Comment by alwillis
> I use AI daily as a senior coder for search and docs, and when used for prototyping you still need to be a senior coder to go from say 60% boilerplate to 100% finished app/site/whatever unless it's incredibly simple.
I know you would like to believe that, but with the tools available NOW, that's not necessarily the case. For example, by using the Playwright or Chrome DevTools MCPs, models can see the web app are it's being created and it's pretty easy to prompt them to fix something they can see.
These models know the current frameworks and coding practices but they do need some guidance; they're not mindreaders.
I still don't believe that. Again yes a boilerplate calculator or recipe app probably, but anything advanced real world with latency issues, scaling, race conditions, css quirks, design weirdness, optimisation - in other words the things that actually require domain knowledge i still don't get much help with, even with Claude Code, pointers yes but they completely fumble actual production code in real world scenarios.
Again it's the last 5% that takes 95% of the time, and those 5% i haven't seen fixed with Claude or Gemini, because it's essentially quirks, browser errors, race conditions, visual alignment, etc etc. All stuff that completely goes way above any LLM's head atm from what i've seen.
They can definitely bullshit a 95% working app though, but that's 95% from being done ;)