Subscribe
Share
Share
Embed
AI After Dark is a podcast hosted by Alex Gras, venture capitalist at Mercury, focused on how real companies are built once the hype fades and the hard decisions begin. Through candid conversations with founders, CTOs, and operators, the show cuts through buzzwords to talk about people, systems, risk, and the trade-offs that actually matter. Alex brings a background as an operator, founder, and revenue leader, with a belief that technology matters, but people come first. This podcast is for builders who care less about trends and more about what lasts.
00;00;00;00 - 00;00;30;08
Unknown
I gotta ask, like, did you ever think you would be a CTO? Absolutely not. Yeah. You're kidding. I was, I was historically like, designer, the first designer always in. Right. So because of that, I had to be technical, but, like, no man, like, even starting Dream Base. It's Dream Inc. is the company. But when we were starting out a couple of years ago, it was like, I was trying to pull my other friend, the other, eye guy from New Orleans.
00;00;30;13 - 00;00;49;06
Unknown
Actually. That's funny. The other two I did. Apparently, Louis Landry is my my old time friend, and we used to be joined at the hip and two sides of the same coin. He was always the CTO, and I was. I was like the CDO chief design officer, kind of a thing. And, since I was always in the room with those guys by osmosis, I became smarter about technology.
00;00;49;06 - 00;01;08;02
Unknown
But like it was arm's length away until I all right then I was just like all the things that they've been saying that I've been designing around for the past 20 years, I was like, I can do this now. And I don't even know if we're actually rolling for the show yet. But like, Andy, my co-founder, was just like, do we do we need a CTO now?
00;01;08;05 - 00;01;25;25
Unknown
Like, I think that you and I was like, shit, maybe it isn't you, man. So it actually just happened organically. But what was what was the like, isolated moment where you were like, oh shit, I can't do this. I don't need, like a formal title necessarily, or to have some kind of formal education pedigree or whatever. Yeah. Yeah.
00;01;26;02 - 00;01;47;15
Unknown
That's it's it's a great question. Am I actually. Degree is 3D animation, which is moderately technical. We were the first to date myself with the first 3D animation class out of Louisiana, graduating class of six people. Around the year 2002. But I think I didn't feel like I had permission. And, man, like, designers are great at imposter syndrome.
00;01;47;17 - 00;02;09;27
Unknown
So, like to to become the CTO was, like, the biggest imposter syndrome I think I've ever had. But it was just that I had had the privilege of, like, working at AI company before generative AI started. We were doing AI compliance. The world's worst, the world's first AI compliance platform. And I got the front row seat and I started prototyping with them.
00;02;09;28 - 00;02;31;01
Unknown
Like, I've always coded. I've been like a designer developer, but never full stack. Right. And then those four years from the pandemic to 2004, I was already steeped in coding again and building platform. So I was like, yeah, I think, I guess I could do this. I just thought I needed permission. You know. Yeah. What? Who I guess.
00;02;31;08 - 00;02;45;11
Unknown
How did you find the permission? I think it's one of the reasons it's, one of the biggest things I would tell people in the startup is like, go find a good co-founder first of all, right? Because if I would have done it myself, I would have never done it. Because I would not, like, felt like I had permission.
00;02;45;11 - 00;03;03;22
Unknown
But with Andy, my co-founder, he was just like, dude, you're totally doing it. He was like, we really wanted my buddy to do it just because he was the best, you know, technologists we've ever known in our lives. But he was like, you're the same thing, and now you can actually do it. And I was like, well, plus I have opinions.
00;03;03;24 - 00;03;20;08
Unknown
So I was like, yeah, I mean, I was stumbling over my words on that question because I was like, is it a who is it a what is it a, you know, moment? So it definitely was Andy, your co-founder that like motivated. Yes. Nudged you. You pushed you Andy and my consultant. You were talking about you consultant Claud earlier.
00;03;20;08 - 00;03;40;25
Unknown
Yeah. Like, oh, my consultant seem to think I had a lot of capabilities now. So, honestly. And it's just confidence. That's one of the things I love about what's happened with AI is giving people confidence to do things. Now, there's a lot of false confidence that speaking of compliance and governance, like we should be terrified of some of it, but if you do it cautiously, man, I think anybody can be whatever they want to be.
00;03;40;25 - 00;03;59;29
Unknown
Now it feels like, you know, we're talking about engineers doing design. Designers doing engineering. Cat sleeping with dogs. Dogs sleeping with cats. Like, it's. It's whatever you want to be now. So I feel like if you're a designer, developer or whatever they call it now, a designer, engineer, design engineer. When I was younger, we didn't have those terms.
00;03;59;29 - 00;04;29;07
Unknown
It was just design. If you have those inclinations, I think now you can be that right. You can step into that role, and you're seeing, like, the rise of the solo founders two for sure, which I still I just kind of said you should find a co-founder, but I think we're going to see that if you have imposter syndrome, yes, you kind of need someone to be able to balance it, to tell you the things that you can't see because you're in it, you know, or sure that someone told me it was like some, you know, some guru or whatever.
00;04;29;07 - 00;04;45;19
Unknown
He was like, yeah, you know, we're sitting 3ft or 4ft from each other. If I don't brush my teeth for a month, like you're going to smell it and you're going to notice, but my nose is like an inch from my mouth, I wouldn't be able to know, you know, like you need someone to be able to, like, tell you that, and but it is wild.
00;04;45;25 - 00;05;08;17
Unknown
So our episode two is with the Keeley CTO of a company called Pi. He was, it was very much a technologist, you know, fastest coder at toast kind of thing. Build modular. And he's like, what I realized is like, man, design was much more critical. You know, the UI, the way a user engaged with the product and figured out things with the product on their own.
00;05;08;19 - 00;05;25;06
Unknown
Yeah, it was way more important than anything. I was doing. He's like, I want to surround myself with the best design people possible, because that's what's going to move the needle in a, in AI. And like they were generalizing some terms, not to like people would just say builders or like, operators is also like the product side of it.
00;05;25;06 - 00;05;44;17
Unknown
Like, I like this term because it captures the hard thing to describe about people that just get shit done. And, you know, you want critical thinkers that can think in systems. Right? And historically, we call those people designers. Or on the technical side, we call them architects, which is architecture is design. Right? So like those people, this is what I was made for.
00;05;44;23 - 00;06;13;24
Unknown
If you can clearly articulate what you want and like call B.S. on AI when it's not doing exactly what you want, you can do whatever you want. Now. Especially, I don't know if you just saw opus four. About seven just came out a few seconds ago. Yeah, yeah, yeah. This morning. Right. Yeah. And with like the rise of 4.5 and 4.6 and now, GPT three, like where they're going with, like with Codex and the model 4 or 5, like we're here now, like I think you can actually autonomously go build something and just go check in on it.
00;06;13;24 - 00;06;30;21
Unknown
Right? Dude, I, I didn't even like I don't think I've opened terminal ever until about a month ago. And now I've got three terminals open with, you know, cloud code running, building different things. I, my wife the other day, like, was looking at like, if she's like, who are you? Like, what is going on? What are you trying to hack?
00;06;30;21 - 00;06;54;20
Unknown
And it's like, no, it's just like the AI's lowered the bar. So much for people to to build. So some something I guess the flip side of that, another CTO we had on here, I won't name because he's like, he's still working through it. And he was like, man, the way I've built my team is wrong. Like, and not just that, but like the he's like, I was going through a one on one review with one of my team and, and everything I'd asked him to do.
00;06;54;25 - 00;07;28;24
Unknown
He executed and he grew. But none of it matters anymore. And he's like, what do I do with that? And so I've organizational design has been like a really difficult thing to think through now. Oh, man. Look, I hate to be fair, it was it was ready for a revolution. I before all this, the reason I jumped into that previous startup I talked about was because I was, at Teradata, a 40 year old company, and I was running global design and UX, and I had a team over 50, and I was so far away from the things that I was good at and love to do.
