They delve into the journeys and insights of distinguished leaders in the development world. In today's episode, Michael engages with  Brian Vallelunga, the visionary CEO of Doppler. Brian shares his compelling journey from early tech innovations to leading multiple startups and eventually founding Doppler, a centralized cloud secret management tool.

Brian emphasizes the importance of making security tools enticing for developers, comparing it to making vegetables taste like candy, to boost productivity. His team’s strategy revolves around seamless integration into developers’ workflows, featuring a VS Code extension and automatic syncing akin to Dropbox, enhancing efficiency and ease of use.

They explore Doppler's competitive edge and how it partners with major cloud resource managers, making two-click integrations effortless. Brian also discusses their customer-centric development approach and the release of enterprise features like two-person approval and config inheritance, designed for complex organizational needs.

Brian's entrepreneurial journey is marked by significant pivots driven by frustration and market demand, rather than strategic planning alone. He shares candid thoughts on the impact of founders and success, cautioning against the allure of celebrity status and emphasizing team contributions.

Join them as they dive into insightful discussions on building developer-friendly security tools, the nuances of secret management, and Brian's perspectives on startup growth and innovation. Discover how Doppler is revolutionizing secret management, one integration at a time.

Michael Berk [00:00:05]:
Welcome back to another episode of Adventures in Machine Learning. I'm one of your hosts, Michael Burke, and I do data engineering and machine learning at Databricks. I'm joined by my amazing, wonderful, and beautiful cohost,

Ben Wilson [00:00:15]:
Ben Wilson. I draft design docs for sandbox safe execution of generating code in Python at the end of race.

Michael Berk [00:00:26]:
Okay. I think I followed that. Cool. Anyway, today we're talking with Brian. He started his career as a web developer. And over the next 6 years, he founded 4 companies. He then joined Uber and quickly became a lead software engineer for the safety team. Then after Uber, he founded his 50th or something company called Doppler where he currently works as CEO.

Michael Berk [00:00:50]:
So kicking it off, seems like you've been in entrepreneurship for a while and potentially enjoy it. So why entrepreneurship and why so much of it?

Brian Vallelunga [00:01:00]:
Good question. I don't know, I think for me it's been much of an instant calling, like I've always known that I've wanted to do something, regardless of the word entrepreneurship, I love the idea of like you have an idea in your head of what you want to what you want to exist in the world and then it's your job to go make it happen, and, like, that process I really love. And entrepreneurship is very much, like, it's a very, very hard game because the game keeps changing, right? Like, a good example, you build the product and the product starts working, all of a sudden now you need to start building out the team. Okay, the team starts working, Now you need to build more layers of the team. You need to build more products to keep growing at the rate needed to keep raising capital and keep moving up in the market, and to stay competitive with your competitors that also start coming out of the woodwork. And so it becomes just like it's a very lively game. Like, the game keeps changing in the the dimensionality of it, and I I just love all parts of that. I it feels like, a three-dimensional press match, and a lot of the rules haven't even been defined in the game.

Brian Vallelunga [00:01:58]:
So you're, like, you're learning the rules while playing it. Great example of that is, like, you build a product that, you think people will love and all of a sudden realizes that the market actually doesn't want it that much. But if you change one thing of your product, all of a sudden it's loved. Right? And there was nothing on paper that said that that was gonna happen or no no blog post you could've read. It it was like you it's a lot of trial and error, and I just love that entire process.

Ben Wilson [00:02:22]:
For you, where do you think that started? Was it, like, early childhood stuff?

Brian Vallelunga [00:02:26]:
Yeah.

Ben Wilson [00:02:26]:
Was that your core personality, or is it something you learned as a professional?

Brian Vallelunga [00:02:31]:
I got very lucky. My dad was an engineer, and we did a lot of engineering stuff when I was a kid. Like, I played chess every single night, and then on the weekends, we just built random shit in the, the garage. And, eventually, I started competing in science fairs, and went all the way up to the the state level science fair, twice. And so we got very, very competitive at science firm. The the art of and the way we did science fair was not like, let's put some Mentos in a Coke bottle and see what happens. It's more like we built devices that could fly with no wings or engines using ionic propulsion. We tried to build a pawn game controlled by your mind.

Brian Vallelunga [00:03:09]:
Like, we wouldn't we were trying to push the limits as much as we could for, like, being middle school, high school, kids and with very low budget says, like, a middle class family. But I just love that. I was like, that was so cool. And there was a moment for me where I think really kicked off the entrepreneurship journey was a little bit before that where I was walking, to school with one of my classmates in elementary school, and we had, these, like, old school, Motorola Razors, if you know what I'm talking about back in the day. And this is the day this is before anyone can, like, watch movies on their phone. It was like that thing was used for, like, calling and texting, and that's about it. Maybe, like, some parental controls on top. And he goes and he's like, why can't I watch a movie on my phone? It's it's stupid that this is impossible.

Brian Vallelunga [00:03:56]:
And it got my mind, like, racing. And what I realized was that you could record on the phone. Right? Like, you there's a camera there, and you can record and you can do playback. And it's like, well, the phone really doesn't care if it's playing back something from a from the camera or from something I tell it to to record. And so what I ended up doing was ripping a DVD that, that night, figuring and then recording a a video from my phone, seeing what the format was, and then changing the video to that format and then uploading it on a micro SD card and and putting it in the phone. And I was able to watch a full blown movie the next day, and showed it to my friend. And at that moment, I was like, okay. You can basically build anything you want at this point.

Brian Vallelunga [00:04:33]:
It's just what do you wanna build? And that's what really got my mind racing of, like, okay. There's so much opportunity out here. We just gotta find the opportunity and see it and then ruthlessly go after it and execute.

Michael Berk [00:04:46]:
Cool. So Doppler provide well, actually, I have seven questions. I'll start with this one. Doppler provides centralized secret management via cloud. Why this application? Was it born out of a real pain point, or is this your true passion?

Brian Vallelunga [00:05:00]:
Both. So I was at the time, I was kinda doing another startup. It was Doppler. It started out as a crypto machine learning marketplace, kinda all the buzzwords in one. And, I guess this is pretty relevant for this audience. And, over time, I realized that it it it was before the AI ML hype, and it very much felt like pushing a boulder up a hill. Like, you move 1 foot forward, slip by feedback from exhaustion. It was just really freaking hard.

Brian Vallelunga [00:05:31]:
And at some point, I just got to my breaking point, like, 8 months in trying to get this thing to work. Turns out it did work. Right? Like, there's, what's it called? There's a an AI model marketplace today that's great. I'm trying to

Ben Wilson [00:05:42]:
I'm Hugging face.

Brian Vallelunga [00:05:44]:
Hugging face. Yes. We were so similar to Hugging Face. I mean, like, even to the design, it was so similar. But we were just way too freaking early. Like, again, going back to the three-dimensional game. Like, we were just too early, and they didn't work. But, yeah, we were working on that, and, it was a it had crypto in it, and then for and then at some point, we actually ripped off the crypto, and it was literally just like hugging face.

