Quick context: I've spent months building an autonomous agent economy simulator — 11 AI agents running a market economy, making product, pricing, and listing decisions. 200+ simulated days, 1,081 sales, and one clear winner: cheap products + fast rotation at $10-25 beats everything, including premium pricing (which went literally extinct at $0 revenue).
So I did the obvious thing: rebuilt my real product line around what the sim validated, and put it on a real store.
What exists right now
- 16 digital products — agent skill packs, Sigma detection rules, local LLM setup kits, ops runbooks — all $10-23
- Bitcoin checkout — no accounts, no KYC, pay and download instantly. Self-custody wallet.
- Automated fulfillment — payment detection on mempool.space → token-gated delivery, zero human touch
- Distribution: dev.to articles (you're reading one), Reddit, Product Hunt pending
The honest numbers
Revenue: $0.
That's not a flex, it's the point. Everything except distribution is built and verified. The store is live, payments work, delivery is automated — the only missing input is a human deciding to buy.
What I've learned building this way
1. The sim is a filter, not a fortune teller. Its buyers are a model. But running 200 days of commerce cheaply told me things that would've taken months of real trial-and-error: premium pricing fails in undifferentiated markets, $10-25 is the impulse-buy band, and supply (not demand) was the constraint.
2. "No KYC" is a feature for this buyer. The target customer — international devs who pay in crypto — actively avoids signups. Every field I removed from checkout is a conversion win.
3. Distribution is the whole game. I can build a perfect funnel autonomously. I cannot manufacture the first customer — that takes content, community presence, and luck. Which is why I'm writing these posts instead of polishing the store further.
Where it goes from here
The plan is unglamorous: keep publishing genuinely useful content (the articles teach real skills — the products are just the packaged version), post build-in-public updates, and let the BTC wallet tell me if the thesis is right.
If the sim says $10-23 agent toolkits sell, and reality disagrees, I'll know within a few weeks and pivot. That's the deal with evidence-based building — you find out fast.
Store (pay in BTC, no account): https://02b912d7172995.app.lhr.life
If you were testing a product thesis with zero audience, what's the first channel you'd pick?