00;07;28;25 - 00;07;45;17
Unknown
You know, I turned into more of a therapist. It seemed like. And no offense to some people babysitting, and, like, getting back to be hands on. Yeah, I think you're totally right. We have to rethink how do we form the best team, and we're doing it right now. We just got the five people in our team.
00;07;45;19 - 00;08;04;08
Unknown
And the people where we're selecting are just people that, like, are passionate, that get it done, articulate collaborators. But like, how far should we overlap is a question, because everybody should probably own a domain that they can just go crush, right? Like so. We don't need to be doing the same thing, but we need to be working throughout the day, checking in on each other.
00;08;04;08 - 00;08;23;21
Unknown
So we're trying to figure out, like when we go from 5 to 10 and more, what does that look like now and how big do you need to get right? I'm just wondering what the max size of of an optimized AI team looks like. And of course, enterprises that exist have to figure this out and retrofit, and that's why they'll take longer and be crappier at it.
00;08;23;21 - 00;08;42;07
Unknown
But yeah, I don't know. I don't know how to do, to be honest with you. Like, I, I, I look at some of the, the enterprises that, you know, portfolio companies at Mercury sell to and and they'll have I mean this time they've they've been maybe even built data science teams ten, 15 years ago with the same understanding of what the use case they were trying to solve is.
00;08;42;07 - 00;09;04;04
Unknown
Right. Like, how do I prompt my data? How do I figure out, how do I disintermediate this intelligence layer? And and still aren't, you know, quite there yet. Aren't and especially now aren't able to keep up with like, the technological exponential curve that's happening. Well, I think a lot of team members weren't those people, right. Like you, you you needed more bodies doing different things with teams like that.
00;09;04;04 - 00;09;21;12
Unknown
I was just reading something yesterday about this, about like it's sad to say, but there's a lot of people in software engineering that were just doing what can easily be replaced by some background agent, right? So they need to up their game and evolve, and they're probably going to be the first ones to get pushed out a lot of the times.
00;09;21;12 - 00;09;44;17
Unknown
So I think if you're not like embracing this fully, and you don't have an entrepreneur mindset, somehow, which is unfair, right? You're born with a series of, probably mental issues if you happen to be an entrepreneur, but, like, they work together. Well, and you're and you're productive. So. I don't know, man. I think everybody it turns into a director kind of of their own team.
00;09;44;19 - 00;10;02;22
Unknown
And you have to have vision for your own area and you're like a mini CEO now, I feel like. And CEOs are more like the board now. It feels like, oh yeah, I hadn't thought about it that way. The only reason I, I'm stealing that from an actual product that I love, called paperclip and paperclip. And like, that was the mental unlock.
00;10;02;22 - 00;10;25;09
Unknown
Speaking of UX, like that was unlock for me using that kind of a product. I was like watching open call from afar. I tested it when it was called back and I was like, no, this is dangerous. And really rough. But it put like the polish and UX around it, and they designed it to feel like linear with GitHub issues and you're operating a team that's hiring people, right.
00;10;25;09 - 00;10;46;04
Unknown
And like you can when you put yourself in that mindset, you know, how to hire somebody and how to describe it. And when they're working together and handing things off, like you think of yourself as the board makes you think about resources differently and like direction. So it's funny, Andy's our CEO now, like the boards barking at the CEO, basically directing them.
00;10;46;06 - 00;11;05;15
Unknown
So I feel for him now after after actually using that product for a while. But no, it's I think embracing the humanity is the easy UX adoption pattern. And like everybody getting in the mindset of being a board or at least a CEO is where we're all going to have to go. Yeah, because it's clear that the agents are the employees, right?
00;11;05;18 - 00;11;29;23
Unknown
Not the actual employees. So let's all follow that pattern. I think what managing people is, is, is funny. I've always I've always enjoyed it. I've always liked people problems. Oh God bless you. Yeah I know yeah. Which is probably an issue. You talked about being a screw loose. I think I have that. Now to your point, it's not just like the agents as employees that you're building, but it's also how they engage with each other.
00;11;29;23 - 00;11;50;06
Unknown
Right. Orchestration of that, is I mean, have you gotten to that point now where you're like, how do these even talk to each other 100%? I think that's everybody's going through that right now. Anybody that's that's like, I feel like we're there's 1% or zero, 1% of the world developing the software. Are all moving around the same themes at the same time because, like, you see, skills have become everything.
00;11;50;08 - 00;12;16;05
Unknown
Even Chrome now has skills, which is like that's how you know, it's it's it's one the adopting. Skills is a thing and skill support agents using those skills. And then there's these agent harnesses which everybody's talking about. And the only point of those thing is like orchestration. So we're I think we and every AI product will have the thought of are we the place that orchestrates everything or are we part of the orchestration.
00;12;16;07 - 00;12;36;26
Unknown
So the thing that we're building is we're calling them the Dream Team because our product, our dream base, dream team of data agents, and we'll have, specialist. And really, the specialist is their access to context and their skills. So we'll have a data engineer, we'll have a data scientist, a data analyst and a business insights agent, obviously different focuses.
00;12;36;26 - 00;12;54;22
Unknown
So what we're working on now is like letting users say a simple task. And that gets handed off to an orchestrator, which is like the manager of the data team. And it's like, well, we need to modify the queries and we don't have all the data. We need a data engineer, go do this. And then once we get that we're going to configure a dashboard.
00;12;54;22 - 00;13;11;01
Unknown
So data analyst go do this. But once that's done the user is actually just asking for insights. They don't care about all that. All they want is like the output. So business analyst Business Insights go go create some useful insights that are actionable. So it's like being able to orchestrate all. That's what we're literally going through right now in our product.
00;13;11;04 - 00;13;37;08
Unknown
Yeah. And it's fun because at least for data nerds like us, I've been 20 years in the industry and surrounded by these actual physical teams. So since models are trained to be like humans anyway, thinking about how you orchestrate those teams to me seems like the right way to do it. Right. But is is it orchestration within the agents that you build yourself, or how are your agents also engaging with Andy's agents or you talked about overlap at the human level, right.
00;13;37;10 - 00;13;51;25
Unknown
How are you looking at the overlap at the genetic level? Yeah. Like within our own product. Now you're kind of referencing given like custom agents, what we want to do, like personalized agents amongst the people in Dream Base. But let's just talk about Dream Base has just agents for, like, a company like you have those for agents.
00;13;51;25 - 00;14;15;26
Unknown
I just talked about. How do you orchestrate those agents with whatever other thing you said? You're using cloud code now, right? Like cloud code. Everybody's using classes to orchestrate agents with other agents. So how do we plug our data agents into your company or your organization? Right. Whatever your agents have happened to be, that's what we're interested in, because I don't think we have, we're optimistic and we're a little delusional.
00;14;15;26 - 00;14;33;07
Unknown
I think everybody has to be in startups, but we're not so delusional to think that people are just going to come to Dream Base and just use our agents by themselves, like we're going to be part of an ecosystem of agents. And the thing that separates us from other startups and other companies is that we're focusing purely on the data piece based off of our actual, you know, 20 years of experience.
00;14;33;07 - 00;14;50;27
Unknown
So we're mimicking that to say, like, now you go do cool stuff with this. Like you've got this whole team that can deliver, what we're calling AI Native intelligence, which is like replacing by this is the new thing, right? So that's how we're that's how we're thinking about it. We'd be delighted if you just want to use it within our product too.
00;14;50;27 - 00;15;11;20
Unknown
We're going to design for that as well. Yeah. So so you and the rest of team are building Dream Base AI that currently sits on top of Super Base, right? So what what is, what is the delta between a, you know, old bi type dashboard and a dream based dashboard today? That's one question.
