Once a product launches on PH, most of us obsess over upvotes and comments but increasingly, people are discovering products through AI chat answers, not just search or PH itself.
Curious if any founders/marketers here are actively tracking how (or if) their product shows up when someone asks ChatGPT or Perplexity a relevant question. Is anyone doing this systematically, or is it still mostly guesswork?
"build it and they will come" is the most dangerous startup advice for first-timers. sell it before you build it. then build. I posted an idea with zero product behind it. 300+ people said yes in 72 hours. that's when i started building. Did anyone else build first and wish you'd checked demand earlier? or am i just late to this realization...
Mine: I m building a guitar practice tool, and I can t legally show my own app working. Every screenshot has copyrighted lyrics in it.
Spent three weeks working around it cropping, blurring, avoiding the best feature. Then realised the actual fix was to just write two original songs and own the demo files outright. One afternoon, 3 week problem gone permanently lol.
Every maker I talk to has one of these. The thing that looked like admin turned out to be the constraint shaping everything.
What was yours
When I first started using AI, I treated it like any other online tool. I was cautious about what I shared and assumed everything needed to be anonymized.
Today, my first instinct is often to open ChatGPT or Claude before I even search the web. I use them to brainstorm ideas, review code, rewrite emails, summarize documents, or think through problems. They've become part of my workflow rather than just another website.
Has anyone else had a bug where the code runs with zero errors, returns a totally normal-looking response, and just does nothing? Spent hours tonight tracing something back to a missing database permission that silently blocked a write instead of throwing a crash, no red text, just quiet nothing. What's the quietest bug you've ever chased?
Not the big obvious stuff. The thing that is quietly building in the background while you are dealing with whatever is loudest today.
For me it used to be a slow-paying client in one business or a staffing gap in another, the kind of thing I would only notice once it already cost something. Not because I did not care, but because I was looking at whichever business was yelling the loudest that morning, and the quiet one does not yell.
A few thousand views with single-digit profile visits is a stranger result than either zero views or a normal click rate. Something is working and something isn't, and I can't tell which end.
Numbers from the last seven days on a short-video account: a few thousand video views, four profile visits, zero comments, zero shares. On a longer-form channel over four weeks: five thousand views, four and a half hours of watch time, one new subscriber. People are being served it and they're watching it. They just don't do anything afterwards.
Three explanations, and I've argued myself into each of them.
The copy is the problem. My captions are written to sound like a person instead of an ad, and I think they've ended up so soft that they never actually ask for anything. Testing this means writing harder captions, which risks the reach that currently works.
I strongly believe the first 1,000 users should be chosen for learning, not reach.
At this stage, raw signup volume can hide more than it reveals. What matters is whether users experience the problem often, try the product in real workflows, and give feedback that leads to better onboarding, positioning, and product decisions.
As entrepreneurs, we spend so much time hustling, agonizing, organizing, and pathologizing. The burnout rate for founders is high. When you finally launched your product or service, how did it make you feel?
When we launched our beta (not on PH, but to our beta group), I felt peace. Finally, I could rest my mind on problem-solving based on facts instead of stressing over things that weren t real. I m curious to know how your mental health changed post-launch and what you re doing to keep yourself sane in the meantime.
Every credit system I've used bills on the API call, not the outcome. So when a model returns something plausible and wrong, the user pays, then pays again for the version they actually wanted. You've priced your own failure rate and handed the bill to the person who noticed it.
The counter I keep hearing is that inference costs money either way. True, and beside the point. If a generation is unusable because we routed it to the wrong model, that's our decision, not their prompt.
Two things fall out of eating it yourself, both good. You have to define unusable narrowly enough that it can't be gamed, ours is a same-prompt regenerate inside a short window and nothing else. And you start caring a lot about routing, because a bad route now costs you directly instead of quietly costing someone else.
Curious who's actually doing this and where it broke for you. Most pricing pages I read still bill per attempt and call it usage.
