[toddh]

You got the gist of it. Then it just kind of took off.

The TLDR:

Where it started: me standing in front of a long aisle of dog food bags at Costco, racking my memory for which kind I was supposed to buy next for my finicky golden retriever, Kilo.

How it ended: me building Rotation List, so nobody will ever have to forget again.

I explained more in:

https://www.indiehackers.com/post/ai-killed-the-mvp-whats-ne...

With an AI assist, there was no reason to stop adding improvements as I thought of them, so I didn't.

[vhcosta]

> With an AI assist, there was no reason to stop adding improvements as I thought of them, so I didn't.

That is the mantra of today, yes. In a similar vein, I made my initially personal C course into something I could make public-facing, with some refinements, I also kept adding features but them my "sanity"(?) kicked in and i squashed a couple to be only in dev mode. Perhaps they will return to the public version after a round of refinement.

[toddh]

The thing is, it's real. Or it was for the golden era of cheap tokens. When tokens get expensive, we'll see how that changes things. With AI, you can think in terms of features rather than implementation; that's a totally different way of working.

I first programmed in C on a PDP 11 many years ago. I'm curious what you mean by "public vs. dev mode"?

[vhcosta]

I started 5 years ago via a programming bootcamp here in Portugal. After finishing it, I dreamed of being able to make my own software, and solve all my programming and life problems in one fell swoop. Then the AI boom happened, I got laid off after 2 years of working at a software and consulting co., and swallowed my dream of making enough money to buy a house (lol).

Currently working in a place where I actually need to talk to people and do real life stuff: a store clerk. Not glamorous, but it allows me, in my own time, to tinker with the things i know, learn new things, and more recently, vibe-code stuff and release at a pace I've never worked before.

What i meant by the "public vs dev mode" is: i gated a feature or two behind only being present if im running my app locally (via npm run dev or what have you), because i was insatisfied with how that feature is currently implemented. I intend to refine it at a later point.

I did that because, even in the age of thinking "in terms of features rather than implementation", i am trying to be selective, and find a headspace where, if i feel that a certain aspect of a project (be it fully/partially vibe coded or not) is not ready to be presented in the live app, then it does not need to / should not be in prod. Could be that i was aiming too high and the feature was lacking, coule be my prompts that were misguided, or could be feature creep making me nervous.

I both welcome and fear the age where tokens get vastly expensive. I also hope for the time where we can run proper local models that don't require me to live in a data center.

Sorry for the long comment. It's my own writing at least. "if i had more time, i'd have written a shorter letter"