Brian Vallelunga [00:06:07]:
And we were at a breaking point, and we decided to take a trip to Cancun and kinda just take a break from it all. And the first day there, I realized that I was never gonna make this project work. I just was not destined for it, which is funny because if I if I stayed in the game, like, 2 years longer, I was actually gonna probably make it work. It was just a timing problem, but I didn't know that at the time. It's kinda hard to predict the future. And so I started looking all the problems I was facing, and managing environment variables and secrets is just one of them. They kept popping up in all these weird interesting ways. And so I come back from Mexico and I go to this dinner, that Stripe is hosting.

Brian Vallelunga [00:06:41]:
I had a bunch of founders at it, and developers called the Stripe Atlas dinners. And there, I basically asked the room, hey. Am I a shitty engineer or a world broken? Like, you tell me. I can't tell if it's a real problem or not. And they were all struggling too. And so that made me realize that, like, this is a real thing. And the answer to the problem is not to go build another security tool. It's to build another tool that developers genuinely love using that also has benefits security benefits.

Brian Vallelunga [00:07:07]:
Kinda like this theme of, like, make vegetables taste like candy. Vegetables being a security, candy being a developer productivity. You want it to taste like candy so that everyone uses it. And that I felt, like, uniquely qualified for and got me really excited. I was like, oh, man. I I feel like I'm seeing the matrix here. All of our competitors are building this, like, ugly, hard to use security product that, like, everyone hates using. And I see the magical bullet over here or the magical candy that's gonna make people love to use this tool.

Brian Vallelunga [00:07:33]:
And that just, like, excited the hell out of me. And we built the first version in 3 weeks, got our first set of customers on the 4th week with what I call Chipotle sales, where I basically take any friend or anyone who knew my friends, that were developer, and I'd say, you're gonna I'm gonna take you to Chipotle. I'll get you anything you want on the menu, and I'm gonna rant at you for 2 hours until you basically say, yes, you're gonna use a product. And it was I would never recommend it again. It's a terrible strategy, but it works. It was just really expensive and really hard. Lot of failures on that one. But, we got our first set of customers from that, and that's what got us going.

Michael Berk [00:08:09]:
How do you make it taste like candy?

Brian Vallelunga [00:08:12]:
Great question. In short, it comes back to productivity. Developers are extremely sensitive to productivity. If you, like, block their flows, they will, like, actively avoid using your tool. And so the the trick is to actually deeply integrate it into their workflows. So an example would be, we have a Versus code extension, where you can edit your secrets side by side with your code. So you don't need to go to another tool to edit your secrets. It's almost like editing an EU file, but now it's secure.

Brian Vallelunga [00:08:38]:
And the second you click that save button, it will sync it to your entire team, And then all your other team members will now get their applications automatically restart, kinda like a Dropbox level sync, with those new secrets. And so there's no need to go ping your coworkers and be like, hey. Did you add this secret? My code is now relying on it, and I can't run locally until I get that new secret. Those days are behind you because if a developer if someone on your team adds it, you automatically get that update. And it's all happening in the tools that you love.

Ben Wilson [00:09:09]:
How many integrations do you need to build?

Brian Vallelunga [00:09:13]:
Oh, it was a lot. I mean, I think we're at, like, 40 now or something like that, like, deep integration. So it's I wouldn't say it's like in the 1,000. It's, like, 40, and then you need to support every language. Like, actually, one critical learning for us, going back to, like, this is a journey is we've built the first CLI in JavaScript. Mhmm. And we were like, okay. We need a CLI that can work on any language.

Brian Vallelunga [00:09:38]:
But we made the we made JavaScript dependency. So if really what you needed was, you could support any language, but you also needed JavaScript. And so we were like, okay. We gotta rewrite this in Go and make it a binary so it doesn't have that dependency. And then it was like it that was like a big unlock for us. We were able to, like, get a lot more customers through it. I think the CLI supporting every language is, like, really helped. Then we had to build some framework supports.

Brian Vallelunga [00:10:05]:
We had to build some, SDK supports and then a bunch of, like, native integrations. And the one that I find the most interesting is, like, our Kubernetes sync, which is really cool. Like, it'll sync to Kubernetes secrets and then automatically redeploy, their app their deployments with whatever rollout strategy you have. So you get this, like, nice, really clean model of, like, secrets stored in Kubernetes secrets. So if Dialpad was ever down for any weird reason, you you're not you still have, like, secrets to use, the secrets in a secure place, and your applications automatically get those secrets whenever, they change.

Ben Wilson [00:10:42]:
Yeah. I have some questions about that. When you're talking about a service like this that directly competes with native stuff that exists within cloud providers, how challenging is it when you have a customer that's, like, I just want this working in, like, Azure, like, directly in Azure? How does that conversation go when you're talking to a tech giant, and you're like, hey. I've got 37 customers that say they want this integration. Do you wanna play ball?

Michael Berk [00:11:13]:
How does that go?

Brian Vallelunga [00:11:15]:
Yeah. I think, the first thing we do is we try to flip them into thinking the Doppler way. And the Doppler way is that AWS, secrets managers, your QVault, or GCP secrets manager are not competitors to us. They are actually partners. We actually have formal relationships with all of them. And our view is kind of more like more like GitHub. Right? Like GitHub, you pull your code locally, you you you change that code, then you push back up to GitHub, and then when you're ready, you push that code up to AWS. And AWS will store that code and deploy it to your application.

Brian Vallelunga [00:11:46]:
Same exact model. Right? Pull your secrets down, change them what you, to what you need, push them back up, push them into, into AWS, AWS, roll it out to your applications. So AWS Secrets Manager, Azure Key Vault, GCP Secrets Manager, and and all the others are just partners for us. We have direct integrations with all of them when and our job is to keep them in sync. So that's where that that orchestration layer is one of the core value props we provide. So, I would say 90% of the customers that come to us already are using one of the cloud providers, and it's perfect because now the installation is so much simpler. Instead of saying, hey. We're about all the old ways of fetching secrets and put in this new way.

Brian Vallelunga [00:12:24]:
We just say, hey. Take AWS secrets manager, connect Doppler on top, and now anytime it changes in Doppler, it's gonna change in AWS secrets manager. And there's no more installation than that. It's literally, like, 2 clicks in your dot. You have production infrastructure setup.

Ben Wilson [00:12:36]:
Yeah. I I couldn't hear it. Like, a better business proposal, actually, and a better explanation of, like, what like, how you made something into candy. Because as somebody who's had to do those integrations, where you're like, okay. I need to connect to every single cloud provider for doing something. Where I was just working on a prototype, like, a, like, a month ago or so for this similar situation. It's like, okay. How do I do this in Boto and assume I am role authentication for the user and then make this request? Well, what about these edge cases? And how do they actually get this to sync to prod when it's deployed Yeah.

