I have a dream since March 2026. It's to create or join a startup. Over the course of 6 months, I have done many things to learn from the startup world. Create a startup "Le Steu AI", studying Y Combinator, startup funding process, learning how to manage projects. Finally, after 6 months, I got to join a startup as a technical co-founder. And after a week, I got to see and learn a lot.
First off, it was a mess. The product, an AI sales bot, was vibecoded and not working. There was no documentation, no PRD. There was no project management, so the launch goals were all over the place. So I came in, immediately created Trello and started gathering information. What the product was, its status, workflows, all consolidated into a knowledge base. I split whatever the vibecoded "PRD" is into multiple files for easier viewing (though I never looked at it, this was just for the coding agent to read). And within two days, me and another guy managed to turn it on, so there was hope for launch, huh.
Three days later, online communication integration and AI chatbot system were able to work, with the AI system being fixed by me with a lot of duct tape. I mean, literally duct taping the code, because it's so convoluted. OpenAPI description on top of Hono backend routes, with frontend using Orval generating TypeScript API clients (types + TanStack Query hooks + fetch functions) from an OpenAPI spec. For instance, on the frontend, there was a command file that called the same backend API directly, which confused me at the start and I was stumped at the time not knowing that it's deprecated (no documentation, you know).
In addition to that, we had a landing page, but it couldn't send email to the clients who registered to the waitlist. And at the time I left, they hadn't worked it out why yet. At least there were people coming in interested in the product, so that was a plus.
After working with the startup for a week, came in an offer. But I declined the offer. I made it clear to them that because all they could offer was equity only, I could only contribute part-time. I could never fully commit until there is sufficient funding for a salary, then I can go full-time. And they needed all hands on deck to rush a launch. Right, a launch while the product just barely worked, with no one comprehending the product because it was vibecoded without specifications, context to understand its creation. So I parted way, wished them good luck, and left the startup.
All in all, 10 out of 10. Would do it again. It's thrilling, and gives me motivation to wake up in the morning. But I would only do it if that position is part-time, and the right project comes along.