A lot of software that was once innovative is now mature. 20 years ago, the bar for what you could sell was really low, because there just wasn't a lot of soft

A lot of software that was once innovative is now mature. 20 years ago, the bar for what you could sell was really low, because there just wasn't a lot of software available, full stop. If you had something that was even a little bit useful, somebody would probably want it. People who bought software from you once were also eager to pay you for updates, because annual improvements could be dramatic. So what happened? Features that were once unique selling points of one commercial application were copied by other vendors and by open-source implementations. Clever techniques were studied carefully and are now well understood. Smart people looking for a challenge would study every publicly available implementation of some algorithm and then come up with one that beat them all. But why did the rate of software improvement seem to slow down? Don't we have more people working on software than ever before? Don't they have better building blocks than ever before? Consider that everything that has ever been published as open source remains in the public commons forever. This body of work grows without bound. Both improvements and regressions continuously show up, but (eventually) the improvements are identified and the regressions are cast aside. So this pile of software theoretically improves without bound as well. Yes, we have more people writing software than ever before, and yes, they have better building blocks than ever before. But they're not competing with other software developers and other companies -- they're competing with every piece of software ever written by anyone. So it shouldn't be surprising or controversial that the bar for marketable software is going to keep getting higher forever. These licensing shenanigans and the steady march towards "Everything as a Service" is a result of companies pivoting from attempting to innovate to rent seeking.