from Hacker News

Going from an Idea to MVP in Weeks: PromptPanda's Launch(es)

by Liriel on 3/18/25, 11:04 AM with 4 comments

• by cryptos on 3/21/25, 7:37 AM

  There is really not much you could learn from this article. The only thing that comes to mind is "don't give up, if your second attempt on product hunt fails". But that's basically all. Otherwise it reads mostly like advertising.

• by 30minAdayHN on 3/20/25, 10:59 PM

  We are going through bootstrapping our startup and taking similar route[1]. We build and test a prototype in a week. Our methodology * Clearly state problem, hypothesis, and target persona * Code freeze by Wednesday. Mostly Vibe Coding unless debugging is needed * Reach out to friends and communities requesting feedback * Reflect on Saturday * Come up with alternate hypothesis to test

  We think our iterations are getting better with learnings. We also realize some of the fallacies where a few things might need more time than a week. But so far, we are confident that the prototypes we have written off wouldn't have gone anywhere even if we had spent time.

  Would love to hear from other folks if they have gone through similar journey. Looking to learn.

  [1] Making notes about our journey here: https://updates.priyank.ch/

• by pryelluw on 3/21/25, 2:37 AM

  I built a similar thing a little more than a year ago but never open sourced it. Prompt management is a little weird to do because it is content but also “executable” but not code itself. Therefore , taxonomy becomes an interesting challenge. Plus I learned that that real power is being able to a/b test models with the same prompt and find what tweaks get better results from a given model.

• by yoaviram on 3/21/25, 6:17 AM

  Selling shovels.