I've built a number of businesses over the years and sold some of them, and I learned different things from each one because each one was different and attracted different kinds of people. The most profound experience for me was my tech company. We were building a competitor to Slack, very tech heavy, very dense on the product side. We started in 2018 and really picked up steam in 2020 when COVID hit. It was my first fully remote business, which back then was a genuinely new thing, so we attracted really great developers in the Philippines who had been commuting two hours each way to office jobs and were thankful for the chance to work remotely.

These days I'm really focused on AI and on helping companies raise money. A lot of the money is going into AI right now, and I split my time between vibe coding my own things and interviewing business owners and investors, raising for AI startups, and helping VC funds raise their next fund. I want to be deep in what people are thinking about in AI, how they're using it in their personal and professional lives, and then partner with great people so that everyone can help each other and everyone can make money.

When you can't see people in person, it's really easy to forget they're even there, and they start to wonder whether you notice they exist. So I worked hard at the FaceTime. We had about 16 people, which was small enough that I could do one on one calls with each of them once a quarter where they could ask me anything, business or otherwise, just so we could get to know each other. I also set up virtual reality nights where a few of us with headsets, and everyone else on a web or desktop app, would get together to play games with one rule: do not talk about work.

The thing I kept coming back to is that these people are putting their time, their energy, and their skills into building my vision and my future. Of course they learn new skills along the way, but it's really my future they're building. So I want to know who they are, what drives them, what makes them happy or sad, and how I can make their career more enjoyable. A lot of founders fail to do that because they're so focused on building the product or the sales that they forget about the people doing those jobs day in and day out.

I'd put over half a million of my own money in before I raised a cent from anyone else, because I thought I could seed strap the business. The idea is you fund yourself past the pre-seed stage so your first real raise is at the seed stage, at a higher valuation, where you give away less equity. On paper you're securing your value by putting your money in now instead of later.

For a first tech company, that was probably a bad idea, and that's the honest lesson. I underestimated how much things would cost and how long they would take, by a wide margin. I'd planned to bootstrap on around a hundred and twenty-five thousand, and by the time I was six hundred and fifty thousand in I wasn't anywhere close to where I needed to be, which is when I knew I had to start raising. If I were doing it again as a first-time founder, I'd be a lot less ambitious about it. Be much more efficient with the capital, get to launch faster, get to revenue faster, and don't bet the whole thing on a vision you haven't pressure-tested with the market yet.

I don't think there needs to be a blue ocean. We're in the early phase of a massive transition in the global economy, which means a massive opportunity for people to make their existing service offerings intelligent, to productize their services, or to automate them. It doesn't matter that a thousand people are doing the same business, because some serve a certain region, some serve a certain industry or business type, and there'll be a lot of niching down before the wider players consolidate through mergers and acquisitions. What matters is: can you do something that's good, and can you convince people to pay you for it.

The part a lot of people miss is the automation middle layer. Everyone jumps straight to AI, but to actually use AI in an existing business you first have to automate aspects of it so that data is being generated, flowing correctly, and labeled correctly. Only then can you turn it into something intelligent. So the real questions to work through are what can I automate, what should I automate, where do I genuinely need a human in the loop, and where can it be fully automated. Get the automation foundation right first, because if the data underneath isn't clean and flowing, the intelligent layer on top has nothing solid to stand on.