Non-developers are writing apps instead of buying them | Hacker News Reader

Non-developers are writing apps instead of buying them

techcrunch.com

For now, there is a lot of novelty at work, AI and being able to make computers realize ideas for the first time.

Let's see if it holds up in the long run. Building a prototype that makes feel-good hormones run is one thing. Slogging, or slopping, through the 20 remaining pareto points to a more polished experience is another thing. I suspect people will go back to full featured apps they don't have to constantly maintain and update.

I'm sure someone will chime in AI will solve all this, but I'd pre-counter by saying that the inherent limits of LLMs and agents based on transformers preclude this reality from materializing

The counter isn't that some hypothetical AI will come along and solve three problems with the current generation. That argument, and the response, is tired as dirt. No, the counter here is to point out that the prototype is fine for small values of production. I wouldn't take something Ai generated and serve it on the web, and have the world try and hack the site, but some small app that isn't exposed to the world? We can throw around claims of security and scalability, but Amy's (local) personal cookbook app doesn't care about either of those.

"Amy's (local) personal cookbook app" sounds great in theory, and that people will do this for the novelty, but my point is that this will not be a long-term trend. Largely because people will want more features in their apps beyond the prototype, they will find it hard and decide they'd rather be doing the cooking than cajoling, and go back to using the same app their peers do so they can be social and share, or even just send the recipe from their phone to the tablet on the kitchen counter

Humans are largely emotional creatures. "I made this" is enough of a driving force on the small scale, and I want the button here, and not there. No app developers since winamp have offered that level of customization at their level. There will always be people who refuse to use technology, and others that dive in deeply, but most are in the middle. So if it's 20 mins a week after an initial push to add a feature and send it to TestFlight or Dropbox an .APK then I see that being way more common that you believe.

Reply on news.ycombinator.com