Everyone loves the cliché that being a founder is like jumping off a cliff and building a plane on the way down. It's not wrong. But it’s incomplete. Try doing it while juggling flaming torches, a live alligator, and the crushing weight of your investors' expectations. The ground gets closer, fast. Forget the glossy magazine covers and the unicorn valuations for a moment. I want to show you the real story, the one from the trenches where the code gets written, the deals get made, and companies are actually built.
I skipped business school. My real education was forged in the Silicon Valley trenches. After two exits—RemoteTeam to Gusto and MovieLaLa to Gfycat—and writing checks to over 200 startups like Anthropic, OpenAI, and Scale AI, I’ve learned a thing or two. I’ve made a comical number of mistakes, but each one was a lesson that cost me real skin in the game. This isn't a textbook theory. This is a raw, unfiltered look at a single week in my life building an AI-first product, and the five hard-won lessons that came with it.
Monday: The Myth of the “Fresh Start”
Forget the calm meditation and green smoothie. My Monday started with a fire. A big one. A key third-party API we depend on decided to fall over. At 7 AM, my phone was already melting with alerts. Our main dashboard was a wall of red. For an AI product, data isn't just important—it's oxygen. No data, no product. Our models were churning out nonsense, and our customers were, quite rightly, starting to get pissed.
The old me, the engineer, would have dived straight into the code to fix it. I had to beat that instinct out of myself years ago. A founder can't be the hero firefighter; you have to be the fire chief. I pulled the engineering and product leads into a call. Triage time. How many customers are hit? What's the blast radius? Can we failover to the backup provider? The team was already scrambling, but they needed air cover and a clear decision, not another cowboy coder.
It threw me back to the early days of MovieLaLa. We built this beautiful, complex model to predict box office hits. It was my baby. Then one week, it just went completely haywire. We burned days debugging the algorithm, thinking we'd made some fatal flaw in our logic. The real cause? Our main data source had changed its HTML structure without telling us. Garbage in, garbage out. A painful, expensive lesson: the most brilliant AI is still at the mercy of its data pipeline.
Lesson 1: You’re not the hero, you’re the strategist. Your job is to zoom out, direct resources, and clear roadblocks so your team can do their best work. Resist the urge to dive into the weeds on every single problem.
Tuesday: The Lonely Work of Deep Work
With the fire contained (for now), Tuesday was a sacred day: deep work. As a founder, this is a luxury you have to fight for. I blocked off a four-hour fortress on my calendar, killed all notifications, and physically moved my phone to another room. The mission: nail down the product spec for our next big bet—a new generative feature that personalizes the entire user experience.
Building an AI-first product isn't like your typical SaaS playbook. You're not just mapping out user flows and button placements. You're wrestling with a fog of uncertainty. Is the model even going to be accurate enough? What happens when it goes off the rails and says something bizarre or offensive? How the hell do you design a UI that explains the fuzzy, probabilistic nature of AI to a normal human being without giving them a headache?
I spent hours whiteboarding, dI spent hours at the whiteboard, lost in a forest of decision trees and user stories that sounded more like philosophical koans. 'As a user, I want the AI to just get what I mean, even when I suck at explaining it.' That's not a feature, that's a freaking research paper. For me, writing is thinking. By the end of the four hours, I didn't have a perfect spec. But I had a much sharper, more intelligent set of questions for the team, and that's half the battle.f questions for the team.
Lesson 2: Embrace the uncertainty of AI. You can’t A/B test your way to a breakthrough AI product. You have to make opinionated bets, build prototypes, and be prepared to be wrong. Your job is to define the right questions, not to have all the answers.
Wednesday: The Art of Selling the Future
Wednesday was investor day. Back-to-back calls with a few of our angels and a check-in with our lead VC. When you're building something truly new, you're not selling a product. You're selling a time machine. You're selling a ticket to a future that only you can see clearly.
