Building AI Tools to Solve My Own Problems
How a $100/month subscription became my personal software factory
One of the best ways to embrace AI is to build tools that solve your own problems, because your motivation tends to be higher than when you’re solving someone else’s problem — say, by following a YouTube tutorial.
With a subscription of around US$100/month, AI is more than capable of building a simple web application out of the box. I haven’t tested its limits yet, since my problems can usually be solved in fewer than 10 steps and the business logic isn’t complicated. Here are some examples where I used AI to develop, test locally, and deploy to Google Cloud:
Expense tracker — parses bank statements with an LLM. You upload a statement and get insights back.
SG Property Monitor — pulls data from URA to track the price index, resale index, rental trends, land sales, and new-launch take-up rates.
US Metrics dashboard — pulls online data to track key fiscal and economic indicators.
Digest app — monitors RSS feeds and newsletter subscriptions, then sends a digest every Sunday.
Poker Trainer — helps me memorize preflop call/fold ranges.