Agent Orchestration: The New "Full Stack" in 2026 | Mike Piccolo founder of Motia Dev

Shamil Malachiyev: Good morning, everyone. Welcome to this week's episode of The Founder's Code podcast. Our guest today is Mike Piccolo, founder and CEO of an open source AI orchestration platform called Motia Dev, and in the past a co-founder and CTO of an IT development company called FullStack. Hi, Mike.

Mike Piccolo: Hey, thanks for having me, Shamil. Excited.

Shamil Malachiyev: Thanks for coming to the podcast. I'm really excited to get into this one, because we have a lot of similar experiences, and you've taken the path that any software development company founder wants to take in the future: starting his own product company. Shall we start with, can you tell us a little bit about your vision at Motia and where you are at the moment?

Mike Piccolo: Well, now an AI company, right? Everybody is. So, the way that I think about Motia, and the way it came into being, is that over my career I've seen these paradigm shifts. If you're in the software engineering space, every 10 or 15 years you see these paradigm shifts happening. Fifteen years ago, when I started my career, we were in the middle of digital transformation. I was a software consultant for a company, and we were taking brick-and-mortar companies, plumbing companies, businesses you wouldn't really think of as software companies, and digitalizing them. They were becoming full-blown software companies.

Prior to this, everybody was building their own stacks from scratch. This was what's called the PHP spaghetti code era. If you went from one company to another company as a software engineer, you'd have to relearn the stack from scratch: I have no idea how they're doing routing, how they're doing the data layer, how they do modeling. The level of complexity grew to a point where this was just unsustainable. You're taking all the business logic of a company and putting it into a bespoke system per company. That level of complexity wasn't possible to sustain throughout the industry.

So one of the big shifts was when Ruby on Rails dropped. This was the beginning of the MVC server-response era, and every programming language basically adopted this paradigm shift of MVC. Depending on how technical you are, the details don't really matter, but the point is that everybody adopted this model-view-controller pattern that became the backbone of digital transformation. Ruby had Rails, Python had Django, PHP had Laravel, Java had Spring, and .NET had .NET MVC. That was one big paradigm shift I witnessed, and that was all in the backend.

Then, once that problem was kind of solved, everything interesting, all of what's hot in software engineering, moved to the frontend. Basically you're taking all the rest of the complexity and throwing it into the frontend, ever-increasingly complex applications being built in the business world. And this triggered a new paradigm shift: the frontend framework wars. This was Knockout, then Angular, then Ember, then React, then Vue, and there are a million of these now.

But React won this battle by orders of magnitude. React is ubiquitous; it's definitely being used in the components on the screen that we're looking at. And the reason why they won this battle is really important. It's a lesson that I learned; I was involved in the framework wars and had very strong opinions on it. What they did is they made two sides of the house really happy. They made the implementation engineers really, really happy, but they also made the library maintainers, the framework developers, the principals and the architects really happy. And they did something kind of ingenious: they created a single core primitive that encapsulates, that is an abstraction over, anything that can be UI. In React they call this a component. I'm not going to dive into the technical details, but they made it so an engineer, somebody building an application, could think of their entire application as just a component with nested components underneath it. You only need to know one thing, and maybe three things about how a component works, and you can build as complex an application as you want with just that mental model and those tools. But they also made the other side of the house really happy, because they gave these really strong core primitives that you could build tools on top of. That's ultimately why we have Next.js and Redux; the React ecosystem is massive, and Vercel is a giant company built on the backbone of those primitives. So that was a very interesting lesson.

If we jump forward to today, everyone knows we're in the middle of another paradigm shift because of AI. And Motia is my bet on where software engineering is headed and what is going to be necessary. We're doing something very similar. If you think about most of the interesting things that happen in the world right now, they happen between a human brain and a user interface. In the business world, you're looking at two pieces of data, you're entering a form, you're hitting a button. This is between you, your brain and the user interface. But with AI, you can effectively take that interaction and move the whole thing back to the backend, in the form of automation. There are many different ways to describe that: agentic automation, workflow automation, durable execution. But what you're doing is taking that step in the process or the workflow and moving it back to the backend.

So what does that mean, if we're thinking in terms of these paradigm shifts that go backend to frontend to backend to frontend? It means the backend is now hot. Backend engineering is where all the action is going to be, and it's going to be in the form of asynchronous compute: asynchronous workflows, background jobs, queues, things like that.

