact that AI, you know, every time you give it a task, it gives you like a lot of stuff. Like it's a huge chunk. So instead of like being in a flow and going from instance to instance or feature to feature or whatever, you just have to sit there and review the code and it dumps more and more and more on you. So at the end of the day, lots of developers, like you said, they're just exhausted and it's not fun anymore. So how do you see it like in your own experience and maybe how do you deal with that in terms of developer experience in the team?

Speaker A: Yeah, yeah, that's a great question. So you know, our. There's many ways to answer that. Right? So first of all, you know, in our organization there's many people adopted these technologies in different ways and they still use them in different ways. And I'm a big believer in like. And maybe this is like a Little old school, but I think that you let the people use the tools they want to use and do work the way they want to do the work. You gotta have some standardization so people can work together, obviously, so it's not like completely chaotic for the most part. Like, if you didn't care what IDE people use, like, why do you care, like, how they use AI? So like what we tried to do was tell people about it, tell them about the best practices, and then what ended up happening was just like the pace. Certain teams would just like take off and have like really fast pace and other teams would be slower and then we'd be like, okay. We would just look at it as basically like, hey, how do we get everybody up to the same pace? And that required a lot of sharing of best practices and enablement and stuff like that. But it's still a challenge, right? And I think that maybe I'll eat my words or whatever, but it does seem to me that like we're in this weird intermediate period where AI is generating lots of code that still has to be reviewed and eventually the code's not going to be reviewed anymore. I'm having trouble seeing a world where that doesn't happen. You know what I'm saying? Like, there's just like too much code. Like, we just have to accept it, I think, and think, okay, like, in a world where like slop lands in master, like, how do we mitigate? Like, how do we deal with that, right? And so I do think that there is going to be a lot of exciting work around like feedback loops. How do you get the AI to test itself? How do you make sure that the specs for the thing that you're building, like, are actually followed by the model? And like, at the end of the day, our job as humans is going to be to like, basically do AI alignment work. Like, you've heard about all these labs talk about alignment as if trying to prevent humans from like AIs, from like having a Terminator style event blowing up the world or whatever. I think of it in a less dramatic way. Like, how do like a alignment to me is like for my given domain job, how do I make sure the AI is aligned with what I wanted to do? How do I make sure that it like hits the quality checks that we need to hit, does the thing on time, at cost, whatever, or on budget. And I think there's a ton of work to be done there that is going to be exciting and motivating for some people and it's going to be Miserable for other people. And one of the things what I've been telling people is like, you know, I made my career early on react, which is a front end thing. And I hated front end. I had a focus on distributed systems in school. That's really what I didn't enjoy Facebook to sling JavaScript. I enjoyed Facebook to work on like the next great distributed database. And what I ended up doing was like I really enjoyed working on products and I had to do this really unpleasant part of my job, which was work on front end, hard to test. It's very like squishy. The technology stack back then was awful and I basically tricked myself into enjoying that work. And I did that for long enough that I just ended up enjoying the work by the end. And I think that if people really try to hypnotize themselves into liking the work, they'll actually get into it by the end. Does that make sense at all? I don't know.

Speaker B: Yeah. But I do have follow ups. And you know, first of all is, I mean you've probably seen it, right? Because you said run your budget, right? AI. I don't know, does AI run your budget or not? But it's a great example, right, because there's a lot of uncertainty on predictability or better yet, unpredictability of AI, right. And trustworthiness. And I've been interviewing quite a few folks, I mean, in the last year, nobody sat here in front of me that said, oh, we don't use AI. They all do to some extent. And my thought is always, okay, so we didn't have that thing and now we do. And it's kind of magic. We understand the concept, how it works, right. And we do have this general wish desire to keep it in place and to have constraints and blah, blah, blah. And yet when we were adapting AI, mostly me, myself and people I know, but we just basically went there and dumped everything about us in their budgets, taxes, personal information, family information, where we're traveling, what we like, our work, what we do. And I'm just thinking, because you guys are not just working with this kind of very basic information, you're handling tons of data. How do you deal with that?

