There are some good points here but I wonder if we aren't missing the deeper problem. Innovation is not required at any level in modern software development. This essay is, not to put too fine a point on it, a summary of my book Famine in the Bullpen about the death of innovation and the reasons behind it:

https://medium.com/marker/the-death-of-specialized-skill-a3c69029f693

When I was at Sun MicroSystems, we still believed that innovation would yield value and profit. I was on the Architecture Review Committee which reviewed all Sun projects insuring that they all conformed to Sun standards. When, during the dot-com bubble collapse, Sun management seized control from Engineering, innovation died. Agile became the new religion and everything became simple procedure.

I remember with sad convulsing nostalgia sitting with other engineers around a table cluttered with drawings and code. The problem was impossible. No one anywhere had ever done such a thing. An engineer would suggest a solution and objections would tamp it down. Suggestion and rejection. Suggestion and rejection. Everyone was despondent. The impossible problem must be addressed.

Finally someone would stand up and say something to the effect of, "I'll take it."

Someone at that table had had a revelation. That person's unconscious mind had made a connection. It had formed a network around the problem and found a solution. That person did not yet know what the solution was but the signal had risen and sent a ping from the unconscious mind to the conscious mind. The bold engineer left the meeting and returned to the dimly-lit office to ponder -- what have I done?

So many times I was that bright-eyed engineer. In my office, terrified at my promise of a solution, I would scribble on my whiteboard. Mutter to myself, jazz playing in the background. Sleeping, eating, taking long walks, returning to my office, scribbling for sometimes weeks. Sun allowed this because Sun understood how innovation and true agile worked.

Then one day, deep in meditation, scribbling and erasing, there would be an ache. A real and deep-in-the-brain ache. Staring puzzled at the latest scribbles, the ache would throb. Unconsciously, I would turn off the jazz. In the silence, the ache would subside. The creative (music-loving) side of my brain had been released. I would stare in awe and suddenly the astounding joy as I fully understood the problem and the solution. Everything was clear now. I, like so many other Sun engineers, experienced the astounding joy of the engineering solution.

Next would be the writing of the Functional Specification and submission to the ARC. Then the project and finally the product.

Sun abandoned that approach and turned to Agile and existing open-source solutions. That wasn't the full reason for their failure but it played an important role. Now, open-source has been largely infected by twisted architectures from HP, Apple and IBM. Linux is a pale ghost of its former self. My servers running OpenSuSe 13 are indestructible. They never reboot and always provide the required services. My ultramodern Debian 12.8 systems require a reboot every few weeks. Why? Not sure. I'm busy writing and conducting public readings of the Constitution because our politics has been overtaken by madness.

One day I hope to look into why Linux is now crap but for now, the bigger problem is that innovation has been killed by capitalism. We need to fix that.