At Motia, we're basically copying React's model. We created one core primitive called the step. The step represents all the existing steps in the backend, API endpoints and background jobs and scheduled tasks and queuing and workflow, but it's also going to cover all of these things that are coming over the line. History doesn't repeat itself, it rhymes; but if we're looking at these paradigm shifts, it follows the same pattern. A massive amount of complexity gets thrown into systems that weren't built to support it, then somebody builds the abstraction with the correct mental model that makes the two sides of the house happy, and then it takes off. And then you're able to build way more interesting and complex things on top of that simplification. So that's where Motia came from: following these paradigm shifts and hopefully understanding where this is headed. Motia is a backend engineering framework that does that abstraction and simplifies the mental model of building these complex, intelligent backend systems.

Shamil Malachiyev: So back in the day, we could see a lot of development teams building ERP systems within these companies, trying to automate everything themselves. With this AI orchestration, you can basically create your own AI-powered connectors, maybe to an existing ERP system or a new one.

Mike Piccolo: Well, I don't actually think of it in the way that these systems are going to be all AI. There are two sides of AI, right? There's code generation, using tools like Claude and Cursor. And then there's the other side, which is what you're talking about: the agentic solutions. Those are things you've deployed that are running and using stochastic reasoning in the form of LLMs. But you're still going to have hard-coded, deterministic code that is running. Another prediction of mine is that we're going to stop thinking in terms of, this is an AI system and this is an API and this is a background job queuing system. It's all really just mixed together. Whether one of your steps is using stochastic reasoning to move to the next step, or it's using deterministic, hard-coded code, is going to become less relevant. We're going to stop thinking in those terms. It's all going to be one interconnected system.

Shamil Malachiyev: And can you bring a bit of clarity for a lot of the people listening to this podcast who are maybe not necessarily technical, CEOs in companies whose understanding of the whole AI world could be quite shallow? What is AI engineering? Because a lot of people say, okay, it's people who use Claude Code, it's people who run the orchestration workflows.

Mike Piccolo: Yeah, it's fuzzy. It doesn't exist; that's the big answer. I actually just posted on LinkedIn about this: AI engineer means like 20 different things. And it's because it's so new that we haven't come up with the terms yet.

There are experts who are Cursor experts, using tools to generate code. That is a whole skill set. It used to be called prompt engineer, now it's called context engineer, and in a month it'll probably be called something else. So I don't have the name for it, but right now we're going with context engineer, and that describes somebody who is setting up your software development lifecycle and the tooling around it, and how your team is generating code using AI.

But then there's a whole other side of it, where you're talking about solutions. These are deployed things that are integrating with OpenAI and Claude and doing whatever the task is. Before, you would have had to hard-code that particular task, and it would be really hard or potentially impossible to do. Now you're integrating with these LLM tools that are doing the task. But those are different skill sets.

And further beyond that, there's the actual AI engineering: building LLMs, doing image recognition, all of that. That's a whole other field. The problem is, when you say AI engineer, you can mean any of those things.

Shamil Malachiyev: Yeah, the thing is, everybody online seems to be like, AI is hot, I'll put AI in front of my position, now I'm an AI engineer. And nobody knows if they're trying to portray themselves as these hundred-million-dollars-a-year senior LLM creators, or just somebody who's discovered that you can use Claude Code.

Mike Piccolo: Yeah, it's really hard to hire for that, right? Because you say AI and engineer, and you get a machine learning engineer and somebody who uses ChatGPT, and you're like, I didn't want either of those.

Shamil Malachiyev: And taking it back maybe three or four years, using an AI assistant to code was not as developed as it is right now, and a lot of people were still trying to get into the IT world as engineers, backend developers, frontend developers. For everybody trying to analyze their own place in this field over the next 10 years, and for those starting out thinking, okay, I want to do something with technology, what would you suggest as the core stack to follow or the main technology to focus on?

Mike Piccolo: Right now, if you were specifically going into software engineering, it would just be: double, triple, quadruple down on Cursor or Claude Code, and learn these tools really deeply. It actually is pretty hard to find somebody who's legitimately an expert at setting these tools up properly and then ensuring that your team is using them properly. That is a very hot commodity right now, and it's really hard to find, surprisingly.

