Sales. Some people can make it rain and generally nobody begrudges paying them well to do so.
Mathematics. Terence Tao is as far ahead of the average Math professor as they are of the average high school Math teacher.
Law. If you give the average lawyer twenty times as much time and all the research resources of a partner at Skadden Arps the Skadden guy will still come out on top.
Finance. Whether you’re talking VC, PE or Hedge Funds the people at the top are a lot more than ten times as good as the average, never mind the marginal member of their reference class.
All of academia as far as research is concerned. Research impact as measured by citations is a power law. People at top departments routinely get hundreds of times the citations of those at no name ones.
In any human endeavour that isn’t limited by severe bottlenecks in material input you get people who can do much more with much less than others can, and who are capable of doing things that those less skilled simply can’t do.
The big difference is maybe that bad vehicle technicians are singled out by the boss or customers and quit? While neither the boss or customer can pinpoint a single programmer.
There's a freedom that comes with that, we're unbound by physics - layer upon layer of abstraction make it a lost concern.
It feeds into itself, growing in power at a rate proportional to its surface.
Until that is, it collapses under its own weight. When that happens is a matter of structure, and the extent to which the developer understands what the structure is, where it should be and how it must expand to cater for new requirements.
For simple cases a single largely linear block of code will be fine.
Eventually things get more complex. You'll need to move the lower level functionality to another file and call through to its methods.
Some of the methods in these files will become so heavily parameterised - and with repeat parameters - that you might want to introduce state. Welcome to OOP.
Then with the proliferation of the files you've created you'll need to introduce a directory structure.
Then a module structure.
Then a separation of API and implementations.
Then a division across projects.
At any of these points some will falter.
Some can't even write a single class without copying a pasting massive blocks of code. Some write class after class without consideration of abstraction or structure.
Yes there are 10x developers - those who understand structure and can implement the code accordingly.
Why does this happen in software and not other industries? As I said, we're unbound by reality.
If the best of us were engineers, we would be more tightly constrained by the physical limitations of the world.
If the worst of us were engineers, they would get people killed.
Finance is another industry that supports 10x employees, its common for traders with a natural ability at pattern recognition to produce 10x - 100x returns. Like technology, their minds can be leveraged by increasing the amount of capital they can manage.