Speaker A: Well, fortunately, Dagster as a business doesn't actually touch a ton of data. We have a deployment architecture called hybrid deployment architecture where we handle mostly metadata. So like the names of tables and the last time they were updated. But in terms of the contents, the customer data that's within the tables, we generally don't have access to that. Unlike previous roles that I'VE had where, you know, we were working in trust and safety, for example, trying to like identify fake and compromised accounts. That was a lot of sensitive data. So it's an area that I'm very familiar with for our business. Again, like, the risk profile is probably a bit lower than a lot of other businesses, but we do have policies around it. We are like, there's new agent security products that are coming out all the time that we use to kind of make sure that we're not putting customer data into. You can't just take customer data and throw it into some new, um, system AI or like regardless whether it's AI or not. Like, you just need to make sure that you understand where customer data is flowing to. And this is like. It's just that there's many more people a lot more excited to throw it at AI. And so we just made sure that we're clear on the policies and we, we only use like kind of a couple of approved vendors. I do think that these local models are pretty cool though. I don't know if you played around with any of them. They work pretty well.

Speaker B: You mean like Open Claw?

Speaker A: Well, I'm thinking like Gemma or Quen or Deepseek. Uh, any of these local models.

Speaker B: No, I haven't tried them.

Speaker A: Um, they work pretty well.

Speaker B: I chickened out.

Speaker A: I don't use openclaw. I'm too scared for that thing. But yeah, the local models could be, could be an elegant solution to these problems.

Speaker B: Probably a bit out of my scope, but it's heard this hypothesis about new thing that's coming to Apple, right? And the fact that maybe they're plotting something a lot more intelligent that we all understand as in bringing all those little machines with our own LLMs to our own home where it just runs itself. At the same time, I'm trying to explain to my mom what LLMs do and like what my job is and she's like, so. So it's like talking to a phone and it talks back and you're like, yeah, okay, that's it. Unreal. Yeah, it's. Half of the planet is still like this. I'm pretty sure, you know, the bubble is more inflated than we think it is and it's more. People haven't really seen pre version of ChatGPT than we realize.

Speaker A: Yeah, yeah, I know, it is crazy. A lot of people's only work experience is with Microsoft Copilot and like Office and it does not work as well as Claude it at that.

Speaker B: But okay. So we kind of grasped a little Bit of just the general concept, but since you guys are like, you need to know where the data is coming from and where it's going. Right. To kind of just put it all together. I wonder how much AI changed the way you're building product. Or it's a good mix of yes, there is AI in the product and how we're building it, but it's also just demand and a promise.

Speaker A: Yeah, it's simultaneously changing very quickly, but also like not at all that makes sense. What I mean by not at all is when you go into a buying cycle with a customer, the buying cycle is like pretty much the same. Now that it's been last couple of years, we're seeing a couple of things change. The first is there's many more of these buying cycles now. There's a lot of lot more data being used, a lot more data intensive applications being built out. The second is like the way that we build internally is changed a lot. And so we have teams of three people doing what would have been a team of 10 or 20. Basically 20 might be excessive, but 10 I think is realistic. We tackle product areas that would have been like very irresponsible and insane to go after pre AI. And there's a certain amount of knowing what is amenable to AI and what's not. For example, stuff that's not core infrastructure and if you push a mistake, there are different categories of bugs and mistakes. For the stuff that's lower stakes, you can like automate huge chunk of that and have one or two people move super, super quickly for categories that are like critical infrastructure, database migrations, writing basically anything that writes to a database, like a source of truth system, you still need to approach that pretty carefully. And so we use a lot of AI to kind of like write the code, but it's like reviewed super, super carefully and you don't get the same degree of speed ups that you do for the lower stake stuff. You can also build features now that were like impossible to build a couple of years ago by putting AI into your product directly. Whenever we want to do like a, we can put recommendation systems all in our product now, like, oh, you know, like, looks like you looked at this thing, maybe you should go look at this other thing too. And that's super easy to do now because we have LLMs. Um, right. And so it is a pretty, pretty brand new world and we're still orienting and trying to figure out what's going on. Frankly, as many are as.