00;15;11;22 - 00;15;25;09
Unknown
The reason I say it's fun is because I was on teams that, we were the first team to use Tableau when it was coming out of a project. I was at eBay and like we had, we bought so many seats for Tableau, like to test it out. And it was great because it was a better user experience.
00;15;25;09 - 00;15;50;12
Unknown
And MicroStrategy, which was the older one funny thing, I ended up working at MicroStrategy after that, trying to design MicroStrategy like Tableau. When that seemed like it's easy. Now, you look at Tableau and it looks terribly complex from what we can do in dream based. So like the thing I would say is we do it all in 100th of the time, but with 100th of the people in seconds, and you just really describe the system you want and we help you build it.
00;15;50;12 - 00;16;11;03
Unknown
So sitting on a super basis is both our go to market wedge, because of course everybody is using super base now, but also we love them because of their back into the service. They've got authentication APIs, edge functions, storage like it's not just database. So we pick them so we could just sit on top and immediately deliver insights.
00;16;11;03 - 00;16;38;26
Unknown
And not only was that easy, it was unique because most people, use products like Post Hoc or Click House, which we integrate with, but they get data from copies of the data from JavaScript and click trackers and event trackers and stuff like mixed panels. Another one is that actually from the database, we were like, what if we could just sit on the database with all the new tools they have now and just read directly and give you a dashboard, and it turns out to work great.
00;16;38;29 - 00;16;53;08
Unknown
It worked a lot better than trying to get AI to do analytics by itself, because I are horrible at math. So the first version of dream, based on a so-called dream, we tried to do it and it was always inaccurate. We were like, we just can't go to market with this. And we were actually targeting rev ops.
00;16;53;08 - 00;17;11;01
Unknown
So we were like, we really can't go to revenue. People with like inaccurate numbers. As soon as super buyers came out with their MCP and we got early access, because they were actually early customers of ours, we were like, oh, letting the models have access via MCP doing SQL queries is great because the queries are doing the math.
00;17;11;01 - 00;17;31;00
Unknown
And like Postgres has great built in stats and analytics for queries. So like that's the story of why we chose Postgres, why we chose super base and like why we want to go to market with that. That was, you know, time flies. I was already like a year ago and now we're already now adding OAuth integrations so we can add click House analytics, post hoc stripe.
00;17;31;01 - 00;17;47;14
Unknown
We're excited. We're going to stripe sessions at the end of the month in San Francisco. And we're announcing some big stuff there. But like so now you can pull sales data. You can pull CRM data from HubSpot and all this, and you can leave it where it lives. We're not creating new data silos, right? We're actually just pulling it where it lives.
00;17;47;16 - 00;18;10;06
Unknown
But the source of truth is super base. And you want the source of truth to be your database, not a copy of the data in the cloud that was derived from JavaScript. Click analytics. Because if you've ever used click analytics, you're always like, why are the numbers different than our database? So after a long career of that and seeing like data lakes and all this crap happened, we were like, let's just rethink analytics from the ground up.
00;18;10;09 - 00;18;31;06
Unknown
And starting with Postgres. Wow. So I want to ask the, the, the new version of the old Google VC question, like, why can't cloud just build this? Yeah, it's a great question. And look what we see that there are they are on a streak right now. They're on a heater. Holy crap. It's like every day there's two announcements.
00;18;31;06 - 00;18;52;00
Unknown
You know like man that was that put out a startup out of business or that we wanted to do that. The great thing I love that they're doing is and I actually they're all copying each other. If you pull up, the latest UI that just got updated yesterday for Claude Desktop and now they have Co-Work, which is like their cool local thing, and Codex.
00;18;52;02 - 00;19;13;07
Unknown
And now cursor has a new thing. They were calling glass, but now it's like, an agent window. They're the same UIs, and not only are the same UI, they have very similar taxonomy and terminologies like marketplace of plugins that have skills for the agents to use and all this. So just like I described earlier, we're not delusional to think that everybody's just going to use dream based.
00;19;13;08 - 00;19;32;22
Unknown
We're designing our our software and all of our features to follow those patterns and to integrate with those things. So back to what you just said. Turns out we never thought dashboards were our unique thing. They're just our Trojan horse. Like, people need dashboards and they need them to be easy. So that was the first big feature that we came out with.
00;19;32;24 - 00;19;56;02
Unknown
But our secret thing that we were doing was actually getting the the AI semantic layer and the connections and understanding of your business. We do a lot of work in like and getting the context right and understanding your database and staying on top of it. So now we actually don't care soon if you use our UI at all, since we're coming out with our own CLI and MSPs and ways to pull our features into other products.
00;19;56;02 - 00;20;11;05
Unknown
If you want to use cloud, use cloud. They're the reason we started this in the first place. When they came out with cloud artifacts, which was the first time I could build a thing. It wasn't just a chat. It was like, oh, I built a PDF or I built a UI. That's when Andy had come down to Galveston.
00;20;11;05 - 00;20;30;20
Unknown
We were at a coffee shop and, we were talking about doing this thing, and he had a deliverable to do, and I was like, why don't you just try quad artifacts? And he was like, what is that? He's like, I don't know how to build stuff. And, 30 minutes later, he turned around his laptop at the coffee shop, and he built three different apps to do what would have taken him months with the team of engineers.
00;20;30;22 - 00;20;50;06
Unknown
And it's like, yeah, that's we should do that bit better. So that was the first version of dream, for analytics. That's a long story to come back to. Like they're always going to innovate ahead. They're always going to set the patterns. People are always going to like want to use them as their daily drivers. I do, but there's there's features and unique things to use for both.
00;20;50;06 - 00;21;11;11
Unknown
Right. Like we're going to specialize in that data layer. And even though they have connectors to those data layers, they do everything that's a generalist product, right? They will never do just data. They would that would be dumb. So we're going to we want to be the winner. The like the default choice for that stack. Just like you would use shad skin if you're going to use a UI these days are super bass.
00;21;11;11 - 00;21;32;16
Unknown
If you want a back end and database, right? Like we want people to use dream base for the data piece. When I almost feel like the because these these foundational models are moving so fast, they are like it's kind of shifting the adoption curve where like, I don't know, maybe it's not as a normal distribution anymore. And you have like these early early adopters.
00;21;32;16 - 00;21;52;25
Unknown
That's like getting, you know, squeezed because you even even if you are a technologist and you understand the space and you want to keep up with it, it's just too much, right? It's too much too fast. And so you have a actually a longer tail or a longer segment of the population that's like, yeah, I'm not I'm not even going to try that.
00;21;52;25 - 00;22;09;23
Unknown
Like, just give me something that that works over here because this is all I really need. And it almost becomes like the cloud for the layman, you know? And to your point, from a data being able to engage with data in a very natural way. Yeah, yeah. It's focused. Right. Focus is is a good word for it.
00;22;09;23 - 00;22;28;27
Unknown
And like the fun exercise we always put to people when they ask us questions about that or the natural question to ask is like, why wasn't super base build this? I first of all, I'd say they're all focusing on different things. Super base is putting all their energy into growing up with the customers that the customers don't leave, so they can keep the enterprise sized customers.
00;22;28;29 - 00;22;45;21
Unknown
So they're investing in something called multi Rus, which is this like read replica, amazing enterprise quality thing, which is funny. I worked at Teradata which did a version of that 40 years ago. So everything that's old is new again. But like they're focusing on that problem and they want you to use their platform and more of it.
00;22;45;21 - 00;23;05;28
Unknown
Right. So they have a version of reports. But if you sold, sporting goods, and now you have a history in like, sporting startups, like, or if you sold pickleball paddles, if I asked you go show me a really good product analytics dashboard and super base for that, you'd be like, where do I start? First of all, I don't feel like typing sequel, and I don't want the agent to help me type sequel.