Ben Wilson [00:13:12]:
And make sure that their IAM role is propagated there? It it quickly becomes it's turtles all the way down. The complexity becomes really, really bad. So you can push that onto the user, but that's putting a plate of vegetables in front of somebody

Brian Vallelunga [00:13:26]:
who wants to

Ben Wilson [00:13:27]:
eat a hamburger. And they're like, dude, I don't wanna I don't wanna deal with this.

Brian Vallelunga [00:13:31]:
They want API forming. It's actually it's kinda crazy. So we started off with, like, we're gonna make this, like, like, 2 to 3 button clicks, and that's it. Like, there is, like, a hard line in Doppler of, like, this integration has to be 3 clicks, fire and forget, you set up once you're done. And then when we got feedback was, like, that actually wasn't enough. Because when you start rolling out to, like we have some very, very large organizations that use this, like, organizations that are on the road map to deploying up to, like, a 125,000 engineers on Doppler, like, very large organizations. And they go, well, we don't wanna do that 3 clicks for every single team. That's a lot of teams.

Brian Vallelunga [00:14:08]:
And happy's teams don't even speak English. And so this is where we started adding Terraform supports. So they can basically add it to their Terraform specs, and the integrations will automatically get standard up along with all their projects, environments, configs, the whole DOPLAR model, along with it. And so that has actually become, like, an even more powerful thing of, like, they don't even need they need to do one click. Right? They they copy and paste a couple template code into their Terraform spec files. They click go and all of a sudden that they are now working in dot work with all the orchestration.

Michael Berk [00:14:42]:
Question for both of you on that front. I've seen this with MLflow a lot where it's sometimes challenging to work with new customers, but it's also challenging to work with the biggest customers. And finding value props for sort of novices and the, like, super, super experts, they're they typically have a very fundamental user journey. So how do you guys think about catering to all users? Do you have a a specific user you focus on and then support the others as you can? How do you think about it?

Brian Vallelunga [00:15:13]:
It's a good question.

Ben Wilson [00:15:15]:
I can take a stab at the MLflow side.

Brian Vallelunga [00:15:18]:
Go for it.

Ben Wilson [00:15:19]:
Or just like open source ML stuff. Depends on where your product is in its journey. And at the early stage of it, you're trying to roll something out. You focus on making it as simple as possible and intentionally make future decisions where you know, like, oh, this isn't gonna appeal to this group. But the development effort in order to get that to be out there is so significant

Brian Vallelunga [00:15:47]:
that

Ben Wilson [00:15:48]:
it makes more sense to focus on the basic use cases first in order to just gain that that intuitive knowledge that people will will develop that, hey. This is an easy tool. And if you get that into their head of, like, this is this solves my problem, and it's really cool, and the demo works great, and and everything's functional for the the novice. Then later on, you respond to what requests come in from the big enterprise customers. They're like, hey. We need this layer of security here. They're like, okay. We'll build that.

Ben Wilson [00:16:23]:
But don't like, we don't do stuff like focus on that hard problem at first because it runs a huge risk of you build something, and then nobody nobody uses it. Nobody asked for it. And then the worst case scenario is somebody asking, why did you build this? Yeah. We need something completely different. Like, this doesn't work for us. You know, like So

Michael Berk [00:16:46]:
in summary, it's it's request

Brian Vallelunga [00:16:47]:
dev work.

Michael Berk [00:16:48]:
It's request driven development. That's the one liner, basically, irrespective of the user?

Ben Wilson [00:16:55]:
Pretty much. Cool. I don't know how Brian does it.

Brian Vallelunga [00:17:00]:
For us, it's been a journey. And, like, the way we kinda think about it is, like, going back to actually our pricing plans. We have the developer plan, the team plan, and the enterprise plan. And another way of, like, saying that is, like, we have a plan for developers, we have a plan for teams, and we have a plan for team of teams. Right? And we have started building those up starting with the developer. So before we even have the team plan, we had the developer plan that was free, and DOP was starting free. And we built out and we built a really great offering for that. And once that was solid, we started building the team plan app and the feature set in that team plan.

Brian Vallelunga [00:17:31]:
And now we're kinda, like, in that journey of building enterprise side. That has been the journey. Now when we get we get feature requests all the time from developers, we get it from enterprises, I think what we look for is we look for that common thread of value among all of the customers. And we try to, like, ignore the direct ask, and we look for the underlying problems that they're at they're talking about and seeing how widespread that problem is. And we gotta bucket it. Is this, like, an enterprise only problem? Is this a problem that every developer will face regardless if you're on an enterprise or not? Because, like, even though our champions, which is, like, the DevOps team in our case, are very, very technical people, when they roll it out to all of the engineers at the company, not all of them are that technical. There may be a bunch of junior engineers that operate very similarly or mid level engineers that operate very similar to, like, a a a start up, engineer or, like, a start up intern. Right? And so there's a lot of parallels there.

Brian Vallelunga [00:18:30]:
And so that's why you kinda go back to, like, what are the the problems and how widespread are those problems across our user base? And we try to focus on the deepest problems that have the the highest lever, highest coverage or the deepest problems that will like like, it's either gonna be breadth or depth, basically. You know what I mean? So, like, if you're on the breadth side, it's gonna, focus on, the entire user base, but it's not that deep, but it's, like, deep enough where, like, it's a real pain. Or on the enterprise side, it's so deep that they're willing to they're willing to pay for this. Like, they're willing to, like like, we've had I'll give you an example that's, like, really powerful for, like, our enterprise is that we literally built for enterprise. And I would say in a year from now, if we fast forward, that will eventually be on, like, the lower tier plans. Like, what we we like to take features and move them down the tiers, as we add more features to the enterprise side. But it's been a pull request or change request. And the idea is, like, there's 2 primary use cases.

Brian Vallelunga [00:19:27]:
1 is an enterprise wants to require that there's 2 sets of eyes anytime secrets change, especially in production, just like how you have the same thing for code. So that was, like, a really heavily demanded feature. The other one is they wanna lock down production and staging access, so only the DevOps team has access to it. But the developers saw the ability to get those kind of secrets into production. And so they can create a change request to production even if they can't see the production secrets, and then someone on the DevOps and security team will basically be able to review and approve it. So they can kinda, like, chuck it over the wall to the DevOps and security team. Both require 2 sets of eyes, just as one has, like, an access control model difference. And that was something that was, like, every enterprise is, like, we want this now.

Brian Vallelunga [00:20:12]:
Like, we will churn if you don't provide this to us, kinda, like, level of desire. And so that we listen very deeply on, and we think it actually has so much value that it can definitely go down on the team plan. It's just a matter of time. I think what we're gonna be doing is, like, we're gonna build more policies around it, like, kinda like get it as a policy where, like, if this is a production environment, it has to go through a a pull request. There's no other path. Like, building those kinda, like, policies into it is, like, something we'll do, and then we'll make the base feature set of, like, you can have a approval flow on secrets go down to the team plan. But that's kind of our model. It's, like, either go super deep on, like, an enterprise feature that eventually can, like, ripple slowly down into, like, the lower tiers or go something that's, like, very, like, breath wise.