Speaker B: Yeah, it's all of us. Yeah, okay. I wanted to come back to what you said about the teams, right. Because I had a chat with Eric Rees from Lean Startup and he said they've been running some research analysis on productivity with AI and like the self, what is it? What's the word I'm looking for? Self reported productivity increase is about 20% but the real one measured in the systems and amount of code and features shipped was actually minus 18, which is a huge difference. I don't know if it's ethical question to ask, but how do you measure if at all developer productivity? That's a thing. And if you can for sure say that, hey, with AI there are certain things that can be absolutely automated and then re automated with reviews and blah blah blah. And there are certain things that we're maybe better look at it yourself. What are the parts that better not be automated and better still be just human owned? This episode is sponsored by Rewardful. Looking for new ways to find customers for your SaaS business? Consider building an affiliate program. Rewordful is the easiest affiliate tracking platform to set up, manage and scale skill. For SaaS companies, building a successful affiliate program can be a little bit intimidating at first. And that's why Rewardful has taken what they've observed from their most successful customers affiliate programs and distilled that into an exclusive online course. The exciting part, their affiliate marketing course is absolutely free. Start the course at Academy Revolution rewordful.com academy.rewordful.com and turn your biggest fans into your best marketers.

Speaker A: Like an objective measure of developer productivity, really? I think there was this like period of time where kind of people thought like, oh, you know, it is impossible to measure developer productivity if somebody's not committing at all. That's usually a problem, right? Depending on what they're working on. We know Commit Velocity has increased a lot. We don't use that for performance reviews or compensation or anything like that. That'd be like pretty weird. That was one thing that we did observe increased. We don't really look at lines of code. I think it's like a PS metric. But Commit Velocity is kind of like a little bit better. What we've basically found is that I can't like slap a percentage on it, but we just think that like we're just way more ambitious with projects down. And if I were to really reflect on our expectations now for like how long things take, it's like what used to be a week long project now takes a day. When I think about like recent features that we built in some of our products I would say it's yeah, five to one compression or three to one compression, something like that for the low stakes stuff. For the high stakes stuff, there's not a ton of compression.

Speaker B: Okay, I'm uh, going to ask the question. So a lot of companies have been saying, okay, I now allow us to not hire people or hire less people, right? Or lay off some people. What is your philosophy here? Because now it's a little bit like VC versus bootstrappers. Now on social media, right? Some companies say, okay, with AI we can hire less and some are uh, going to be better, quote unquote saying no, with the AI we're still hiring more because eventually it's people who are taking responsibility over what's going on. So what's your stance?

Speaker A: We haven't really changed our plans at all from a hiring perspective. The way that I think about it is for venture backed companies, they need to deploy capital quickly in order to tackle a large market opportunity. And they basically want to deploy that capital as efficiently and quickly as possible. And they're not as concerned with like cost cutting, you know what I mean? I think that like, if you're really focused on profitability then maybe extreme profitability, maybe you start to look at that. But for a lot of these like high growth startups and these AI native companies that are tackling large markets, they like were already so constrained by labor that even massive productivity gains, like they, they just deploy those against the market. That's how we've looked at it anyway. It's like there's basically like so much of a market out there that to tackle for us that like we, we didn't like reduce hiring or anything like that because of AI. We just like got more ambitious basically. Now I don't think that applies to every, I'm like a realist, right? I'm not going to be like say that there's not any AI layoffs out there because like I think there are a couple. But I think different businesses have different constraints and for a lot of these high growth places that like people really want to work at, there's a big market that they're going after for a reason and people shouldn't be too worried about it. That's what I think.

Speaker B: Okay, all right, so coming back to the fact that you guys raised something about 50 million, right? Correct me if I'm wrong, it's 62.

Speaker A: Yeah.