00;23;05;28 - 00;23;27;21
Unknown
I just want the dashboard. So it's like, what if a product was engineered to do that easily is the first thing. Then I'd say, go try it. And Claude and Claude, look, I love it. Like actually these are new schedules thing that run every morning. And it's a good way to mentally exercise. Like where I was going. Like I want to integrate with five things just by telling the agent what I want and do it every morning at 9 a.m. and deliver a report to me or whatever.
00;23;27;24 - 00;23;45;13
Unknown
And you could do something like what we do. But like, now if I'd say, share it with your team, get in there and build dashboards together, like derive insights together like build some data agents. You could, but you'd have to do it by hand and it's not engineered for that. There's 35 other things in the UI that's going to distract your team and like not make collaboration easy.
00;23;45;15 - 00;24;08;00
Unknown
So we're like, if we just purely focus on that piece and on that vertical part of the process, we can actually then go out horizontally to different use cases. But like first we've just got to crush this like data experience, you know. Yeah. No, I in fact I, I'd be curious if you could just stay there. I mean I feels like we said Jim Barksdale quote, it's like business is just bundling and unbundling, you know, and it just that's the way the world works.
00;24;08;00 - 00;24;28;05
Unknown
I feel like we're having all of that at once right now. Yeah. You have companies trying to bundle while other companies try. I'm going to focus on a thing. I think ultimately, it does it does allow a user today AI to, to leverage specialized products. Right. And in a way that they may be felt like they couldn't.
00;24;28;05 - 00;24;50;06
Unknown
I mean, going back to how we start this conversation, and I think I think as human beings, we will get limited in our capacity to like, think through what are the individual things that we want to focus on, in that moment. Maybe it's a day, maybe it's a week, whatever the unit is, and therefore, like a generalist platform, will not be able to accomplish everything that we want to get done in that moment.
00;24;50;08 - 00;25;10;13
Unknown
Right. And add to the fact that we want short term satisfaction, we want things to be done immediately, blah, blah, blah. You're just not going to get there with the clock or with, OpenAI or whatever. Yeah, there's a lot of psychology around it, right? Like, and it's not a new pattern. Like, we can keep saying that nothing's new, but like, think about, like Apple with iOS, you know, like, why do they eventually create the App Store?
00;25;10;13 - 00;25;33;17
Unknown
They weren't going to at first, but then they like they opened it up and they never wanted to give full access for integration into blank. Like it could be like the fingerprint ID and the face ID now it's like the AI model and all the things. But then they started to realize, like, the whole thing's better if you can swap out or give access to the guts of the SDK and then keep people in your ecosystem and platform.
00;25;33;19 - 00;25;49;16
Unknown
So I think we're going to see people of like very accepting of integrating to things and looking for specialist. And we want to be a specialist. So like I do think that well, first of all, we all have AI fatigue already. I know you do like I do I every day I'm testing everything and it's just like, man.
00;25;49;19 - 00;26;06;08
Unknown
And I'm like, all right, now I'm going to focus on work. And then I open up Twitter like I always call it Twitter stuff, but like, and I'm like, now, like literally, yesterday, Curser came out with an update where they brought in canvases, which is basically like cloud artifacts, and their demo was creating a dashboard.
00;26;06;10 - 00;26;23;17
Unknown
And I first thought I was like, have they replaced us? And then it's like, no, they're actually just ephemeral little canvases that you can go do something, but it's not like the point of the product. Right? So like if we integrate into that, the that's immediately where my brain goes, like I want cursor to feature us for proper integration.
00;26;23;17 - 00;26;42;13
Unknown
And now that M.c.p.s have MXGp apps so they can render a UI that we control. You know I want that. So like if you just don't want to use our product again, you're gonna hear me say it over and over, please don't don't log into our UI, just integrate our product into wherever you want to be. Because I think that's where we're getting, that's the personalized nature of AI.
00;26;42;16 - 00;27;03;11
Unknown
You've got a million choices now of integrations. Integrations are easier than ever. Use. The thing of the platform are the experience that you want yourself. I still like email, and I know I'm weird about that. Like, everybody wants to like, I was going to say most, most CTAs I've talked to have been like, you know, our customer base, you know, where they like to work, email, they like to receive email, they like to send email.
00;27;03;11 - 00;27;21;26
Unknown
They like to read email. Like in fact, our UI should just be email. Dude. Oh my God. The very first precursor to Dream and Dream Base was a product about email for that reason, because we were like, every phone on the planet supports it. Every you want have to build the client is already the email client they want.
00;27;21;28 - 00;27;43;10
Unknown
That's the thing that people look at every night and every day, right? So one of the things we're just releasing actually next week for Dream Basis Email Insights. So we're automating the data refresh of your dashboards. Again the dashboard was what was the sneaky reason for you to build the data pipeline. Now that we have your data pipeline refreshing, we can actually deliver insights based off that.
00;27;43;13 - 00;28;03;03
Unknown
So you wake up in the morning sipping coffee at 9 a.m. and saying, oh, look at these stats like these things are up. But more importantly, these things are down or these things are flat. Why? And then you take action. Now it might be viewing a dashboard, but it might be kicking off some other data workflow. Right. And in cloud code, since you're an engineer now.
00;28;03;06 - 00;28;20;13
Unknown
Yeah, that's what I meant. So I something that I've been struggling with is like you, it's it takes so much time and so much energy to, to keep up with everything. To your point. You start getting involved. You start looking at Twitter. You start looking at whatever medium, whatever article, and then you're like, oh shit, I got to do my job right?
00;28;20;13 - 00;28;40;19
Unknown
And you start working. At what point, like, what do you do to detach to like, not to be able to, like, remove everything or digest everything, and have a more holistic vision of of what to do with dream based, what to do with your life, like touch growth. Yeah, yeah, man. I, when I find out, I'll let you know.
00;28;40;21 - 00;28;58;11
Unknown
We were just talking about how for people that have some, some version of, you know, lack of focus or ADHD or whatever you want to call it, it's this is your your way to finally truly realize that, like, I've always had it, I've always done ten things. And I had a problem with focusing, and I like knocking a lot of things out.
00;28;58;13 - 00;29;15;23
Unknown
It makes me look productive, but it's actually probably a negative sometimes. So if you can see my crazy monitor with all my agents working on all the things I have complete lack of focus, I think there are times in the day and thank God my wife comes down with our miniature docs and makes me take a walk with her.
00;29;15;25 - 00;29;34;27
Unknown
Since we live in Galveston and get some some beach air. But I like walking. I get to mentally unplug and process and reset my context a little bit because humans have context too. And it's almost like when cloud code has to either compress or clear the context window. I got to decide what's the crap that I was just doing.
00;29;34;27 - 00;29;50;11
Unknown
That doesn't matter right now, right? Oh my God. Like, there's so many things that I go down the rabbit hole of building a new tool that the team might need myself. I was building a tool for background patterns, and I was like, somebody is building this right now. And then paper came out with shaders and I was like, Thank God somebody else can do it.
00;29;50;11 - 00;30;08;18
Unknown
So I think it's that it's like stepping back for a second and unplugging as much as you can, even though none of us can truly unplug now and saying, like, all right, now, end of the day, it's already 2 p.m. what are the things I have to finish today? And for me, that always gets down to like the last two things.
00;30;08;18 - 00;30;24;27
Unknown
And my goal is to get off the computer by seven so I can cook dinner and again unplug until I look at my phone and my background agents are running again and they're reporting to me. So it's like, this is a real problem we're going to have. I think we're going to reach a max. Yeah, amount of stuff that we can process.