Brian Vallelunga [00:20:56]:
A good breath example would be config inheritance for us, which is, like, imagine you have a config, which is, like, a list of secrets in in the Dappo world. We now built it like Legos, so you can stack them on top of each other. So you can build a config that is only you has secrets in it that are shared across other projects. So, like, maybe you have one database that's used across the entire company. And that database, that logging, that ports, all that stuff is gonna be the same across every team, every feature. So you can put all those credentials in there, and then you can have every other feature team's projects consume, like, stack on top that config so that you, so that the DevOps team can control that. The DevOps team says, this is what the database URL is. And for production and for local development, everyone gets that update to click the second they click the save button because it ripples out.

Brian Vallelunga [00:21:40]:
Like, that was, like, a feature we launched for everyone, but it was, like, heavily demanded.

Michael Berk [00:21:46]:
Got it. Crystal clear. So so shifting gears a little bit, I was wondering what your experience was like with the Forbes 30 under 30, which you were nominated for in 2021.

Brian Vallelunga [00:22:00]:
Weird experience. Woke up one day to a flood of calls and emails from friends saying you got Forbes 30 under 30, and I was like, okay. Didn't know I was even in the running. And that was I knew it was just, like, so random. And I found out, like, many months later, it was, like, someone y c recommended me, in Y Combinator. But, yeah, I didn't apply for it at all. Didn't really have, like sorry. Actually, I take it back.

Brian Vallelunga [00:22:28]:
Someone reached out to me a couple months before Forbes 13 and 30, and he said, hey. There's, like, some Forbes thing that you may be nominated for. Fill up this, like, application. And that's all I knew. And, like, I was like, I don't know. I don't care, but I'll do it anyways. And I kinda, like, just wrote some stuff in there and moved on really quickly. And then many, many months later, I wake up to these, like, calls and texting about Forbes are gonna do it.

Brian Vallelunga [00:22:47]:
But I had no clue that it was, like, one was for that. I thought, like, we do a lot of work with Forbes. There's, like, a ton of, like, articles they've done for Doppler. So I thought this was just gonna be another one of those, like, articles or bios or something. It it didn't really click in my mind that this was, like, a 30 under 30 thing. But, honestly, I try to tune out for most of that. Like, it's kinda like it's like a buzzy thing that people really care about. But, like, in the in I wanna say almost in the real world, it doesn't matter.

Brian Vallelunga [00:23:15]:
Like, what matters is, like, does is your company successful? Right? Like, I don't care if I get Forbes or 30 or 30. I want Topo to be successful. I want our customers to have real impact. If I fail on both those things, it does not matter if we got Forbes 33. If anything, it's worse. Because now I'm like, shit. I got this amazing metal, and it don't mean crap.

Michael Berk [00:23:35]:
Got it. So you're not cooler than everyone now?

Brian Vallelunga [00:23:39]:
No. I mean no. I mean, I like, for me, it's like you look at it and you're like like, at at least for me, there's, like, a bunch of companies in our y in my y c batch, which is, like, a group of, like, 200, 400 startups, something like that. It's some big number. I forgot how many it was. And it's like you look at that and you see in that batch, some companies died, some companies are zombies, some companies are thriving, and then there's a couple giants. And, like, I look at the giants, and I'm like, man, I wish it could be you guys right now. I know, like, the scale of success because I know what you guys are doing, and we're not doing that.

Brian Vallelunga [00:24:12]:
Like and it's very hard for me to go, oh, yeah. I'm a Forbes 30 under 30 crushin' it kinda guy. When I'm looking at the guys that I was just having, like, cookies and coffee with, or cookies and hot chocolate in my case because I don't drink coffee, during that batch, and now they're, like, 10,000 miles above us in success and impact. And I'm like, it's kinda hard. It's it's quite humbling to be like, oh, yeah. Those 2 guys, we used to just chill with them, and now they're, like, multibillionaires. And their companies are worth, like, 1,000,000,000 and 1,000,000,000 of dollars and have so much impacting there's, like, billboards everywhere you go with their their name on it. It's like it's, it's very hard to be like, oh, yeah.

Brian Vallelunga [00:24:51]:
That Forbes 13th, everything, I'm crushing it.

Michael Berk [00:24:53]:
What are some companies you look up to?

Brian Vallelunga [00:24:58]:
GitHub, big one. I mean, just like and also, like, I know this is gonna be, like, almost cliche to say, but, like, Microsoft, like, the the scale that Microsoft has gone to with how much funding they've raised, Doppler has raised more funding than Microsoft has, and they're enormous. They're like a trillion cap company. Like, that's I don't that's crazy. Like, that's, like, genuinely crazy. Couple companies that I, like, really, really admire, in today's like, for today's generation of companies, Perplexity, it's actually an AIML company. They're basically just like how you have Google for search, they've kinda take it a step further, and they give you the knowledge behind that search so you can ask a question, get an answer. Kinda like what ChatChiucci was supposed to be, but, like, these guys, I feel like, are like that promise delivered.

Brian Vallelunga [00:25:48]:
I really love Perplexity. Like, it's my default search engine. Anytime I have a question, I just go to them. Highly recommend even paying for their pro plan. It's so worth it. But Proplexity is probably my favorite company today right now. There's also another company called Lightmatter, which is rebuilding chips using, photons. So instead of, like, electrons running through, it's photons, and that's gonna lead to significant compute, performance and then, increases and then also a lot less energy consumption, which is gonna be really big.

Brian Vallelunga [00:26:21]:
Because, like, you can imagine, like, every time you see, like, chat gpt3, chat gpt4, like, for us as a consumer, we see just a small little number change, 3 to 4 to 5. But behind the scenes, that's, like, 1,000,000,000 of dollars of GPUs. Right? And those 1,000,000,000 of dollars of GPUs require 1,000,000,000 of dollars of energy. Right? And so it's, like, enormous amounts of energy. They were, like, estimating that by, like, chat g p t 6 or 5, it's gonna require 10% of the entire US power brick. Just think about that for a second. Like, Amazon just got in the business of building nuclear reactors. That's how much energy is required for these, like, models.

Brian Vallelunga [00:26:58]:
And so if you're able to, like, downsize that, right, we're, like, you don't need that much energy anymore to do the same amount of computer, even more compute. Think about how much more offering there is, now that AI can provide because you could just have so much more compute for so much less energy. Like, that that curve also changes quite dramatically. So I'm really a big fan of that too.

Michael Berk [00:27:19]:
Cool. So it sounds like you're very product focused and results focused. Do you care about the process of these companies or the founder's moral character or whatever it might be?