Because to do it right, it doesn't work, certainly not yet, where you can just give your team Cursor and then say, we're now 50% more productive. It doesn't work like that. You have to really learn the tool, learn how it works, learn how it works across a distributed team, and how to incorporate it properly into the software development lifecycle for you to get the gains. You can get there, but it's actually really hard. The tools right now have a tendency to write a lot of code, and if you don't understand how to rein them in, when to use them and how to constrain them, you'll end up with either too many PRs or a lot of slop, or something like this.

So the advice would be to double and triple down on that. If you're going to college now, I don't think being an expert in the particular syntax of a particular language is going to be a valuable skill. But that's not really what software engineering was in the first place. That was what you ended up doing to produce the result, but there's a lot more to engineering than that.

Shamil Malachiyev: Can we go back to your early days? Was there anything specific in the way that you were brought up that showed that one day you'd take the founder's path?

Mike Piccolo: Mostly, I'm just not good at being an employee. I have high disagreeability. I have all the traits that edge you in that direction, so it never really was much of an option for me. The other thing that helped me really make the decision was that I graduated in 2008 from college, which was literally the worst job market in 90 years, I think. There were just no jobs. In hindsight, that's kind of a blessing, because maybe I would have gotten a job that I wasn't quite satisfied with, and I wouldn't have built FullStack and I wouldn't be on this path. But it definitely did influence me.

The funny story is, when I was in college and picking my major, everybody who was giving me advice cut their teeth in the '80s and '90s. Their advice was: oh, you should get a generic business degree, then you have all these options, this is what you want. I graduate, the financial crisis happens, and everybody's like, what do you mean? You need to be specialized. What are you doing with a generic business degree?

Shamil Malachiyev: So you didn't know where to apply it directly, without having any specific skills in any particular domain.

Mike Piccolo: Yeah. I guess I would also say: be careful of the advice of the previous generation. The market is going to be different than when they graduated.

Shamil Malachiyev: And did you also feel the influence, I remember I was in college around the same time, graduating around 2008, 2009, and this was just the era when everybody started using Facebook. Everyone was playing these games in browsers, Mafia Wars or something, and everybody wanted to be like Zuckerberg. Before that, being a tech billionaire was something you only heard about. After Facebook came, it became more public; you could see those people, everybody knew the stories, and new software kept popping up around the world. Did that influence you, in terms of, well, if these college kids can do this, maybe I can too?

Mike Piccolo: Yeah, I think so. I tried to do multiple startups that were influenced by the Zuckerbergs of the world. But what I fell into was IT services, which is a very known business model. There are a million IT services firms; the business model is very known, the go-to-market is very known, so there are a lot of things you can compete on knowingly. My advice to anybody, especially for your first one: you should not be aiming at making the next Facebook. You should find a business model where you have some kind of take on why it would succeed and how you're going to do it better than the competition. In general, I wouldn't suggest right out of the gates trying to be the next Facebook.

Shamil Malachiyev: Yeah, I think we've all learned that lesson. Back in the day, when I was working on my first startup, my co-founder came to me and said, we can make a social network like Facebook, but better, with all these features. No, that's not going to work. And nobody was talking about picking something niche back then. It's now that all of this knowledge is all over the internet, and you can easily get access to it. And how was the start of your journey with FullStack? Where did you meet your co-founders?

Mike Piccolo: It was called Hacker Lab, a coworking space in Sacramento. After college, I went and worked construction; like I mentioned, there were no jobs. I liked construction, but I figured out that wasn't my long-term path. So I went to a coding bootcamp, and I met one of the co-founders there. Then I went off and did the software consulting that I mentioned, and the two of them went off and tried several startups. Then I left software consulting to join them and build these startups, which had marginal success, but not enough.

We actually did exactly what I'm saying not to do: we built the Uber for lawn care. It had a lot of success; I think it got listed as number one on Yelp. My co-founders built the majority of that, and I joined on to help them. There were like four or five companies doing that at the time, and several of them got VC money. The business model just didn't work at all; it wasn't a functioning business model at the time. So we dropped that one, and then we did a couple more, and eventually we're kind of looking at each other, we'd both recently gotten wives, and it's like, maybe we should make some money.

