The bottleneck flipped and most indie hackers we haven't noticed

Five years ago, the hard part of shipping a product was the code.

You had an idea. You understood the problem. Maybe you even had users telling you exactly what they needed. But turning that into a working product took weeks or months. You got stuck on auth, on database schemas, on deployment, on that one CSS layout that refused to behave.

And because code was the bottleneck, the entire indie hacker culture optimized around it. Remember the endless debates? Rails or Next.js. Firebase or Supabase. Should I learn to code or use Bubble. Every other post was "what's the fastest stack to build an MVP" .. as if picking the right framework was the thing standing between you and $10K MRR.

No-code tools exploded for exactly this reason. Bubble, Adalo, Glide . their entire pitch was: skip the code, get to market faster. The community treated implementation as the enemy. If you could just get past the coding part, surely the users would come.

And that framing made sense at the time. Building was genuinely slow. The gap between "I want this feature" and "this feature exists" was measured in weeks. So we obsessed over closing that gap for the MVP phase.. faster frameworks, simpler stacks, no code at all.

Then AI closed it for us.

Today you can go from idea to working MVP in a weekend. Cursor, Claude,Codex.the code generation tooling for building MVP's is so good that the implementation bottleneck is essentially gone. The great stack debate is over. It doesn't matter anymore. You can build anything in anything, fast.

And that should be great news. Except something weird happened.
Instead of spending the freed-up time on understanding the problem deeper, most of us just started building faster. We ship MVPs in 48 hours for problems we spent 30 minutes thinking about. We build first, validate & understand the first principles later . We have a working product and no idea who it's for.

The bottleneck flipped. It used to be: I know what to build but I can't build it fast enough. Now it's: I can build anything but I don't know what's worth building.

Five years ago, the typical failure was running out of energy mid-build. A half-finished app, an abandoned repo, a framework migration that killed momentum. The founder understood the problem but couldn't ship the solution.
Today the typical failure looks completely different. A kinda fully working MVP app with a polished landing page, clean UI, proper auth ....... and zero users. The founder shipped in a weekend and spent the next six months wondering why nobody signed up. The product looks done. The thinking never started.

The no-code crowd from 2021 and the AI-vibe-coding crowd of 2026 made the same bet ... that removing the coding barrier would unlock success. It didn't. It just revealed the real barrier that was hiding behind it: understanding the problem deeply enough to build something people actually need.

The founders winning right now aren't the fastest builders. They're the ones who got slower at starting and faster at finishing. They spend weeks in conversations, forums, and customer calls before they write a single prompt. They know the exact workflow they're inserting into. They know the objections. They know the budget. By the time they sit down to build, the spec writes itself because they actually understand the problem.

The code was never the product. It was always just the delivery mechanism. We just couldn't see that clearly when it took three months to deliver.

Share your story .... what does your split actually look like right now? How much of your time goes into building versus talking to users? And if you shipped something fast with AI, did it change how you think about validation? Would love to hear what's working and what isn't.