I've sat on both sides of this table. As a founder, I'm the guy with the crazy dream asking for the cash to make it real. As an investor in companies like Hugging Face and Replit, I'm the guy betting on other people's crazy dreams. The one constant? Conviction. You have to believe in your vision with a borderline-delusional intensity. It has to be infectious.
One of the calls was with a newer angel who was starting to sweat about our burn rate. He saw the AWS bill for our GPU cluster and the word 'profitability' started getting thrown around. I didn't bother with a spreadsheet. Spreadsheets are for boring companies. Instead, I told him a story. I walked him through the exact user experience of the feature I'd been wrestling with on Tuesday. I showed him the mockups. I made him feel the magic. I reminded him we're not building another damn CRUD app. We're building a new form of intelligence. That burn rate isn't a cost. It's the fuel for the rocket ship.
Lesson 3: Sell the story, not the spreadsheet. Investors are people. They’re moved by stories and conviction, not just numbers. Paint a vivid picture of the future you’re building, and they’ll follow you there.
Thursday: Culture is What Happens When You’re Not Looking
I started RemoteTeam on the simple belief that talent is distributed globally, but opportunity isn't. We were 100% remote from day one, way before it was trendy. But building a real culture when you can't bribe people with free lunch and ping pong is tough. You have to be deliberate about it.
Thursday is for my one-on-ones. I noticed one of my best engineering managers was off. He was quiet, withdrawn. I poked a little. 'What's really on your mind?' He finally admitted he was fried. Completely burnt out. He was carrying the weight of the Monday fire and felt like he'd let his team down.
We talked for a solid hour. I didn't jump in with solutions or cheap platitudes. I just listened. I told him about the time I almost flamed out at MovieLaLa, trying to carry the whole world on my shoulders. I told him his job isn't to be a shield for his team, but to be a coach. To give them the tools and trust to solve their own problems. By the end of the call, he wasn't just feeling better; he had a concrete plan to rally his team. He didn't need a CEO; he needed a human to listen to him.
These are the conversations that matter. This is the real work of a CEO. Culture isn't the bullshit you paint on the office wall. It's the sum of the conversations you have and the behavior you model every single day. It's about showing up as a person first, and a CEO second.
Lesson 4: Your team is your product. You can have the best AI models in the world, but if your team is broken, your company will fail. Invest in your people, listen to them, and build a culture of trust and psychological safety.
Friday: The Dopamine Hit of Shipping
Friday is for shipping. Nothing in the world beats the dopamine hit of pushing code to production and watching the first analytics events roll in from real users. This week, it was a small but meaningful tweak to our onboarding flow. Not a sexy feature, but it was a huge quality-of-life win for our new users.
We have a simple tradition: every time we ship, someone posts a celebratory GIF in the main Slack channel. It's a small, silly thing, but it's our moment. It's a reminder that we're all in the same boat, rowing in the same direction, building something that matters.
Post-ship, the rest of my Friday was classic 'CEO-mode.' Clearing the inbox, staring at spreadsheets, and prepping for the next board meeting. It's the janitorial work of being a founder, but someone's got to do it to keep the machine running.
When I finally shut my laptop on Friday night, I wasn't tired. I was wired. It was a brutal week. We took some punches. But we survived. We shipped. And we were one week closer to building the future we're all dreaming about.
Lesson 5: Find joy in the process. The founder journey is a marathon, not a sprint. There will be moments of despair and frustration. But there will also be moments of pure joy. The thrill of solving a hard problem, the camaraderie of building with a great team, the satisfaction of creating something new in the world. Savor those moments. They’re the fuel that will keep you going.
That's the job. It's messy, it's chaotic, and some days it's a straight-up fistfight. But I wouldn't trade it for anything. This is the greatest game in the world. And I can't wait for Monday.
Frequently Asked Questions
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.
Do all experts agree with this view?
No, and that's fine. The best ideas in business are often contrarian. I share my perspective based on my experience and data, but I encourage you to seek out opposing viewpoints and form your own conclusions.
How has this view evolved over time?
My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.