Brian Vallelunga [00:27:36]:
I think it's incredibly hard to have a real moral opinion, like, a strong and informed opinion about these founders. As being one myself, you get told to write a lot of things. You get told to show up in a lot of places. A lot of times, you take a lot of shit from a lot of people that you don't wanna be taking shit from, and sometimes you react negatively to it. It's a hard game being found. I'm not saying it's impossible or that, like, it's an excuse for everything, but I just at the end of the day, I think, like, a founder's job is to create impact, and the best way to judge them is by the impact they create. Fuck all the words they say. What is the real impact? And so that's why define impact? For me, it's about what their products deliver.

Brian Vallelunga [00:28:19]:
That's the biggest thing. So if, like, a founder's job is to create impact through their companies, what are the what what's the impact of the products that those companies create? And for what it's worth, I also like a couple controversial founders. Like, I'd say the most controversial founder I love is Palmer Luckey from Andro. Like, I know, like, half of Silicon Valley hated on him for a while. I've always been his biggest fan. I love what he's doing with with Andro. Like, we need new defense primes. We need them really grow we need them badly.

Brian Vallelunga [00:28:48]:
Yeah. So I don't know. That's, like, my main thesis around, like, liking founders is, like, what is the impact they're creating regardless of all the other shit on top. Like, they're not meant to be celebrities, in my opinion. Like, I think it'd actually be great if, like, most founders weren't even that popular. Like, let the companies be popular. Let the products be popular. The founders should be in the shadows.

Brian Vallelunga [00:29:09]:
Right? The products and company should speak for themselves. Like, I've tried really, really hard with Doppler not to make your words like my my name is strongly attached with Doppler in the sense of, like, Doppler couldn't be successful if it didn't have my name. Like, Doppler should be able to stand on its own as a brand, as a company. It's super freaking important. And I think a a lot of the best companies do too, like Amazon, Bill, Microsoft, all those companies, like, Jeff Bezos was, like, tied to the brand and now not so much. Bill Gates tied to the brand, not so much for Microsoft, because they were great companies and they built great, brands. And so I very much follow the same way. I don't think founders should be celebrities.

Ben Wilson [00:29:50]:
I agree. Could not agree more.

Michael Berk [00:29:54]:
Yeah, Ben. I feel like you very much live that ethos of not being out and about. I feel like you experienced people bother you a lot, like, take up your time. That's my hypothesis.

Ben Wilson [00:30:08]:
I mean, to be somebody who's known for the words that come out of their mouth and not the the products that they choose to use their life to build. I think that's just a sad state of society to for people to sort of gravitate towards that. Like, I personally don't care, who's the head of any company. Do they are they the the captain of the ship that's driving it in a particular direction, telling the navigator where to go and making sure that the crew is is kind of fed and is on station doing what they they need to get done. Sure. They are the visionary. That's their job. But nothing's gonna happen on that ship without the crew, and the crew is what builds the ship and maintains the ship.

Ben Wilson [00:31:01]:
100%. I I've always felt more esteem towards the actual builders. Maybe that's because I'm one myself. I'm a a nerd like that. But I get more excited when I talk to somebody who's built stuff, and founders like yourself, Brian, who are like, yeah. You built something yourself, or many, many times you're doing that and then moved into a role of of advising and and vision. I respect that so much more than somebody who's like, I got out of Harvard Business School. I went into business management.

Ben Wilson [00:31:37]:
I've been an executive. I run a company. I'm famous because of my name and what I do. It's like, yeah, you haven't built anything. You you've been a captain of ships, but, I'm not saying I I think those people are worthless. They they're very, you know, worthwhile human beings. I just don't personally feel an attachment to people like that.

Brian Vallelunga [00:32:00]:
Also, like, maybe controversial take, I don't think the amount of money you've generated should be something that determines your celebrity status. Like, I really don't. Like, I don't really care how rich you are. Like, it's what you and as as you mentioned, Ben, it's like it's not just about, like, what have you done, but it's like, what has your team done? That's like no matter what I say as a CEO, my team's gonna do a whole lot more hours of work that I'm gonna do, like, cumulatively than than I'll ever be able to do. I just don't have that kind of time. Right? Like, they that's that's why teams exist now. Think about, like, Microsoft. Satya says, like, gives a speech about, like, we're gonna do everything in AI and become the next, like, giant of AI.

Brian Vallelunga [00:32:41]:
Think about how many, like, millions of hours get deployed through all their people and teams and stuff like that. Those people are doing the real work. Yeah. Yeah.

Ben Wilson [00:32:51]:
When you start thinking about large companies and large bets that they're making, As that team grows inside in size, it that's a direct representation of the complexity of what they're trying to tackle. You know, you you look at the exact same product coming out of a start up that and compare that to what does that look like in AWS, or what does that look like in in Azure. If you were to look at both code bases for that implementation side by side, you would not be able to understand one of them. The start up one, you'd be like, I grok this. Some problems here. Like, okay. You're that's kinda hacky. But you look at the, you know, the true enterprise grade cloud provider implementation.

Ben Wilson [00:33:34]:
Like, how many millions of lines of code is this? And how many different code bases? How many people contributed to this? That's why this took 6 months to to roll out. Wow. This is complex. It it's because it has to be in order to work in that ecosystem, and yet so many humans involved in building it. That's why I always, like, silently applaud when even one of our competitors to our company build something cool and people like it. I'm just like, good good on you. Like, congratulations on doing this. It is.

Ben Wilson [00:34:08]:
It's like I I'm just sort of cheering for humanity in general. Like, hey. We we as a species built something cool here. I don't care if we're direct competitors with them. And I happen to work in in an organization in the engineering department that a lot of people kinda see that. Like, nobody's like, oh, yeah. Screw them. They built this thing.

Ben Wilson [00:34:28]:
Like, oh, it's more like, that was successful. Let's go build that too so our customers will be able to use that on our platform.

Brian Vallelunga [00:34:36]:
That is a really great point. Like, I don't think people really understand the asymmetry there of, like, being the first to market versus second, like, on a specific, like, feature or even product. Because, like, when you're first, you have no clue if you're gonna be right. You're gonna you have no clue if, like, the thing that you're building is gonna get, like, is gonna draw market attention and and and and the market's gonna love it or not. And the risk there is that you could be wrong, and you may be wrong a lot. It's far easier as a second player to just go, that worked, that didn't work, that worked. I'm just gonna do the things that did work and bundle them up in a nice pretty way that, like, is slightly different than the competitor that has our own little spinning and taste to it or style or flair. Like, we've seen this with one of our, competitors right now where, like, they just went and ripped 99% of the product.

Brian Vallelunga [00:35:27]:
And then they took a couple of other competitors we have, and they looked at their products, and they said, what worked it out their thing? And they took a couple of the best parts of their products too, and they mashed it all together. So it's, like, the best of Doppler, the best of this, the best of that. Now we're starting to see them invent their own stuff that hasn't been built by anyone. And we're like,

Ben Wilson [00:35:44]:
it's a

Brian Vallelunga [00:35:45]:
little bit harder than it looks. Some of that stuff isn't really resonating the way you think it would. It's just a hard game. And, like, they will build stuff that is absolutely beautiful and, like, will be amazing to have, and we will learn from that. And I applaud them when they do, because they're in the game fighting just as much as we are to to add value to society. Like, I think that's also the other thing that, like, people forget a lot is, like, we're all doing this just to make impact. Like, we only, like the company only gets more valuable if we drive more impact.