00;30;24;29 - 00;30;42;06
Unknown
I what worries me is like, what do you do with this next generation? I mean, you touched on is like the most important thing now is critical thinking is the ability to like, digest data and be like, well, this is this is what I need to push back on. This is what I check on. This makes sense for someone that's just coming out and does have little life experience.
00;30;42;06 - 00;31;11;00
Unknown
And there's an expectation now right on a younger group to be able to do more with less, execute, execute. I, I, I wonder a lot around how a younger generation builds critical thinking skills today. I don't know. Yeah, I know you. You want to believe that like civilizations have always gone through something like we're going through their version of this, whatever the revolution was, the industrial revolution, whatever, like and we will make this out of this hopefully as humanity.
00;31;11;02 - 00;31;33;06
Unknown
I'm hopeful when I see, we don't have kids, but we've got nieces and nephews and they're in college now and graduating. And, you know, you have these fears, but at the same time, they seem to be working it out, and they're way more responsible than we were at that age. First of all, and it seems to be more ethical and like, have more of a sense of morals in, like, managing their time.
00;31;33;09 - 00;31;59;24
Unknown
So I'm hopeful that that they can actually process that since they're raised with it. I would I was the last, year of Gen X and, we had the privilege and, you say that like there's a cutoff and it just absolutely is. It flips the next year. Oh, damn. Millennials. Oh. We got the privilege of seeing the change like the MTV generation got to see the change of cable TV and the introduction of the internet and mobile phones and all these things.
00;31;59;24 - 00;32;23;18
Unknown
Right? So you can see both sides of it. And I think that's also why it's tricky for us, because we can see before and after. I think if you're born with it as a natural thing, you're going to process differently. You know, it's like babies were able to use iPads when they first came out. And it's if you if you take that and now extrapolate like information's at your fingertips more than ever answers and like doing things is at your fingertips.
00;32;23;24 - 00;32;42;21
Unknown
What do you care about? It's actually people talk about taste, which is always kind of a funny way to frame it for me, it's it is it's, if you've got vision and clarity and communication and creativity, I think it's gonna be a fun time. I think of this already as like, since the internet started, that was fun.
00;32;42;24 - 00;33;02;01
Unknown
When web 1.0 and 2.0 happened and we were building social networks and doing all kinds of fun stuff back then, it felt like anything was possible, and now it feels like again, I was getting really bored with tech. Honestly, if I'm going to be fully honest, and this all happened, I was like, man, this feels like the early days of the internet and we can just build whatever we want.
00;33;02;02 - 00;33;22;00
Unknown
Now I actually challenge us to be more creative. Let's do Winamp again. Like let's, let's do like non-boring looking processes and interfaces and get really creative. So I, I think the kids will be all right, man. What what was causing the boredom. It just got very repetitious to me. You know, like we wanted to do bigger things. We wanted things to be easier.
00;33;22;00 - 00;33;37;12
Unknown
So like, natural language came out and I was trying to do NLP interfaces with data, and it was so clunky and horrible. And I was like, well, this is never going to work, so we can't have it. Every time I design some kind of interface that had like speech as the forefront, it was crappy. So we were like, that sucks.
00;33;37;12 - 00;34;01;26
Unknown
Let's put that on the shelf. Let's go back to the same old thing that we've built a thousand time, like content management systems and CRM and whatever it might be, it just felt like it was templated, you know? And this feels like it breaks the mold. And this is why I think younger people might have the advantage. Like, if you can do anything from when you're young, you can do stuff that are never going to like you and me would probably never imagine what they're going to build, right?
00;34;01;29 - 00;34;20;23
Unknown
Yeah, well, I think I think you touched on something that was super interesting, which is this impact piece, right? Where you were saying like, oh, they're more ethical, they're drinking less alcohol or maybe less Catholic, all that kind of stuff. And they, and so they have maybe time to think about what is the impact that I want to have in the world.
00;34;20;23 - 00;34;37;29
Unknown
You know, I mean, it took me until I was 30 before I was like, like, felt like I woke up and my my first daughter was born in maybe overly romantic. I remember thinking like this. I can't just be doing this. Like I can just work in oil and gas and cell pipe. Like what? You know, what is the meaning?
00;34;38;01 - 00;34;54;10
Unknown
Yeah. And, and so to your point, maybe, maybe this allows a younger generation to, like, start thinking about that a lot sooner, start plotting data points. About what? Yeah. What is the impact they want to have. And now they've got all the tools to to not just like execute, but also digest all this information that's out there.
00;34;54;14 - 00;35;11;13
Unknown
Yeah. To be able to drive that, I, I hadn't thought about it that way, but that's, that's super exciting. I mean, you hit. Yeah. Now you just have to, to choose what's important to you. Right. Since you can do virtually anything, it does make it a little bit more tough. Like it's overwhelming, I think, to think that you could do anything and everybody's going through this building stuff with AI.
00;35;11;15 - 00;35;33;23
Unknown
We've got so many friends that are building a new tool every day or a new thing, and I actually feel privileged that I'm so consumed by dream based. Like I can just focus on this thing, because if I didn't have one thing that we were focused and enjoying and having a lot of fun building together, like I, I'd be out there peddling a million side projects and like, I'm already frantic with the amount of I had taken.
00;35;33;23 - 00;35;55;17
Unknown
So like, I like that I have some level of focus and clarity and also like I tried to not do analytics for a startup first, like we were going to do AI generated playing cards and just something different. And eventually we got pulled back into the thing from our careers. But it's like it also feels good, like you need a little direction in life and like you hope to find purpose that aligns with that direction.
00;35;55;17 - 00;36;15;22
Unknown
I think that's the ultimate gift. And like if you happen to enjoy that thing and you can like really focus on it, that's what every kid should be trying to learn from elementary school now. Like, I like video games. When I was younger and computer graphics was this new thing. So I was in that first graduating class of 3D animation in Louisiana because it was the only choice for me.
00;36;15;22 - 00;36;30;20
Unknown
Right? And then I couldn't get a job in Austin in 3D. So I started building websites. But like every dot was connected, looking back and I hope and I think that's what parents should focus on with their kids. Just like help your kids find the thing that they love and they want to do to make a dent in this world.
00;36;30;20 - 00;36;55;29
Unknown
And, you know, because you can do anything. But like, if you think you can do anything, you can do nothing. If you're just kind of staring out the oblivion. Sorry, I didn't mean oh, no, that was awesome. I, I think you're right. I, I love my parents. My parents were amazing. But their doctors and I didn't want to go down the medical path, and and so I never, you know, I never got the chance to really, like, look or not.
00;36;55;29 - 00;37;18;28
Unknown
Not to. I never gave myself the chance or forced myself to look about, you know, what? What is it that I get excited about? It was like. Or it was too rudimentary to be like, you know what, I like math, so I'll do something engineering, right? Know? And then it was like, well, I'm an engineer and I'm in Houston, so I'll just do some oil and gas, you know, and then, you know, it took a long time to be able to realize, like, you know, I can go out and explore and try things and see how it makes me feel.
00;37;19;02 - 00;37;38;12
Unknown
And to your point, plot data points. And then because you can't connect data points that you don't see. Right. And, and so I, I, I love the sentiment and, and it's exciting. I guess the flip side of it, besides, like your agents pestering you about how great they are, like, what keeps you up at night?
00;37;38;15 - 00;37;53;16
Unknown
What keeps you up at night? Oh, man. I, I am happy to say that right now, I feel like we're very much in the honeymoon phase of the startup. It's Andy and I. It's our first startup, so we're getting to, like, discover all these new things together for the first time. And we have this team of awesome people that we're building.
00;37;53;16 - 00;38;16;02
Unknown
So it's like, I look forward to that every day. The things that keep me up aren't existential for our business so much. And that, like, I have every confidence we're going to figure out landing the plane for AI native analytics that people adopt for things, all that kind of stuff. What keeps me up is just thinking about like, moving quickly enough.
