“Your process shapes your product. Build sloppy, ship sloppy. Build with intention, ship with confidence.”

Question 01

What made you start what you're building?

I've been a product designer for about fifteen years. In that time the thing that has killed the most projects — mine and everyone else's — was never the code. It was building the wrong thing. Beautiful execution of an idea nobody wanted. Now AI agents can ship code faster than any of us can think. That should be a gift, but it's actually made the old problem worse. When shipping is cheap, building the wrong thing is cheap too — so people do it at ten times the speed. They vibe-code their way to a polished MVP that solves a problem that isn't there, then wonder why nobody shows up. So I started building the layer that's missing. LaunchChair is an AI MVPer and a living product spec — it turns your market strategy into a spec-driven agent loop before a single line of code gets written. ICP clarity, competitor and substitute mapping, positioning, pricing, MVP scope. Then that same spec drives the build cards your coding agent works from, and the launch pages and SEO after. One document, alive the whole way through.

Question 02

Why three products instead of one?

Because the pain doesn't stop when you ship. It just changes shape. LaunchChair fixes the front of the loop — market research that's never thorough enough, MVPs scoped by vibes, prompts that lose the plot halfway through a build. But then you launch, and two new problems show up. The first is the distribution gap. You get one day of attention on launch day and then you fall off a cliff. BuildHop exists for that — it turns product discovery into an ongoing loop instead of a single spike. Live leaderboards ranked on real attention — votes, watch time, momentum, repeat visits — so you stay visible long after launch day. The second is that half of indie launch pages are quietly broken. Missing meta, no schema, security headers wide open, accessibility a mess, a hero that doesn't convert. FreeScan is the free audit for that — paste one URL, get 36 checks back as fixes, opportunities, and insights across SEO, AEO, GEO, security, accessibility, and design. Now it even hands you non-destructive prompts to fix each thing.

Question 03

The stack, in one place

How the three products chain into one loop:

1. 01LaunchChair — validate the market and turn it into a living spec + build cards
2. 02Agent API / MCP bridge — let Claude, Codex, or Grok run the spec directly, no manual prompting
3. 03Starter boilerplate — customer-owned Next.js + Supabase repo with auth, SEO, and security scaffolded
4. 04Launch workflow — the same spec becomes landing pages, schema, and SEO/GEO/AEO content
5. 05BuildHop — stay visible post-launch with ongoing discovery and dofollow exposure
6. 06FreeScan — audit the live page, catch launch-breaking issues, get fix prompts

Question 04

What's the one belief the whole thing is built on?

Your process shapes your product. Build sloppy, ship sloppy. Build with intention, ship with confidence. That's not a slogan I put on a landing page — it's the reason the spec sits at the center of everything. A messy, undocumented, prompt-by-prompt process produces a messy product. A structured one produces something coherent. The spec is just intention made durable, so your agents inherit it instead of guessing. The rest of what I believe follows from that. Talk to users before you build, not after. Launch before you feel ready — you never feel ready, and the feeling isn't information. Distribution beats product; a worse product that people can find beats a better one nobody sees. Boring consistency wins. And your first ten users are everything — they're not a metric, they're the people who tell you whether the thing is real.

Question 05

What's the biggest mistake you see indie founders make right now?

Treating the AI as the strategy. The agent is unbelievable at execution and completely indifferent to whether the thing is worth executing. So people point it at a vague idea, get a working app back in a weekend, and mistake speed for validation. The fix isn't to slow the agent down. It's to give it a spec that already encodes who it's for, what it competes with, and why anyone would switch. Then the speed is on your side instead of against you.

Question 06

What rule have you kept that's paid off?

Boring consistency. Show up, ship a little, talk to a user, post the thing — every day, whether or not you feel inspired. Most of the founders I watch flame out aren't short on talent. They're short on repetition. They want the launch-day dopamine and none of the quiet Tuesdays that actually build an audience. The quiet Tuesdays are the whole game. That's why I built a distribution product instead of just telling people to 'do marketing' — visibility should compound in the background while you keep building.

Question 07

What would you tell yourself twelve months ago?

Ship the ugly version and start distributing on day one. I spent too much of my career polishing before anyone had ever seen the thing, because I'm a designer and polishing feels like progress. It isn't. It's hiding. The spec-first stack I'm building now is basically a letter to my past self: figure out if it's worth building, build the smallest real version, and get it in front of people the same week — not the same quarter.