So we just took whatever we could. I think our first project was a $200 WordPress site that I built myself. And I'm not a very good frontend developer; building websites is not my forte, I'm much more on the backend. But that was our first project. It was just like, hey, let's do it. And then the next one was $1,000, and the next one... we just moved the ball along.

Shamil Malachiyev: Yeah, I can relate. I think my first one was optimizing a WordPress website for speed, for 250 bucks, around 2016. I think that's the area everybody was getting their feet wet in the technology world. And can you talk a little bit about your leadership style?

Mike Piccolo: In terms of leadership, it's really all about ownership. I like to give people massive amounts of ownership and leeway, and let them just take it. Responsibility is very rarely given. It's generally just taken or earned. So that is how I operate the company. I want to bring people in who just see the void: no one's doing this, I'm just going to take that, and now that's my responsibility. And I'm like, yes, you own that now. You're the DRI. Good. It's about finding people who thrive in that environment and really want to be in it, who want to be taking ownership and more responsibility, doing important things and not having to ask permission to do them. You come to me and say, I have this whole go-to-market thing set up, it's ready to go, I really think it's going to work, and I'm like, yes, go. Let's test it.

Shamil Malachiyev: And what would you say, do you remember the toughest years or the toughest months you faced at the start of FullStack? Were there any moments where you thought, shit, we're going to go bankrupt in two months?

Mike Piccolo: I think it's basically cyclical. There are really, really hard points, and they come up every couple of months. Consulting is very much a grind, and there are very, very high expectations on consultancies to deliver. It becomes very stressful to get a release out; somebody's not performing on the team, and yourself or other people have to carry that weight and get it over the line. Then you get over the hump, and things stabilize for quite a while, they can stabilize for a year at a time. But FullStack has like 150 clients or something like that, so there's always something going on that has that feel to it. I don't think there was any one particular time; most of it, in consulting, is around project completion and the level of stress involved in project completion.

Shamil Malachiyev: Do you also feel like you build resilience over time? When you just stop taking everything close to heart, and you're like, well, we'll manage anyway.

Mike Piccolo: Yeah, I think that's true. You get resilience, and you also prove to yourself that you can do it. You're like, we've been in this kind of spot 20 times, and we figured it out 20 times. It's very likely we're going to figure it out. I don't see exactly the path over the next month by which we're going to figure it out, but we're going to figure it out. And you can tell that confidently to your team, because you have the track record under your belt.

Shamil Malachiyev: Yeah. And everybody's trying to make sure that the people around them are the best suited to bring about success. You've talked about ownership, which is one part. And when building that culture, it's important, as a founder, to protect it and make decisions. If you see somebody failing to fit within the culture of the company, or starting to be toxic, it's up to you to make a fast decision, so you don't regret making it later. Was it always easy for you to keep the culture in place and say, you don't fit in, I'm sorry, you'll find somewhere else? Or were you trying to be kind and nice?

Mike Piccolo: I've done both over the span of my career. I've let things linger longer than I should have, when I knew it wasn't going to work. And I've also become better at that. But it's not always clear. I've let things linger and it turned out amazing: they actually got over the hump, and they're some of the best team members. So it's very, very difficult to make that call. Some people take three months, six months to get into a groove, and some people hit the ground running. And some of the people who took six months had a higher trajectory ultimately. So I actually don't know how I feel about that. I don't know how to identify it consistently yet. I don't have a heuristic for it.

Shamil Malachiyev: Whenever I talk to founders, there's always this question about whether you should trust your gut feeling. A lot of them say, I had a gut feeling about that person, but I didn't do anything. What is your relationship to that gut feeling? Do you always trust it, or do you sometimes give it time to play out?

Mike Piccolo: I tend to trust my gut a lot, but I do try to validate it with data. I'll trust my gut, and then I'll make a mental note or a physical note of it, and follow up on the gut feeling and reflect on it. Just recently I had a situation where I had a gut feeling and I was wrong. And I was very glad that I didn't react on that gut feeling, truly.

Shamil Malachiyev: So impulse control is something that you practice. Something I know is that sometimes not making a decision is the right decision, just giving the situation time to cool off. Maybe it also allows people to put themselves forward and sort things out without your oversight.

