This is exactly the answer. I've worked on one very old codebase in my career. I was paid well for doing so.

The graph of investment in software starts high with greenfield projects, drops over time to an all-time low as optimizations take hold, and as the software continues existing will rise to the greenfield (or more) levels of cost. The interesting thing to me is what substantiates the rises; for instance, in greenfield experimentation and iteration is what costs the most. In old software it's usually the employees.

I'd love to actually document this sometime.

You won’t get a million per year from my revenue since I’ve built it and I’ll still be there come hell or fire; but you can get 10% at the beginning and I don’t hesitate to increase employees as they maintain the existing one or contribute to increasing the revenue. In fact I generally give them more than they contribute, and I’ve increased my employees 30% this year in average, unprompted.

But there is always this HN comment with, no matter how much you pay, they put exclamation marks about how little the compensations are in France and how 35hrs is too long or having to be in office a few days a week is the bane of their existence. I’m always wondering whether we’re unionized here, it looks like a systematic complaint, no matter the conditions.

Well, I don’t know, create your own company I guess.

You missed the point. You can't recruit because you aren't paying enough to be competitive.

So what is your salary range?