00;38;16;04 - 00;38;34;26
Unknown
And this is again we're in a bubble right. Like we're all in our own bubble. And on if I go on Twitter I'm seeing people like Karpathy and Pete Steinberg and all like the, the ones that have risen to the top in terms of visibility. I'm like, man, how are they getting out in front of this so fast?
00;38;34;26 - 00;38;51;19
Unknown
You know, Boris, that made clog code. Every time he talks about something he says like he has a simple set up and I'm like, that is not simple. Like, that's holy, Holy crap, that's advanced. So I'm like, that need to stay out in front, which is probably backed by a little bit of insecurity. It's that's what drives me.
00;38;51;19 - 00;39;10;10
Unknown
And, so, like, I'm looking at my iPad, a lot of people don't use iPads, which that's my my bedroom device for, like, learning. And I'm just laying, swiping through stuff like imagining, like how we could use this tomorrow. You know, and I'm jotting down notes or you shouldn't do this to your team, but I'm like, posting the slack relate about like, we should we should read this tomorrow.
00;39;10;10 - 00;39;28;12
Unknown
We should do this. So like that guy on the team. Yeah. And and like I'm not the only one. Which I'm glad everybody else does it too. But like, it's it's impossible to not be like that. And and actually, you're not, not going to make it as everybody says. It's like you got to stay out in front now.
00;39;28;16 - 00;39;49;14
Unknown
Yeah. I also have to cut through the noise and say, like, what's the most important thing for us to focus on right now? So that's CTO of a team. For me, that's the hard job that I try to do because I'm the excited guy. But I also want to be like, all right, look, we have to ship email on sites this week like there's a thousand other things we can do with our agent tech platform, but for value for our users, let's just go ship.
00;39;49;14 - 00;40;07;06
Unknown
One thing that really matters to them, and then we'll get back to the cool stuff, you know? Yeah. Well, that's the critical thinking that you're talking about. I mean, that's that's what you bring to the table. What is more, maybe one more granular question. Where do you find signal in your ingestion of of materials, of outside things that people are creating.
00;40;07;08 - 00;40;26;23
Unknown
What's your sources of of news? Man, I wish it wasn't social media so much. I wish I had, such a more of a profound answer than that. But like the one good thing about the algorithms now, social media, is that you get targeted with the things that you want to see. So that's why I'm consuming at night time to read about what we should be doing tomorrow.
00;40;26;23 - 00;40;43;20
Unknown
So for me, it's always searching for patterns, which again, I is really good at searching for patterns that all my little AI assistants are also searching for patterns for me and bringing them to me. But I just want to see where we're going. You know, I feel like I and we, me and Andy and then the team early.
00;40;43;22 - 00;41;00;05
Unknown
Ed is a guy that works with us, that he's. I've worked for the last decade across companies, and it's like we're passionate about the same things, and we feel like we can call the ball a little bit because, like, if we see the pattern of what we knew with our career and the new way to do that, how do they kind of converge?
00;41;00;08 - 00;41;16;15
Unknown
So like when I'm consuming and I see everybody talking about skills, for example, I'm just like, okay, skills aren't just like another word for context. There's actually more two skills. How can we steer our ship towards that to get out in front of it and meet it where like the timing is right? Because timing is everything for startups, right?
00;41;16;17 - 00;41;35;19
Unknown
So it's like that's when I'm reading, I'm trying to read something that that's just a bookmark for us to go to when it's time. And, you know, I'll be fully honest. Like, I usually like a lot of people don't understand what these things are at first. You see everybody talking about skills in the same way that everybody talking about agent harnesses with open claw.
00;41;35;19 - 00;41;53;10
Unknown
And I was like, what the hell is that? Just a stupid name for something that already exists or is that a new thing? And it's like trying to read from those sources and understand them a little bit more intimately and predicting how we might use it, but don't force us to use. It's a hard thing, too, because now that we can build everything, it's almost way, it's way harder to say what we shouldn't build.
00;41;53;12 - 00;42;14;11
Unknown
And I'm very guilty of this. Like Andy is way better, my co-founder, about saying, nah, man. Like simple. Keep it simple. Like we don't need more features. And I was like, but we can build everything. Yeah, well, but this, this comes to like one of my favorite questions, especially at this stage as you guys are at, which is like, you're ingesting all this, you're getting customer feedback, you're doing all this.
00;42;14;13 - 00;42;34;09
Unknown
How do you delineate between like, you know, condensing and synthesizing everything a customer wants and building a faster horse versus, hey, I think this is the way the world is going, so let's go build a car. That was something we went through. We had a hard conversation. Our big feature that we released a few months back. And time moves are differently now, right?
00;42;34;09 - 00;42;50;01
Unknown
It feels like it was years ago, but we came up when we came out with dashboards. That was our first massive feature. We had a previous version of that which was like a formal reports that would kind of disappear after you would do them. And we were like, reports. Yeah, basically you're right. Which they're still kind of neat, right?
00;42;50;01 - 00;43;10;21
Unknown
There's still there's still a use for that. But we were like always like, man, we got to get dashboards done. We got to hire people to build dashboards. And they've got to be like updating live dashboards, everything you expect. But then we had this conversation about like, okay, where do we draw the line in the sand of recreating Tableau or like, let's not go build every dashboard feature that our users think they want.
00;43;10;24 - 00;43;29;24
Unknown
We have another one that's actually a great customer that came from power BI and wanted all the filtering and all the things that power BI does, and like we did a lot of them. And I was like, guys, can we, can we focus on some new stuff like that doesn't exist right now? And I was like, I don't want them to be new, shiny things, but we are going to get mired down in dashboards for the rest of our career, and we're going to go out of business.
00;43;29;27 - 00;43;45;14
Unknown
I was like, how can we focus on then? This is what turned into the agents, the CLI, because this was right when people were talking about the CLI again. So I was like, this is a new thing. And it's not just a trend. It's so agents can use our product and that's like exponential growth for us beyond people wanting dashboards.
00;43;45;21 - 00;44;00;04
Unknown
Yeah, yeah. Plus honestly nobody wants to dashboard. Let's be real honest. Like you don't want to open up a dashboard every day and have to try to figure out the data yourself. But you do want something. You want your dashboard, something to talk to you about it, right? Like you want it to surface. The things that you didn't think about.
00;44;00;07 - 00;44;20;02
Unknown
It's it's visual confirmation, right. And, I think that it's for us, it's like I can look at an insight and then go confirm it in the dashboard and maybe even look at the query that powers the dashboard if I need to. But that's a secondary thing to me, right? Is that I love dashboards. And look, I've designed them so many dashboards in the last 20 years.
00;44;20;05 - 00;44;34;07
Unknown
But it's like there's only so many I can look at. And I'm just like, just tell me what you want me to do. That's what I want from I. If I were just to say, like what we want Dream Base to become, it's like, look, you're connected to all the data in my business. Don't just tell me historically what happened yesterday like a dashboard.
00;44;34;07 - 00;44;50;27
Unknown
Tell me what? Like I need to do for tomorrow. And it's like, do it in such a way where you can prove it with data. And if that means clicking on a link to view a dashboard, fine. Right. But like let's I don't want to start with a dashboard. Yeah. Yeah. I mean almost because maybe even a sale internal sales tool.
00;44;50;27 - 00;45;06;11
Unknown
Right. Yes. You're right. Maybe I don't want to look at a dashboard, but I do know that if I need to convince someone else on my team about something, being able to show, you know, if you just show a bunch of numbers, nobody gives a shit. But if you show them in a way that makes sense and provides an insight, it makes it more real, right?