Mike Piccolo: Yeah. I will say, I am very direct. I'll tell people the expectation. And the cultural piece: I'll tell them, I expect you to go take things on your own, even if I don't tell you. That's how I expect the team to operate. I need a team that is finding things, being creative, and providing value outside of the things that are listed in the role. It's just not going to cut it if I have to say, here's the role and here are the eight things that you do. We're not really living in a world where that is an option, in the world of AI. You can learn and implement and do things so quickly now with AI that you have to be creative beyond just a simple checklist.

Shamil Malachiyev: And can you tell me what's different between the Mike who's right out of college, the Mike who's starting his years at a consultancy, and the Mike we see today?

Mike Piccolo: Well, I went to San Diego State, and I didn't take school particularly seriously. There are a lot of ways to be distracted away from school in San Diego. There's the beach, and there's partying, and there's girls, there's tons of things to distract you from the academic side of college. So the Mike right out of college was nowhere near ready for any reality of life. I needed a lot of different lessons, and one of them was going into construction, at my stepdad's company.

I learned a lot of things from him that I've used throughout my career directly, which is kind of interesting, because you wouldn't think there'd be that crossover. One of the things I learned from him: when we would go into a project, his goal was to become best friends with the super, the person in charge of the project. We'd be on these commercial projects, we're swinging doors, hanging these large commercial doors, but we would leave the project with like eight different things. On these big commercial operations, there are always vendors that just drop out, don't show up, or can't do it. And my stepdad was really, really good at the relationship piece of that interaction, and he would just pick the work up.

We were building the doors at a prison in Vacaville, a large, huge building, and the super comes in one day and says, oh my gosh, the T-bar guy dropped out. Don, can you do the T-bar? And he says: yep. And I walk out and I'm like, Don, we've never done T-bar. And he's like, we'll figure it out. Just cool as a cucumber. We'll figure it out. And we did. And that's the lesson: just say yes to stuff. You will figure it out.

Shamil Malachiyev: I think it's a big secret to growing an IT consultancy too. Because if you have a strong enough team that can persevere, spend a night on it, find things out, all of this is just human-made knowledge, and anybody can master it. It just takes time. Just say yes.

Mike Piccolo: Oh, absolutely. You need that. But you also need a leader who can express to the team the calmness that he expressed. If he'd freaked out, I would have been like, what are we going to do? But he's just: I'll figure it out, don't worry about it. We're going to go research, we'll buy a laser level, we'll do it.

Shamil Malachiyev: And were there any other mentor models for you? People you've looked up to who influenced you?

Mike Piccolo: Yeah, my first boss at the software consultancy was a big mentor, both technically and with all the soft skills, watching him, how he interacted with clients and how he managed a project. He was a big influence on me for sure.

Shamil Malachiyev: And how has your "why" evolved over these years?

Mike Piccolo: That's a good question. For FullStack, consulting is a lucrative business, and it's challenging. It's very fast-paced, and you learn so much stuff in consulting. Every three months you're on a new project, doing something new, kind of at the bleeding edge, teaching your clients how to work at the bleeding edge. There are lots of really interesting things about consulting that are different than working at a large company, where you have one role, and then a year or two later another role, focused on one niche thing. That has its merits as well, but in consulting you're learning at an incredible pace. So the first phase was doing that in projects, actually doing a lot of project work. And the second half of my career at FullStack was working on product and higher-level things as CTO, higher-level business management skills, learned in real life. I went to business college, but like I said, I didn't study much.

And the why now: I've always been really, really interested, from the beginning of my career, in building the right abstractions, like I talked about at the beginning. It's probably hard to understand if you haven't coded yourself, but you can make something beautiful and elegant and artistic in a way, and solve a problem at the same time. So this is a bit of a passion project. I want it to succeed and become a very successful company, but the why is coming from a bit of a different place than FullStack.

Shamil Malachiyev: So would you say the role of money in life kind of diminished after you felt yourself more secure, after being in consultancy for 11 years? Now you feel it's safer to pursue something of passion, rather than a pure focus on short-term returns.

Mike Piccolo: Yes, absolutely. I became successful enough to where finances are not a massive concern at this stage in my life. And it's also not as enticing to make the next X amount. When you're first starting your first business, that's a very different thing.

Shamil Malachiyev: Oh yeah, the first $200: it's a business now. And with being a founder come a lot of things that only other founders understand. One of them is the importance of being ready to brace the day, because most days are fairly complex and hard. You have to make a lot of tough decisions, you're always fixing something that's broken, and you're trying to build a team, do everything. In terms of good days versus shitty days, what is your percentage?