Speaker B: Okay, so I get a, uh, lot of bootstrappers here on the podcast and it's always about, okay, where do you spend and for them, it's really life or death kind of situation. Right. You spend it the wrong way and that's it. So where does a $62 million funded company prioritize their spending? And if product is becoming cheap or cheaper, allegedly with the eye, is distribution where you're spending?

Speaker A: Well, it's definitely not on like water slides and caviar.

Speaker B: Damn it.

Speaker A: I wish. Yeah. You know, I think that what everybody's realizing is, and you've seen this even at the big AI labs, right? They've launched so many products. You know how many products that like OpenAI and Anthropic have launched that have just been duds, Right. Nobody's picked them up or used them. And I think what people are realizing is that product, market fit, attention, distribution, these are the actual bottlenecks in the business. Like, often products, you need enough product bandwidth to like run the experiments and like find the product market fit in the segment. But like a lot of times for most businesses, it's attention, it's sales, marketing, it's all that, all of that stuff. And so I would say, like, we've seen an equal or bigger transformation in how we go to market than how we develop product. And I think it's a super important place to look. And that's really where I think AI is having a big fundamental transformation in our business is go to market and top of funnels specifically, and you ask kind of like what jobs should humans really own? And it's like, I do think for B2B market, enterprise sales, a lot of buyers do want to talk to a person, or at least they spend more money when they talk to a person. So, like, I think that's going to stick around for a while. Forever?

Speaker B: Oh, absolutely, yes, for sure. Like, I've always believed the bigger the check, the more you have to be there as a person and establish the relationships and just make sure that you know your customer face to face. Because nobody's going to write a multimillion dollar check to you after cold email that says, oh, seems like we went to the same university. And you liked hedgehogs. It's a true story.

Speaker A: I get tons of those.

Speaker B: I got that one.

Speaker A: Yeah, yeah, I didn't get hedgehogs, but I got the version of that. Yet.

Speaker B: It's a true story. I shared it on the Internet, apparently. I don't know, maybe AI picked it up and, uh, now people know I had a hedgehog, so.

Speaker A: All right, well, a bunch of BDRs just teed up some emails to you.

Speaker B: Yeah, absolutely. So what Works for you. When you're building that brand and that trust with your customers, what works and maybe even more importantly, what doesn't?

Speaker A: A lot of this stuff is not like super bootstrapper friendly like trade shows for example. They're like a ton of money to sponsor booth, but like they're very, very effective. I think if I were a bootstrapper ticket to like, you don't want to spend 5, 6 digits on a booth, but like you can buy a ticket and then go there with your company T shirt on and then have face to face meetings with all these people in your network and customers that are going to be there. And I would say do something that's not just like going out to dinner on one of the conference nights. Try to do something fun, like if it's in a cool city, go find some cool thing to do in that city and take them there and hang out. But just try to form like an authentic human connection with people. And I think the way to like the way that I've approached it anyway is like, if there is this like aura around you of uh, good things happening and good like people coming away with like good feelings, then the success will just like, like you're basically creating an environment where you're more likely to get lucky or like have somebody take a chance on you or whatever. So like I wouldn't go into the conference being like, I gotta like sell this account, upsell them 30%, be like, listen, I'm just gonna like go have fun with these people and there's going to be a renewal conversation later this year and we'll have better rapport when that happens and you know, maybe that'll result in something good, I don't know. But like really focus on just like people want to have fun at work, even with their vendors, even with their counterparties and just do your best to include them and be, have a decent human.

Speaker B: Oh yeah, a hundred percent. I wanted to ask you how you know because you have an open source free version and probably community behind it. But at the same time for Daxter plus, uh, you're going after bigger companies. How do you connect those two and if one kind of pushes the other?