Ben Wilson [00:36:14]:
Yeah. You can't really have much, in my opinion, like, understanding your product. The one feature that that that thing has that resonates the most with me is on on your website, there's a a title screen just called firefighting 101. And that I don't know how much that's, like, the selling point for your customers. Like, this is why this is so important. But if anybody has ever had to do that exercise internally at their company manually, that is a special form of, you know, Geneva Conventions violation torture event. We're like, hey. We need to search through our code base, and see if any secrets exist that have ever been leaked.

Ben Wilson [00:37:00]:
And then we need to see if those are still active and then try to rotate them and ping the people. And when you look at an enterprise, like, either a monorepo or if you have thousands of different repos, and you look at the amount of code that you have, you're like, okay. We have, like, 1,700,000,000 lines of code. Alright. That's not what you're searching through. You're searching through every commit of that of every repo since the beginning of time, and think about the volume of that.

Brian Vallelunga [00:37:30]:
And also every Slack message, every email, and you have this thing of, like, did I miss something? What's the cost of missing something? Even if you captured everything, you're racing at the clock. Like, think about how fast data moves out of AWS. Right? Gigabits per second. So, like, every minute that you spend looking for a secret, not even just revoking, but just looking for those secrets, that is tons and tons of data leaving your system.

Ben Wilson [00:37:55]:
Mhmm. Yeah. The risk, like, you

Brian Vallelunga [00:37:59]:
It's enormous.

Ben Wilson [00:38:01]:
You get accidentally like, it happens fairly regularly, when we're interfacing with, like, our customers, where we get some support ticket, to engineering. It's like, hey. We need help debugging this. Here's here's, like, a notebook that we're using, and and it contains the exception. You start looking through it. I'd say 40% of the time, I have to immediately send a high urgency priority email back to them and whoever else was from that company, alerting them to please rotate your token immediately. Like, you leaked it to me. I'm gonna delete this thing that you sent to me and make sure that it's deleted off of our like, any place that we would potentially store it, and then please send me a clean notebook when you get this fixed.

Ben Wilson [00:38:50]:
But please rotate your secret right now. It's sometimes, it's not even where you would think a secret would be leaked, and that never would have happened if that was detected when you were putting that into your development environment and preventing people from doing that. I think that's a it's a big it's a big thing to add into a product as a feature that I don't think people always think about what the the lack of that could could, like, enable.

Brian Vallelunga [00:39:21]:
Yeah.

Ben Wilson [00:39:22]:
Leak of service principal key that is basically root admin access to your cloud provider. Think about what you can do with that.

Brian Vallelunga [00:39:30]:
Everything. Yeah. You can do everything with it.

Michael Berk [00:39:32]:
Literally. Yeah. Brian, are you familiar with Truffle Hog?

Brian Vallelunga [00:39:36]:
Yeah. Yeah. Yeah. Big fan of them.

Michael Berk [00:39:39]:
Yeah. I remember, like, a year or 2 or, like, probably 18 months ago, I was put on a security audit project, and I had no idea what I was doing, per usual. And found this cool little library that will actually go parse plain text and, like, do reg x matching on what tokens should look like typically. And then I remember I actually showed it to you, Ben. I don't know if you remember, but, proposed it in the security channel. Like, oh, we should add this to the Databricks product. And yeah. It I think it's sort of there now, but I don't know what they're using on the back end or if I inspired that at all.

Michael Berk [00:40:14]:
But it's a real, real problem. Like, because Yeah. The amount of Slack messages that have plain text tokens, it's it's kinda funny, like, in the in the Databricks database.

Ben Wilson [00:40:26]:
Yeah. It's scary. You know? You get a data breach. You should be worried about, like, what that looks like in the news. But what you should be really worried about is when it happens and you know it's been compromised and you don't hear anything. That's scary.

Michael Berk [00:40:44]:
Yeah. It's

Ben Wilson [00:40:44]:
like, what did they get, and what are they gonna do with it?

Brian Vallelunga [00:40:48]:
Yep. Mhmm.

Ben Wilson [00:40:49]:
Somebody's gonna monetize this in some way.

Brian Vallelunga [00:40:51]:
I think people often forget, like, the impact it has to real people. Right? A lot of times when I see the framing of, like, a data breach, I I see it from the sense of, like, oh, a company had a data breach bad on them. They're gonna get some fine. Like and I'm like, no. There's real people at the other end. Like, this company is in charge of storing real people's data, and real people's data just got out.

Ben Wilson [00:41:15]:
Yeah.

Brian Vallelunga [00:41:15]:
I have a quick story, and then they, for you guys. I don't know how much time we have left, but I'll try to make it quick. Right when I moved to Austin, Texas, I flew my mom out to kinda get a sense of, Texas as, like, Austin, Texas, and I was trying to sell her on it. And I took her to this barbecue spot. And right in the middle of our, glorious barbecue lunch, I get a call, and it's from the Texas, borders and customs. And they're saying, hey. We have, an illegal package with drugs and money in it, and we're fed and, we're investigating you now. You are formally being investigated.

Brian Vallelunga [00:41:50]:
And I'm like, oh, shit. This is the moment where my life just changes. Like, this is where, like, I go from, like, a CEO to, like, behind jail bars. Like, this is the start of that. I was freaking out. And they, like, start they started asking me questions to verify who I was, and a lot of these questions were, like, did you live at this address at this time? Did you do these things? Have you bought these things? Like, so much stuff that I was, like, only the federal government would know even to ask these questions. And at some point, we get our like, about an hour and we get our lawyers on the phone, and we got a team of lawyers working with them, like, because they're, like, really and they're, like, they say some scary shit too. They're like, if you hang up this call, that's, like, presumed guilt.

Brian Vallelunga [00:42:33]:
Things like that where you're like, oh, shit. I'm on this phone, and I don't have time to think now. I just gotta answer this barrage of questions that are coming my way. And it was scary. Like, because, you know, every question that you get wrong is, like, jail time. And at some point, my lawyers caught it where they we realized it was a scam.

Ben Wilson [00:42:53]:
Yeah.

Brian Vallelunga [00:42:54]:
And but they got me. Like, CEOs of a security company, they got me. Like and you think and I'm thinking about, like, okay. What how much information do they have on me to even make this feel legit? And then how much more information did they just get from me because of this call? Right? Now they have more intel to make an even better thing next time around. And it's just like you forget that, like, there's real people on the other side of this. I don't know if my mom would have been able to fight them like that. They she doesn't have a lawyer she can call. She doesn't have, like, people trained in in this kind of stuff.

