an LLM-coding post-mortem: PicPickr

It’s been six months now. Data, backend^°hard word, frontend^°hard word, Electron^°hard word, CI^°hard word, testing, landing page, design, Supabase^°hard word, Resend^°hard word, Lovable^°hard word… the list goes on. What I thought would be a quick sprint^°hard word turned into a marathon^°hard word, and the todo^°hard word list still doesn^°hard word’t seem to end.

My plan in the beginning was simple: build something quick and dirty, and share it with other people. Something about that felt very human, very exciting^°hard word. So I whipped^°hard word up a web^°hard word app overnight^°hard word using Cursor^°hard word. Then^°hard word as I thought more about the problem, details emerged. I learned the hard way that web^°hard word apps don^°hard word’t play nicely with local^°hard word file^°hard word systems, especially when they need to handle thousands^°hard word of large photos.

I then^°hard word thought maybe an electron^°hard word app would help with my objectives^°hard word. Again, I think I gave AI^°hard word too much credit and dove^°hard word right in. I realized the real work had just begun. Packaging^°hard word, distribution, signing binaries^°hard word — all those invisible^°hard word details that make something feel like a product^°hard word. The funny part? I discovered these only toward the end, totally a surprise (not a good one, still waiting for my apple developer account approval) after months of building.

Even the landing page became its own project. I used Lovable^°hard word to generate one, and soon enough, I was knee-deep^°hard word in building flows^°hard word for email signups^°hard word, testing Resend^°hard word integrations^°hard word, tweaking^°hard word copy, fixing layouts^°hard word. It became another app in itself. If I were to do it again, I’d probably try something like Framer^°hard word. I can’t even count the number of prompts^°hard word I spent^°hard word just refining that one page.