NEXT U goes live here tomorrow (July 30). Before that, one decision I'd like someone to argue with me about.

The interface ships in English, , and . Four languages, solo maker, zero paying customers. Every launch guide says pick one market and go deep.

My reasoning: I'm in Taipei. My first users are far more likely to be in Asia than in San Francisco, and an English-only tool is a hard sell to a shop owner in Taichung or Osaka. Localising later means retrofitting it into every component, which is worse than doing it up front.

The counter-argument I can't answer: four languages means four sets of copy to keep correct, and I still have no idea which one converts.

Every AI tool I admire ships a blank prompt box. I shipped a wizard instead: product photo copy images video + music finished MP4. Four fixed steps.

The trade-off is real. Power users hate being walked through things, you lose the "type anything" magic, and every step is one more place to drop out.

I built it this way after making ads by hand for a friend's e-commerce store. The bottleneck was never the tools it was that we didn't know what the ad should be until we'd already made three bad ones. A blank box doesn't help with that. A wizard at least makes the decisions in an order that works.

Still might be wrong about this.

I went with credits for NEXT U (AI ad generation): 1 credit = $0.01, no monthly fee, and you're only charged when a generation succeeds failures refund automatically.

The honest downside: metered pricing makes people hesitate before every click. A subscription makes them forget they're paying, which is better for my revenue and arguably worse for them.

I picked credits anyway, because I'm the guy with four AI subscriptions he used twice last month.

But I keep second-guessing it. If you're paying for AI tools right now do you actually prefer per-use, or do you just say you do?