Brian Vallelunga [00:43:26]:
Like, I do cybersecurity training. I just did one, like, a week ago, and they still got me. Like, I can imagine a whole bunch of random people who, like they they would have given everything. Like, their bank accounts would have been drawn. Their Social Security numbers would have been, like, totally used. Like, they had just enough to really convince me for, like, an hour and a half. Right? And that's a long time.

Michael Berk [00:43:46]:
What gave it away that they were scammers?

Brian Vallelunga [00:43:49]:
We asked some really hard question back that in inflamed them, they got them a little frustrated, and then an accent came out of the blue that was, like, totally, like, out of like, it was, like, an Indian accent. We're, like, oh, gotcha. Like, this is so not it just we were, like, well

Ben Wilson [00:44:05]:
your name is John. Yeah.

Brian Vallelunga [00:44:08]:
And it, like, it changed on a dime. And then, like, we started asking harder questions. And, like, the more we asked inflammatory questions, the more the accent came out. And then we're like, okay. If you're real, just come find you know where I live. Come find me and knock on my door and arrest me. And then they, like, hung up a second later. Yeah.

Ben Wilson [00:44:25]:
That's my standard. Like, I I don't know what whoever had my my Texas phone number before I got that phone. It's funny. It's an Austin number. But that dude must've had some security breach, because I get I get scam calls every week, from usually from Indian call centers. And I'm the clown that actually messes with them just because I I can tell when it's, like, oh, I don't recognize this number, but they're spoofing something, like some other number in my area code. And it's fun to just start, like, asking those sorts of questions to them, and then when they start I've I've definitely had not the the FBI one, but I've had ATF and the IRS before. I'm just like, hey.

Ben Wilson [00:45:15]:
You know my home address. Just send some agents over whatever paperwork you need, and they always just either hang up or say a bunch of word like, very bad words to me, on the call. It it always just amuses me. The more amount of time you can waste of theirs is the better like, the best thing you can do. So if you have free time, which you don't, but, anybody else who's listening, keep them on the line, and then just see what happens. Because, for sure, the federal government does not make phone calls about an indictment or anything or about an investigation. They like to drive their cars up and, talk to you face to face.

Brian Vallelunga [00:45:57]:
As they should for what it's worth. Like, when it gets to that series. Yeah. It it was just, I think we all have some story like that. But remember that these stories feel credible because of the data breaches.

Ben Wilson [00:46:10]:
Exactly. Like, how are they gonna know all of that detail about you unless some company leaked a password, leaked a token, and then some really smart sociopath was like, pay dirt. I'm gonna leak this online, sell this dataset on on the dark web, and I'm gonna make a couple couple $10,000 out of this, You know?

Brian Vallelunga [00:46:32]:
The ones that really get me, like, really get, like, they just bother the hell out of me, is they, like, go after these, like, old gentlemen that, like, are probably unhappy in their marriages, and they boot up these, like, AI fake girlfriends. And these AI fake girlfriend my my one of my uncles just got hit by this. And, like, I'm telling him, like, every time I meet him, stop. Do not talk to this woman any further. And, like, he keeps doing it. He doesn't know how to stop. And, like, they're like, oh, I'm this like, I'm I made all my money as this, like, really attractive 23 year old that's talking to this guy who's, like, 60 about, through, like, Bitcoin and and, like, crypto, and you should put all your money in this crypto wallet. And I'll teach you how to make a ton of money too.

Brian Vallelunga [00:47:22]:
And, like, oh, man. He's lost so much money doing that. And, like, to the point where, like, Fidelity and these other brokerages are like, we've seen that you've done this stuff. We're shutting down your accounts because we don't trust them anymore. Like, it's like a cascading effect, and I'm like, damn. Like, it's Mhmm. Like, it's not just, like, the scary tactics. Sometimes they go these other routes of, like, turning old old gentlemen that are lonely and are taking them.

Brian Vallelunga [00:47:44]:
And this is, like, their livelihood. Like, they lose this stuff. Like, they don't have money to pay rent or or or live. Like, they're they're not they're retired. They're not they're not there's no more money coming in. So you take that from them. They're done. Mhmm.

Brian Vallelunga [00:47:56]:
It's bad.

Michael Berk [00:47:57]:
Yeah. Cool. I know we're coming up on time. I wanna shift gears one more time and ask one more question before we end. I can't end the episode without knowing. So we were talking about iterating in the start up context. You have an idea. You have a product.

Michael Berk [00:48:16]:
Someone's like or you've pitched it on your, like, friends and family, and they're like, fuck. Alright. Fine. I'll I'll try it. And then they start liking it. You start getting traction. How do you think about pivoting within that space of growth? Like, what are the first principles, and does the objective function change?

Brian Vallelunga [00:48:34]:
And I just wanna make sure I understand. When you say pivoting, do you mean moving to a different idea?

Michael Berk [00:48:38]:
Yeah. You were the Bitcoin ML site, and now you're very different.

Brian Vallelunga [00:48:52]:
I list and maybe this is not the most first principle view. I don't think, like every time that I've, like, pivoted a company, I think it was more out of anger than it was at a, like, first principle thinking. You're just so frustrated. But I think there's signals that I look for. The first is, do I feel like I have the ability to grow it, and do I have an understanding of why I do not have this ability? Is it a timing problem? Is it a market problem? Like, it's so easy to build something that looks pretty and adds value, but the market just doesn't want. Like, the market only wants a couple set of things I've generally found, and they want it really bad, things that they do. And so I kinda, like, listen to the market of, like, are you telling me you want this bad or not? Like, I think, Segment's a great example of this. When Segment got started, they built, like, a really quick, like, couple line SDK, and that thing blew up.

Brian Vallelunga [00:49:42]:
It was dead simple. Right? And it blew up. You don't need something super complex and flashy to to to get signal. People really this is a real problem people want an answer to, right, or they they want solved. And I think a lot of times, we kinda, like, get into, like, this product mode of, like, we're gonna build something nice and cool. And then we spend all the time doing that, and we don't realize the market actually we never tested if the market wants us or not. You can always build the sexy thing later, but, like, test the root for principle stuff first. And so I try to, like, listen.

Brian Vallelunga [00:50:10]:
Is the market actually wanting this or am I, like, am I pushing or or is the market pulling? And the only right answer is the market is pulling you. To every founder out there, pull only. If you are pushing, that is a recipe for, unhappy life, in my strong opinion. And then the other one is, am I happy doing it? And I found that, like, after the shine wears off, if I'm not feeling this excitement and this buzz, then it's like, what are you even doing it for, man? Like, the market doesn't want this. You're not excited about it. Why are you doing this? And just give yourself that reality check. And I've usually found right at I get that peak of, like, frustration, another idea comes around. I'm like, oh, that's a shiny thing I should go investigate.

Brian Vallelunga [00:51:00]:
And that's usually my pivot point. It's like out of that frustrating, a new idea emerges. And one of them may end up being really good, and I've gone through that a couple of times where there's some good ideas that came out of that.