Mike Piccolo: That's a good question. I think Elon talks about this a little bit: how, when you own the business and you're the last line of defense, the entire business acts as a giant filter to give you all of the worst possible problems. Because all of the easy ones get solved before they reach you.

So it's pretty often that you're dealing with something that is not super fun. There are those moments of enjoyment, but most days you're going to have to deal with something that is boring, frustrating, monotonous, challenging, something that makes you angry. There are going to be these negative parts to running a business, and it's going to be pretty consistent. So part of it, like you said, is adapting, being able to handle that. I really think there's a huge part of being successful as a business that is just outlasting the people who can't deal with that.

Shamil Malachiyev: Yeah. And this is partly why I love asking this question, because the next one is: what are your mental frameworks and life hacks that help you persevere through the days when you don't wake up thinking, let's nail the day, but rather, fuck, so many problems to deal with today? What do you say to yourself? Are there any actions you take to get back into the right frame of mind?

Mike Piccolo: I don't think I do anything in particular to manage it in the moment. But I will say there are things you can do throughout your life that will make you much, much better at this. Like working construction: I had to sweep the entire floor for hours, these monotonous things. It was not enjoyable, but learning to do really hard work in any capacity will help you get through these things. I've built up a tolerance to hard work.

Speaking of mentors: when I was growing up, one of my best friends, his dad, we would go over there, and he would just make us work the whole time. We'd work for like six hours on a weekend. And it would be completely ridiculous stuff. There's a giant hill in his backyard: go pick all the weeds off the hill. And we're like, why? He's like, yeah, that hill needs to have all the weeds picked off. And we would just do that for six hours, and then he would basically do anything we wanted after that. We'd have milkshakes, we'd go do the coolest thing. He was a really great guy, and he was teaching us a lesson. You need to put yourself in situations where you're purposely going through some kind of difficulty along the way, so that when you get to these things in repetition, when you're owning a business, it's just part of it. This is how you run a business: dealing with these difficulties and being able to manage them.

Shamil Malachiyev: Is there anything else you did to build your character? Because to me it sounds like you were faced with a lot of hard things to prove to yourself that you can do hard things. So when there's a large challenge in front of you, even now, a new investor meeting, a customer acquisition strategy meeting, you know you can handle it, because you've done hard things in the past and you know what hard feels like.

Mike Piccolo: Well, working out and physical activity is the main thing. The hack that I use for working out is that I hire a trainer, and not because I want training advice. I can do all the lifts; it's nice to have one, but that's not what it's for. What it's really for is to make me show up at 6 a.m. every morning and grind and work really, really hard. If I didn't have a trainer, I could probably do it for like six or eight months, but then the holidays come, I get sick, and then I'm off for three months and I've set back. So I do the training three times a week, and I do tennis two to three times a week, every morning at 6 a.m. And I work really hard while I'm doing it. That is a daily ritual: I'm going to start the day by working really hard, right off the bat.

Shamil Malachiyev: And how easy is it for you to keep that discipline?

Mike Piccolo: It's not easy. That's why I have the trainer. I've found out about myself that I can only do it for six to eight months with consistency. I know it about myself: the willpower will not get me through the sickness and the holidays and the other things. So that's the hack. It doesn't need to be a trainer; it could be CrossFit or a group, or you could have a friend. But I need some kind of external pressure. All it is, is that this person expects you to show up and expects you to work really hard in the morning. That's where I found my ability to be consistent year over year, where I'm going every single time. I don't miss.

Shamil Malachiyev: It's adding that extra pressure. I kind of want to try that for myself, because I remember back in the day, when you have a trainer, you feel accountable, somebody holding you accountable for all of that.

Mike Piccolo: Yeah, there are a couple of tricks. There's one that's very cringy. Use a trainer, use whatever tricks, and get into the best physical shape you've ever been in. Then take a picture of yourself and post it publicly. Now, in your mind, you identify yourself as this fit person, and you've socially told people, I'm this fit person. If you go against that contract, you're letting yourself down, you're letting everybody else down; you can no longer identify as the fit person you posted. It's very cringy, but it's an external pressure. Social pressure is very, very strong, and in many people, me included, I care what people think, probably more than I'd like to. So: hey, let's use that. I did it a couple of years back. I posted a picture on Twitter with my shirt off, and I probably got made fun of, ridiculed. But the point of it was: now I have to keep that level of fitness. I've got to do it.

