I'm not a developer by trade. I'm a basic upper-middle management guy with an MBA, that works a cyber project management job by day. So take this as "here's what worked for me," not expert advice.
When I rebuilt an old trading script into a real app — a local tool that runs AI agents to test trading strategies and write up the results — I made one rule for myself: keep it as simple as humanly possible. No extra software bolted on. Just Python and a single file to hold the data.
Why
Because I'm not an engineer. Every extra piece of software you add on is one more thing that can break — and one more thing I won't know how to fix at 1am. The more the app leans on other people's tools, the more likely I am to get stuck. So I kept the moving parts down to almost nothing.
The upside
- Anyone can run it as long as they have Python. No complicated setup.
- All the data lives in one file, so backing it up is just copy and paste.
- Nothing to update, nothing to clash, a lot less that can go wrong.
The tradeoff
I had to build a few basic things by hand that ready-made tools would have done for me. More work up front, and a real engineer would probably roll their eyes at some of it. But for a solo, non-expert like me, "boring and reliable" beats "fancy but fragile."
My honest question
For the actual developers here: when does "keep it simple, no extras" stop being smart and start being stubborn? For me it has worked — the thing just runs. But I'm sure there's a line, and I'd like to know where you'd draw it.
(If you're curious what the app is: a self-hosted paper-trading tool — practice money only, nothing leaves your machine. Demo, no signup: https://premium.deadcatfound.org/demo. But I care more about the "keep it simple" question than the app.)
Top comments (0)