Ben Wilson [00:51:15]:
Would you say that most of the successful startups that are out there that are, you know, financially solvent and are going places, that that initial root idea came from somebody's personal experience instead of just their ideating.

Brian Vallelunga [00:51:32]:
Absolutely. I would say almost all my good ideas have come from experience. And and the I think that comes back to, like, actually really the the middle thing, which is, like, that's the market. You may just be one person in the market, but you're still the market. Right? If you're just ideating without any any grounding at all, then you are really not tapping into the market. You're just, like, throwing shit at the wall, but there's not there's no way to see if it sticks. Right? So you at least having an idea that you're passionate about that would solve a problem for you is, like, at least 1 person in the market, and then you gotta, ask a question, how many views are out there?

Ben Wilson [00:52:09]:
Exactly. I mean, it's something that we do fairly religiously, in our our product road map processes. We, like, force everybody to do user journeys and say, hey. Yeah. We we built this thing or this thing exists. Now let's go use it and build something with it, and how painful is that? And that identifies maybe not stuff that is directly related to what we're doing, but you're just like, man, this one step, this sucks so bad. Like, why is this so complicated? Or why did I mess this up many, many times? And then you check what other people did, and you're like, yeah. Everybody messed this up.

Ben Wilson [00:52:51]:
Do we need a new product here to solve this problem? And if you look at everything that Databricks has as a product offering is the result of that exact process of people being like, yeah. This thing, this one thing, we don't have it. Nobody has it, And it sucks so much that in order for us to fix this, we have to actually design something. But we use that same process of, like, build something fast and simple, get some feedback, and find out what direction this needs to take.

Brian Vallelunga [00:53:21]:
I also think just adding on to that, especially if you're, like, starting at 0, like, it may be different for Databricks where you already have this foundation, the set of customers, all those stuff, like, same with Doppler. But if you're starting from, like, ground 0, you just as much as you need to understand what the product you're building is, you need to understand how you're gonna grow it. Like, if you're b to c, you better have a viral loop in there somewhere, like and that thing better be strong. If you're b to b, you better understand, are you doing product led growth? Are you doing a sales motion, sales led growth? What is your mechanism for growth, and how are you realistically gonna go reach those customers? And how are you are you convincing them to do it, and, like, how hard is it to convince them? What is the activation energy that it's gonna take for them to sign up? Like, I in a way, like, I think, credit card companies like Brex and Ramp are actually great examples of this. They're like, hey. Fill up this, like, 2 minute form, and we're gonna give you a card, and that card represents money. And we're just gonna give you money, and then you'll just pay us back, like, 30 days later. Right? You'll you'll pay us back, please.

Brian Vallelunga [00:54:19]:
And that's, like, their model. And it's like, well, okay. That's pretty low friction. I fill out a form. I get a shit ton of money in tens to 100 of 1,000 of dollars, and then they'll connect with my bank accounts so that I can pay it back. Right? That's pretty low friction versus, like, other tools you may have to spend 20 minutes configuring it, another hour implementing it, another hour, like, convincing a whole bunch of people, and all of a sudden you're, like, days later and you're still not getting any value out of this tool. Right? So, like, what is the activation energy it takes to get someone to use this and get value out of it? And so that's why I'd argue, like, those credit card companies have grown so fast because activation energy is pretty much close to 0 as possible. Fill out a form and get cash.

Brian Vallelunga [00:54:57]:
Versus, like, a dead pool, it's like, well, now you gotta think, should I trust this tool? Is this secure? Especially with like a secret manager. Okay. How is this gonna integrate my infrastructure? Am I gonna be able to convince my developers to use this? There's an enormous amount of activation energy that it can take. I think Stripe's actually a good counterexample on the dev tool side. Does not matter convincing all your all your, friends and developers to or teammates to use this. All you need to do is add an SDK, and you can do payment processing. You know you need to do payment processing if you wanna be a real functional company that process payments on the Internet. Right? So it's dead simple.

Brian Vallelunga [00:55:26]:
It's like, am I using Braintree or Stripe? It doesn't really matter which one. I need the API that process the credit card. Right? So think about that activation energy, and it and it has to be comparable to the value that you drive. Right? People go through a lot of activation energy if you drive an immense amount of value. But most of the time, what you really want is that activation energy to be near 0.

Ben Wilson [00:55:48]:
Yeah. Anecdotally, I still remember at a previous company I worked at where it was the shortest sales motion that I've ever seen a vendor do at any place that I've worked, where I think the meeting was 20 minutes long, which was shocking, and the contract was signed the next day. And I've never seen more engineers happy at the end of that week. They're like, we get to delete our entire payment processing system. Yep. And, like, some we're just gonna pay somebody to do this for us, and it's easy. And the migration was done in, like, less than a month. And I was like, this is crazy.

Ben Wilson [00:56:29]:
Like, no other vendor has ever had the success because the product is so compelling. I was like, hey. We can do this safely and securely, and we're taking the risk here. And we'll just make sure that you get paid. I was like, man, what a great business model. Yeah.

Brian Vallelunga [00:56:45]:
Yeah.

Ben Wilson [00:56:46]:
Okay. Cool. I

Michael Berk [00:56:48]:
think we're roughly at time. Any other closing thoughts before I summarize?

Brian Vallelunga [00:56:53]:
I would just say to anyone listening, if you're trying to build a company, think about things we just talked about. If you're a developer, add secrets management early. Don't tack it on way later. You're gonna be very grateful that you, you add security from the get go because you don't wanna be learning how to firefight when you're firefighting.

Michael Berk [00:57:13]:
Heard. Cool. So lots of cool stuff discussed today. We talked about entrepreneurship, the importance of security. Some things that stood out to me, first, get developer adoption. Ensure you're not blocking their productivity. As someone who codes a lot myself, I freaking hate it when stuff doesn't work, and, it's like your your job is to debug. And so adding another thing to debug is pretty not fun.

Michael Berk [00:57:39]:
In terms of giving customers value, look for the common thread of value and also be request driven in your development. Regarding pivoting, think about 3 things primarily. 1st, do I have the ability to grow it? 2nd, does the market want it bad? And the market must pull you. You shouldn't be pushing the market. And 3rd, am I happy doing the thing? And then finally, founders should not be celebrities. So, Brian, if you wanna learn more about you or your work, where should they go?

Brian Vallelunga [00:58:04]:
Dalper.com. And then for me, personally, you can follow me. I'm not that big on social media, but I still have 1, Powell and GoodBrian, both on LinkedIn and Twitter.

Michael Berk [00:58:13]:
Cool. Alright. Well, thanks so much. Until next time. It's been Michael Burke and my cohost,

Ben Wilson [00:58:18]:
Brad Wilson, and

Michael Berk [00:58:20]:
have a good day, everyone.

Ben Wilson [00:58:21]:
We'll catch you next time.