00;45;06;12 - 00;45;23;24
Unknown
Seeing is believing it. So, yeah. I, I do push back on a little bit on whether people want to see dashboards or not. I think maybe because I come from this oil and gas real estate legacy industry space where people frickin love dashboards, you know, but that is something that, like, keeps me up at night right now.
00;45;23;24 - 00;45;49;05
Unknown
It's like how legacy industries adopt AI, native technology. And I think what's interesting about the way Dreambox is being built is that it's, you know, the the distribution is through the super based marketplace and stripe and other other places. And so maybe I'm making a poor assumption, but I feel like it's a very technical user that's going to engage with dream based first before eventually.
00;45;49;09 - 00;46;07;03
Unknown
Okay. So this is I mean, like you can't you can't rest on any assumptions anymore. I think like that would be an easy assumption if developers were the only people using super base, which was very true for a while. Right? When I was first adopting it, they're only five years old, but they have that crazy chart that just got updated.
00;46;07;03 - 00;46;25;13
Unknown
They're nearing 8 million now. But, I guarantee you that 8 million is not all developers. This is people like everybody from vibe coder from lovable or V0 bolt that get a super base under the hood is people like me that were kind of developers that stumbled upon it because it was real easy to stand up with versal.
00;46;25;16 - 00;46;42;02
Unknown
So like, there's all these reasons for it, for getting to it. And I think that's the tip of the iceberg. I think there's like a large majority that we can't assume. And look, I'm a design person and I'm a user experience person at heart, and I would love nothing else. And to say that our only persona using our product is a developer, because then I wouldn't know what to build.
00;46;42;04 - 00;46;56;28
Unknown
We have such a hard time designing our product because you can't assume that, right? So yes, like for developers who are arguably the people that set up the super base that arguably might know the names of their tables, we need to make it a good experience, and they're not going to trust us if we don't show our work.
00;46;56;28 - 00;47;13;10
Unknown
So we do a lot of work of like surfacing queries and issues with their performance and all that kind of stuff, showing them the data set. We have buttons to click that, open it, and super Base where you get to query like in Super Base to call BS in our dashboard. Speaking of like validating things with dashboards, I think we get them.
00;47;13;10 - 00;47;33;08
Unknown
We are them. But then there's ones in our team that are less technical and like there's people out there that are actually much less tech, more user than our product. So for them they're like, just make it easier, right? So how do I do both at the same time? It's got to be the simplest, easy thing that's got the most validation and verification under the hood as you can possibly need.
00;47;33;09 - 00;47;55;07
Unknown
It's just progressive like disclosure thing we have to design. But then now imagine an AI agent using this, not that user. Now what do we expose? Like what's our transparency and trust via that. Right. So the thing that I think that we'll get to that's beyond dashboards, I still visual. I've always loved infographics love love love infographics. And it's always hard.
00;47;55;10 - 00;48;16;27
Unknown
You can't build a dashboard like an infographic because it's like it's designed. Now that I can also design and build SVGs and all these things like part of the thing I want to build is an infographic that really describes all pieces of the puzzle that it put together, that no dashboard could ever do. That also shows the data so you can trust it, but it's very digestible.
00;48;16;29 - 00;48;37;12
Unknown
And of course it can be interactive. Right. And like, let me drill down on that, whatever that looks like, you know, like turn it into a new infographic that morphs and tells me the story of like, like, oh, that's HubSpot data and that stripe data. And then it landed in super base. But tell me how they connected. And also tell me, like from a CRM perspective, what I need to know and not tell me from a sales perspective and rep ops, what do I need to know?
00;48;37;14 - 00;48;56;19
Unknown
So I think that it's going to get to this like polymorphic, 3D maybe 3D might actually happen. We always wanted like 3D visuals and they're so cheesy. Like, I'm sure you're seeing a whole gas, right? Like, let me see the bar chart. But like in 3D, like wave my hands, it'll move around. Yeah. Let me explore the drilling platform in 3D, but like so.
00;48;56;19 - 00;49;18;16
Unknown
So I think we'll actually had to get really creative on what to do next. And, and then also now I would describe that in a language that an AI can understand and interpret, not just a human. Right. Yeah. Yeah. Well there we go. Back to the bundling. Right? Like you're while you're unbundling the use case. Right. And you're focusing on, on just this data management vertical, you also have to think about bundling all the different users that would be using that.
00;49;18;19 - 00;49;38;11
Unknown
Yeah. Because it's not just the technologists anymore. Like everybody from the CEO to the accountant to the, you know, sales person, all want to be able to engage with the data. Yeah, in a way that provides insights to make them better at their job, all the while making sure that the permissions are right. And then to your point, like, we haven't even talked about the agents that get built by these people and how they engage with the data.
00;49;38;11 - 00;49;56;05
Unknown
That's it's a massive problem. It is the good thing is that, like, again, like agents are modeled after humans. So I think as we build out more of these agents, they're like some of them are dumber than other ones, like the whole Ralph Lupe thing. Like sometimes there's a dumb agent and like, how does it interpret the data that was handed off from the data engineer agent?
00;49;56;07 - 00;50;15;07
Unknown
So I think that like some of these things, honestly, we we will expose those to the humans that are equipped to them. So if you're a data engineer using our platform, we're not trying to replace you. We're just trying to make your world so much easier. Right. So a data engineer can go talk to the actual data engineer agent to do all the menial tasks that they don't want to do, all the DevOps kind of stuff.
00;50;15;07 - 00;50;34;25
Unknown
Right? Like go clean up our data, go wrangle the stuff. I don't want to do it. Just do it. If you're more of the insights person, then speak like insights, right? And like talk to the insights agent. So this is the agents was the biggest gift for for designing software in this generation of AI I think. And again what's next.
00;50;34;27 - 00;50;54;29
Unknown
Hard to see right now. Like we're so caught up in this current area of skills and agents. But there's going to be something next. What's, what's your most contrarian point? That you die on a hill for, man, that's that's a hell of a question. I feel like I should have so many of those at the tip of my tongue.
00;50;55;01 - 00;51;14;24
Unknown
I mean, well, honestly, I, there's people. I have to look at a camera. There's people that would know what I'm talking about and say, like, I like to be right about things and especially, like, calling the ball where it's going to go. And, a few months back and, like, maybe longer, I was saying command line interfaces are going to be the default interfaces.
00;51;14;24 - 00;51;31;25
Unknown
So we have to be able to do like analytics and reporting via command line. Now there's actually really cool command line tools to do charts and stuff in the command line. And then it turned into I consuming command line. So I feel like that was a contrarian thing that became true. Yeah, I was like, CLI is over dashboards for sure.
00;51;31;27 - 00;51;47;12
Unknown
I, I think that's so interesting because like, I remember having a conversation a few months ago where with with an AI engineer and he was like, you know, like CLI is important, but like, are you telling me in a 40,000 person company, like, everyone's going to be using the CLI to, to be able to engage with it with an agent?
00;51;47;15 - 00;52;07;13
Unknown
But I think, I think you, you created the right caveat, which is, no, it may not be that person, but it's definitely the agent that that person created. Yeah, 100%. So like and that's we're going to see that that's just starting right. Like now there's whole management softwares for the CLI and the agents because you can only look at so many tabs open at the same time and cloud code.
00;52;07;16 - 00;52;30;13
Unknown
But I think that CLI is a good contrarian one that that came to fruition. I also still stand on like dashboards and just analytics in general. Like a lot of people create dashboards and data to tell a lie. You can you can manipulate data to say whatever you want. The numbers dance however you need. Yeah. So I think that like in this is not a new problem, but it's a problem I've seen over the last 15 years.
00;52;30;13 - 00;52;47;21
Unknown
It's like we need something to get to the source of truth that's like that's believable. And and analytics lie a lot. So I think that like one of the things we have to figure out is, is how to make that truthful and actually accurate, like not just the sales number. Right? But like the story is trying to tell.