Speaker A: Yeah. So, um, open source raises awareness for the commercial product. We started out as an open source business and I come from the open source world. The open source product has to be really good, especially in a category where we're like the second mover, not the first first mover. And so open source is very key to what we do. And for a long Time we didn't even have a commercial product. And so open source is all, all that we did. But at the end of the day open source is not free. Like somebody has to do the work and even if AI is writing the code, someone has to pay for the tokens and someone has to verify that it works and deal with the security incidents and stuff like that. Right. Your options are either have large mega corps run the open source projects like Amazon is a uh, Linux kernel contributor or whatever. Right. And that can be done via uh, foundation or not, but like it's still a similar thing. Second option is you can have maintainers burning themselves out and not having a good time or you can have an economic model behind the open source project and fund it through a commercial product. Daxter plus is our take on the third model. The basic idea here is if you are a company with thousands of employees and you are deriving a lot of business value from the open source product, you should probably pay us for building it. And the way that we're going to do that is we're going to build adjacent features in our commercial products and capabilities that you would be willing to pay for. So this is like enterprise security stuff. It's hosting, it's multiplayer and team features. So companies that have two or three data people, they don't need a ton of collaboration features. But like we get into teams of hundreds, you do. So we really try to segment by that. Right. Like once you're in the hundreds, uh, or thousands of employees, our commercial product becomes pretty valuable for you. And if you don't want to use it, then don't like it's an Apache 2 project. Like go for it, it's fine, but we think you should. So yeah.

Speaker B: Okay. All right, cool. There is another question that completely forgot about, but somebody told me that apparently, and again, correct me if I'm wrong, maybe that's outdated or something. Apparently you are running your development cycle in cadence as an ship and then stop and work on debugging or technical debt or whatever. First of all is that if that's true, how does it work? Why like that? Why no instant housekeeping after you ship something small?

Speaker A: Uh, we always play around with our development lifecycle. The current one that we're working on that we're doing right now is we'll have a four week development cycle where people will. It's kind of like a traditional sprint, like from Agile, you know, here's what we want to accomplish over this period, let's go do it and then we'll Go and like do the work and ship it and land it and cut a release and then we're done. Hooray. And then we have a two week cooldown period. And the cooldown period is product people and tech leads and managers get together and they think they're like, okay, like what do we want to do next? We're like, thoughtful, we're like, let's look at what we just shipped. Let's take a little bit of time to think about it, do our research, plan the next thing. And meanwhile most of the engineers are able to have time to work on quality work and cleanup work. That kind of stuff that only they know has to happen if that makes sense. You know what I mean? Like if you're an engineer, you've got kind of your like top down business moving work and then you've got like the uh, sometimes it's technical debt, sometimes it's other stuff but like things that you know, have to get done at some point point, but you never have time to do it. We try to build that into this like cool down period. And I think so far like people have really liked how that works. You know, we've got pretty good product quality. I think at this point we've been doing this for maybe six months so it's not like uh, a long thing that we've been doing forever. But so far it's been pretty good.

Speaker B: Okay. I've been, I've been talking to another founder here who's super technical and he said, yeah, like with AI, our product is becoming more and more relevant for even non technical folks. And I asked, okay, well that's a great market. That's probably opening up a lot of opportunities for you. Are you guys doing anything to make it easier to use for my kind of people? And he was like, no, are you kidding me?

Speaker A: No.

Speaker B: I want to talk to a person who knows what a quiry is and not um, have to explain it to them. And it's so funny but at the same time it's very good I guess. So for what you're building and for the community you're serving. I'm um, listening to you all this time and I'm thinking the team is probably like 99% engineers and the audience is engineers. And even if maybe the person who's holding the budget is not, you're still making it easier and more enjoyable for the end user.

Speaker A: Everything does seem like it's becoming engineering, doesn't it? Yeah, it's pretty interesting. My observation is that everybody's becoming like this amorphous maker where designers are writing code, engineers are designing things, data analysts are building front ends to the data because they know what they need to see. And, um, everyone's kind of like mashing together and able to do a little more of their adjacent domains. And you're also seeing this on the customer side too. And so what I think you're seeing in the market is that solutions that are really customizable and that are more like a LEGO set are being rewarded and people are really excited. They want to be able to rearrange things and put them together in ways that the vendor didn't expect. And these kind of all in one monolithic solutions are less, less exciting for people right now. So I don't know if this is a durable trend, but it certainly feels that way for kind of the early adopters of technology that they're really thinking about at time the same this way.