Shamil Malachiyev: Yeah. And how easy was it with age? Because I think after 30, especially, it gets a bit harder to maintain the same level of metabolism. Have you found yourself, especially on bad days, craving something like sugars, quick dopamine?

Mike Piccolo: I went through a pretty heavy Tim Ferriss phase back when I was 25 to 28. I did the slow-carb and the low-carb and all of these diets. So I did a bit of experimenting, and I built up a bit of intuition about what I need. I can kind of feel what is the right amount, whether I overdid it, and I can pull back. I don't do macro counting or a specific diet; I've just done enough experimenting with these diets to find a happy medium that works.

Shamil Malachiyev: Cool. I almost feel like after this episode, I'm just going to go to the gym, because I haven't been in, I think, five months, and I'm getting to the point where it just has to happen.

Mike Piccolo: The trainer, I'm telling you. A trainer or a class. Just pay for it up front, commit to it, or find a friend at the class who wants you to be there, so you can't let him down. That's all it took for me. It was just the external expectation.

Shamil Malachiyev: And I want to also touch on one more topic, which I think resonates with a lot of people I speak to. It's about long-lasting energy; motivation is maybe the wrong word. Everybody's trying to find the way to keep themselves ignited, to keep the fire burning. For a lot of people it starts with, I have to prove myself to my parents, to my mom, to my dad. Some people stay in their forties still using that as the main source of fire, which works, and I think it's one of the strongest types of fire to ignite a founder. Have you ever tried to self-analyze and understand what is driving you from within?

Mike Piccolo: Yeah, I think some of it is that you need to prove yourself to somebody. I haven't analyzed who it was exactly, but it's your parents or something: prove that you are successful, however you measure that.

Another one that's really interesting is the chip on the shoulder. For me, the best example of this is Michael Jordan. In his Hall of Fame speech, there was a good portion of it dedicated to the high school coach who cut him. That's an incredible thing. Everyone agrees he's the best basketball player of all time, he's giving his Hall of Fame speech, and he's like, that guy cut me off the team. He had that chip, and he used it. It's probably why he was out there practicing.

I'm not going to go into the details of my chip, but I had a chip. And the interesting thing is, I think Jordan is right: it's super trivial. It doesn't matter. If I met the person who gave me this chip, I would probably be laughing with them and having fun. I would not go and be like, I told you. It's not really for that. You kind of need that chip. And honestly, I think the person who gave me the chip was actually right at the time. I wasn't on a trajectory for success; there was no indication that I was going to get there. So I think it was actually really helpful, if you look back on it.

Shamil Malachiyev: And have you found that after your success at FullStack, that voice almost doesn't make sense anymore? You can hear it, but it doesn't ignite you anymore, because you're not a loser, look at all this stuff.

Mike Piccolo: Yeah, you've got to get a new chip. Make one up if you have to. There's always something. Like: I successfully built this company with my co-founders. I haven't done it on my own, right? So now I have a chip: what can I do on my own? It's a whole new endeavor. There's always something else you can do that is interesting, that raises the bar for yourself.

Shamil Malachiyev: And what are you most excited about for the next five years?

Mike Piccolo: I mean, it has to be AI, right? I'm pretty deep in this stuff, and it's just insane what's going on. Using these tools and seeing what they're capable of already... I refuse to predict past five years, just because it's too crazy. I don't think anybody knows what's going to happen five years from now, truly. So that's definitely what I'm most excited about. And a little nervous. There's the possibility that it's substantially disruptive in a bad way. There will be negative consequences to it, for sure. So: a lot of excitement, maybe a little bit of hesitancy.

Shamil Malachiyev: And I think with your experience in consultancy, because I can say from myself, being able to work not in one company but across a lot of projects and companies, you can see inside, you can hear first-hand about the real problems with adoption. Knowing all of that, having seen the insides of all the industries, how they operate currently and how they're actually using AI and embracing it, do you believe we are living in somewhat of an AI bubble at the moment?