00;52;47;23 - 00;53;15;20
Unknown
And I tries to please and this is not necessarily contrarian, but like we got to figure out AI that honestly is realistic. So there was a version of called or ChatGPT that came out where it was so pedantic and just like agree with everything. And they had the real back. But like, we see that with our insights, our agents would be like, these metrics are doing this, therefore sales and your store would be better if you change your shopping cart.
00;53;15;20 - 00;53;32;07
Unknown
And I'd be like, we don't even have a shopping cart like you just made that last part up. So it's like a lot of it is making sure the insights are real, which is the hard nut to crack, which is why we're kind of going towards calling ourselves AI Native Intelligence instead of just AI native analytics. Feel like what you want, the intelligence like.
00;53;32;07 - 00;53;52;00
Unknown
That part's the part you have to nail for the real future version. And honestly, as these models get smarter at the core, it's more possible, but it's still going to require a focused product like us to actually deliver that, I think. So that's sorry. That's a long way. The way I know there's country and beliefs that led to what I'm saying.
00;53;52;03 - 00;54;14;20
Unknown
But the fun part is seeing contrarian beliefs become true with AI. You know, I honestly now I think that's a good creative exercise. Yeah. We should sit around and just be contrarians just to be proven wrong in the next six months because it's probably happening. Well, maybe to the point you're making. Like, do you feel like a contrarian almost changes, meaning because you can validate anything now with, with whatever data you've got.
00;54;14;27 - 00;54;30;12
Unknown
That's man, you sound like my dad. We go home and argue about politics and like that. Yeah, that's my AI, cause I was just telling a story he wants said you're not entirely wrong, man. Like, it's because now that they have context of what you like to hear for sure. Right? It's like, that's a great idea. You're right.
00;54;30;12 - 00;54;47;13
Unknown
Let's do that. Is like, man, that's a dumb idea. Oh, dude, I we get in a fight with my wife, recently, she, like, you know, she's got three young kids. It's a difficult season. She's been amazing. And and she's working two, three jobs, like, just, you know, little things. Odds and ends. Sounds like a hero. And.
00;54;47;13 - 00;55;10;21
Unknown
Yeah. And and and so she's struggling with, like, like the stresses of life in this season. And she, like, I don't have, I don't have social media because I, I will get stuck on it for hours. And so she shows me an Instagram video and it's like, you know, this guy is older gentlemen, like, on a panel talking about the struggles his wife went through with their three young kids.
00;55;10;23 - 00;55;29;15
Unknown
And it's AI generated. Right. Okay. And she's showing me she's like, see, this is how I feel. And I was like, well, yeah, duh. Like that is AI generated ad for you based on what you say, what you look for, what you do, you know, and and and she's and she got frustrated. She's like you just like you don't you don't like you dismiss my feelings.
00;55;29;15 - 00;55;48;24
Unknown
I'm like, no, no, no. Like, I get it. I just don't know what what you want to accomplish with this video. Like you showing me a video that's AI generated based on everything you've already told me. What do you want me to do with that? You know, like, okay, like, this is terrifying. Yeah. And but for her, it was like, oh, my gosh, they get me, you know?
00;55;48;24 - 00;56;07;25
Unknown
And I was like, who is they? Like, yeah. To me it's just AI. It's just an algorithm that you fed with your data, you know. Yeah. Yeah. And yeah. So it was like it was a very real emotional argument with the fact that, like, they just ingested everything that she's fed it to be able to reproduce something that was emotionally triggering for her.
00;56;07;27 - 00;56;24;12
Unknown
Right. And, and I, I, I yeah, I after we settled down and whatever it was like man that's scary. That's that's terrifying. Yeah. Like because it's the pro and con of everything with the AI, right. Like it's so cool that when I using these products, they know about who I am and like what I've talked about with it.
00;56;24;12 - 00;56;39;21
Unknown
So if I ask it to help me write a blog for a dream base for a new feature, you can write it with in My voice, and it knows our entire history about what we've done. So like at this point, like most people, I don't write anything by myself anymore, right? It's always assisted by AI in some way.
00;56;39;23 - 00;56;56;16
Unknown
But the flip side of that is exactly what you're saying. Like stuff can be generated just for you that you don't realize is being generated just for you, to get you to react in some way. Like that's and that's freaking terrifying, right? Like, especially when you're not aware, again, we're the less than 1% of the world that are like knee deep in it.
00;56;56;16 - 00;57;17;11
Unknown
Right? And we still kind of get fooled by it. And as these, seed dance and all these new video models come out and stuff, I'm just like, that's AI, right? And to to the point where I don't trust real stuff anymore, which is the detriment, right? Like now people can use that and really harmful ways. But yeah, to like see a video and you're waiting for something unrealistic to happen with the movement.
00;57;17;11 - 00;57;35;16
Unknown
So you could be like, that was I think. Yeah. Yeah, yeah. No. Like, well it goes back to my question, like how do we train critical thinking? And to your point, it's not just young people, right. Like we need to be constantly training our critical thinking aptitude in a way that I it's never had to be a focal point anymore.
00;57;35;16 - 00;57;51;09
Unknown
Right. Like I think you just experience life and and touch grass and do all the things. And just like as you fix these mistakes and failures, you learn and use that to get better. Yeah. You call like street smarts or so, right? Right. Like now I think you have to be so much more intentional about building critical thinking skills.
00;57;51;11 - 00;58;09;06
Unknown
Yeah. And have wherewithal at the same time, which is rare with people. Right. Like if you go out, walk in public, drive in traffic in Houston, like people don't have self-awareness and wherewithal and critical thinking. They're looking at their phone. Yeah. Yeah. Right. Yeah. Yeah. If you're if you're generating eye diagrams on the way to this interview, you're one of them.
00;58;09;07 - 00;58;29;21
Unknown
You know? Yeah. No, you're not you're not alone there. Yeah. Yeah, man. But it's absolutely terrifying. Yeah, I'm shitty at intros, so I didn't. I just start with questions. But thank you so much. Kyle Ledbetter, CTO of Dream Base. I, do you want it? Like, do you want to actually, like, encapsulate everything about dream Base and kind of close out with that?
00;58;29;21 - 00;58;57;22
Unknown
Yeah for sure. Thanks. And by the way, this is a blast. So Dream Base is really just delivering on the promise of analytics. It's it's AI, native analytics and AI native intelligence. Like I said. But what that really means is just that we sit right on top of your data. We're really easy to set up, and no matter what your technical background is, you can have answers to your questions, insights and very soon proactive and future looking, predictions for what you should do with your data.
00;58;57;25 - 00;59;21;08
Unknown
So we're starting with dashboards now, but as you probably heard in this podcast, it's not just dashboards were a lot more than dashboards. And we really feel like we're paving the future of what analytics becomes. And we're surrounded by amazing, investors, including some of the folks, around here. So I dream base is delivering on the promise of analytics is how I frame it.
00;59;21;10 - 00;59;42;18
Unknown
How for someone that like, thinks of analytics, is this, like amorphous, like almost consulting buzzword? How do you cut through that? A little bit to, like, really help people understand what dream base the value Dream Base has? Yeah, I mean, that's why I honestly Stephen King in on intelligence and insights like those kinds of word. Now it's just like, what do you want to know about your business based on your data?
00;59;42;21 - 00;59;57;26
Unknown
Analytics is just moving the data around and making it usable. Really. And there's lots of other fancy words I could say that are surrounding analytics, but that's what this entire industry was propped up to do get data into place to make it usable, clean it up and transform it so it's usable just so you can have some kind of intelligence.
00;59;57;26 - 01;00;04;19
Unknown
Yeah, you just got to try it out. Yeah, yeah. Thanks so much guys. This is awesome. Cheers. NASA's blast.