Speaker B: What are their, the tools that lately got you excited?

Speaker A: Um, I switched to Codex recently from cloud code and that's, that's been really.

Speaker B: It's going to raise some eyebrows.

Speaker A: Yeah, I know it's faster and it's actually up and the model is almost as good as Opus, so. Yeah, I really like that, man. What other tools have I been using? I've just been mostly like writing a lot of. Oh, I, I used 11 labs to rig something together the other day. That was, that was a really positive experience.

Speaker B: Yeah.

Speaker A: Uh, cool. Yeah. Awesome.

Speaker B: Yeah, I love to create little things and it's just for me it's like, it's amazing. Right. But it's also a trap because you're like, oh, so fun. Let's create something else and let's create something else. And it doesn't really get anywhere. It's like this slot machine. Oh, that's so fun. Or like, you know, this, this claw. Oh, okay, we got something.

Speaker A: Yeah, yeah, yeah.

Speaker B: The rest of the time it's just whatever.

Speaker A: Yeah, you gotta have like really good inner game and make sure that you're like focused on the thing that you're trying to do.

Speaker B: Yeah. Nothing that is super niche that we're all missing.

Speaker A: I don't think so. I've been playing around with like the Gemma 4 series of local models. Maybe that would be the only other thing. But no, I don't use whisper flow. Super niche. They just raised a zillion dollars, so I don't think it's super niche anymore. Yeah, no, but pretty, pretty basic, I think.

Speaker B: Oh, me too. Yeah, I got spooked by somebody told me I don't know what it's called. I, uh, forgot. But it's a wearable saying. It's kind of like an air attack, but with the eye. And it listens to you like basically all the time. And Instagram does that also. But we try not to think about that. Right, but that's the thing you wear. And like, you wanted to listen to you. And I was like, oh, no, no, I'm not ready. I'm not ready. It's like that robot where inside. It's actually some dude in India that. That's. I don't know. That's too much.

Speaker A: Yeah, yeah, I'm not ready for that either. Personally, I don't use AI note takers. I use them for like a couple of group meetings maybe, or like an interview, but that's about it.

Speaker B: Okay. That's what's going to be the headline of this episode. Pete Hunt not ready for.

Speaker A: Yeah,

Speaker B: okay, well, awesome. The last question I always have here is about a hack. And it can be anything. You know, running a company, building a product. Stay insane. With all the things that are happening with AI, anything that you think other CEOs, founders would appreciate.

Speaker A: I just think now, um, more than ever, cultivating something outside of work that is equally important to you as work is very important for some people, that's a hobby. For me, it's my family. More important than work. When somebody on the Internet tells you to go outside and touch grass, you can actually go do that in whatever way it works for you. Then you get energy from that because it is a very fast moving market right now. Lots of changes and nobody knows what's going on. So it's important to stay grounded.

Speaker B: Absolutely. Uh, yeah, I saw this. Yet another post about how important if you want to achieve something in life to, to not have work, life, balance, and to just grind nine to nine, basically, and not, uh, prioritize your health or family or whatever. And I thought no, just no.

Speaker A: Maybe for a couple weeks, but not forever.

Speaker B: Yeah, yeah, exactly. All right, thank you so much, Pete. I think the, uh, developer team is going to be very happy about some of your answers and it's great to see what you're building. You know, it's super exciting, but also, I've got to say, very confusing.

Speaker A: Yeah.

Speaker B: But it's good that you admit that. Yeah, it's good that you admit that. Uh, sometimes you don't know what you're doing. It's.

Speaker A: Well, that's fine.

Speaker B: So thank you for that. Yeah, thank you so much for your time and hope to do it again sometime. Thanks for listening. SaaS Unbound is brought to you by SaaS Group. We're a long term home for a great B2B SaaS. We buy, keep the team and brand DNA and help with the boring stuff like hiring and finance so founders can truly focus on building great products. If you're a founder who'd like to be featured or explore an acquisition, reach out through the form on our website or email me at. Uh, Anasas Group.