I was thinking about Codex vs Claude Code.

I see the world bickering over which one is better.

Yet I don't see enough people making any money using either.

Potential is in selling a service that you would not be able to do on your own.

For example, clean a database.

Make a website for someone.

Create a prototype.

Automate something.

Go out there and find where you can get paid to do something.

Then buy the subscription.

Or stay stuck in the loop of releasing what everyone else is releasing.

The only ones actually making money using Claude or Codex are the experienced programmers who improved their workflow, and AI gurus trying to convince you that "You can sell some type of AI nonsense as well". 

As if it's that easy.

Before AI, 50% of all built products solved nothing useful. Now it's more like 99.9%.

Pick the right problem. Ask yourself if you should actually build it. Devote your time wisely.