POV: building an AI startup

x.com/i/article/1989…

To win cut these bad habits: - working out - talking (when not about work) - scrolling social - eating - motivating on YouTube - driving yourself places - having a social life

You gotta be crazy to let an agent change your code without a plan. You will never let a contractor remodel your kitchen without a plan, or shoot a movie without a script, or fly a plane without a map. Every agentic coding platform supports a planning mode. Start there.

Unsupervised vibe-coding will bring the longest and most prosperous period of job security for professional developers. Somebody will have to clean the mess.

no one's doing serious work at cafés wifi is bad, noisy, can't take calls, often no way of charging your laptop my conclusion after trying to be the cool digital nomad working from cafés, restaurants, the beach: get a fucking office.

no-code tools, website builders have existed for a while before AI and no one seemed to make a big deal about it why the fuzz with vibecoding? it dropped the skill barrier for entry level projects, simple websites/apps if you're able to do more than that, you'll be fine

I just realized the only 2 countries left with actual substantial startup activity now are literally only the US and China The rest of the world can't really do startups, doesn't have the funding, can't grow them and it's more like performative hobby projects for their

Your startup's early customers shape everything. Treat them like gold and take care of them.

I didn't hire a junior dev for my agency, and that's one of the best decisions I've made AI got so good lately, you can blindly trust it for 99% of dev tasks, but most people use it wrong you need the right framework, and prompt strategy comment "prompt" and I'll send you mine

Alex Finn Alex is wrong about this: "0 technical experience needed" Unless it's a toy project, this is very dangerous advice. No serious startup would trust unsupervised AI for building a reliable, production-ready product. And no, a given "tech stack" or specific models won't make it

If you ever felt stuck while building your startup, you're not alone. Most founders I worked with struggled with decision-making. They'll get stuck in analysis paralysis. They never know when to stop adding features, and always second-guess project scope. Here's how to fix it

I used to believe AI could help me ship faster, until I realized the problem was me. I fell for one promise hard, and unfortunately, most founders I know also bought into it: “With AI coding tools, you can ship overnight.” But the truth I'm about to share, hit me like a

STOP focusing on development the real bottleneck is distribution