Mike Piccolo: I believe it's a bit of both. I posted on LinkedIn about this. You have your AI-is-God folks, who are simply hyping it: it's amazing, AI is God. And then you have your AI-is-slop, AI-is-trash folks: it doesn't ever produce anything. I think the reality is basically in the middle of those two things.

Like I mentioned before, you can't just give your entire team Cursor and expect everything to work itself out. It's going to be a lot of hard work. There's going to be a lot of red tape, especially in larger organizations: security, compliance, legal, bureaucracy, adoption rates, training. So I think it's going to take quite a bit of time. It's moving faster than any other kind of technology in the past, but I think it will still take some time to proliferate and be effective in all of these different types of organizations.

Shamil Malachiyev: I often think about what Elon Musk said recently: in the future we'll have companies using AI that outperform companies using less AI, but then those are going to be massively outperformed by companies that are entirely run by AI. How far do you think we are from that vision, where we actually have companies running fully autonomously? There's just one person who registered the company, and everything else is done by orchestrated AI engines.

Mike Piccolo: I guess it depends on the size of the company and what its focus is. If you gave a very narrow goal to a recursive loop, I don't know if you've seen this Ralph, the recursive loop thing that's been going around; it's basically: give AI a goal and it just recursively loops until it's done. It's not there yet. I could definitely see that happening within the next five years, and I could potentially see it happening for narrowly scoped things in the next two years.

Shamil Malachiyev: Well, I think Motia is going to be a jump in that direction, helping build that future.

Mike Piccolo: Yes, I hope so. I hope so, for sure. It's a lot harder than people think to get these things to work. To get any particular solution to work in production is a lot harder than they think. For instance, you can take any particular process or workflow, and you can play around with ChatGPT or Claude, and you can get something to work, and you're like, oh, it works. And then, if you eval it, if you use evaluation tools and you run it a thousand times, the question is: how many times does it work out of those thousand times? That's the important part, not that you could get it to work once. Depending on the domain that you're in, does it have to be six sigma? Does it have to work 99.9% of the time? Or maybe 95% is fine? But it's even hard to achieve 95%. You're talking about concurrency: how often is it running, how many can run in parallel, atomic state updates, what if this runs before this, are there collisions and retries? What if OpenAI goes down: what happens to all those jobs that were churning? There are a lot of infrastructure details that go into these kinds of solutions, that you graduate through over time, to build a thing that is durable and resilient in the real world. And especially if you're testing these things out: the tests always work. And then as soon as you put it into the real world, it gets messy and difficult.

Shamil Malachiyev: Oh yeah. I can almost imagine, five years down the line, OpenAI going down for 10 minutes and half the country just stops working. Nothing works.

Mike Piccolo: I had not even thought about that, but you're absolutely correct. OpenAI's uptime is going to be a national security concern. Yeah, exactly.

Shamil Malachiyev: Everybody's just going to be monitoring every morning, like, okay, we're up to schedule. And there was one more thing: maybe you'd share some advice for somebody who is a founder right now, or at a job right now but thinking of starting something of their own. What kind of advice would you give them?

Mike Piccolo: I don't know if I have exact advice around this, but I've been thinking about it a lot, and Chamath posted something about this yesterday. The whole business model is changing, and founders need to think, if you're really truly just starting, you need to think very carefully about the business model, especially if you're building software, like I'm thinking about, like I'm building.

When software becomes extremely easy to produce, cheap to produce, and very fast to produce, it's going to be hard to understand how to monetize it over a long period of time. Traditionally, you build SaaS, and you add incremental improvements to it over a long span of time, three years, five years, and you're monetizing that investment you made in building the system. And I don't know exactly how that works when you can build the whole thing in two months or something like that. It changes the business model dramatically. There will be business models that work in this environment, and I don't exactly know what they are. But if I were another founder, not building Motia, I would be thinking about what those business models could look like in that type of environment.

Shamil Malachiyev: Well, Mike Piccolo, everyone. I hope I pronounced your last name correctly.

Mike Piccolo: You did. Good job.

Shamil Malachiyev: Thanks so much for joining me today, Mike, and sharing some of your thoughts. I wish you all the best building Motia. We'll be following your success and your development, and hope to have you on this podcast again sometime down the line, to talk about where the AI generation has gotten us. Thank you so much for being a part of this.

Mike Piccolo: Well, thank you so much for having me. It was great.