If you are developer or some sort of IT engineer, don't automatically think your job is not BS. Like job of developing thin client to a windows server running licensed software to get around license costs of software, where price is based on number of users... Yes, this types of businesses generate TONS of money from essentially the fact, that supplier and customer could not get in agreement for more reasonable licen…
I'd argue (as would most of the free software community) the creation of proprietary software is entirely bullshit. Because you have to eventually reinvent it with either permissive or copyleft licenses to just avoid the huge black hole that is legal bullshit and bureaucracy around licensing rights to a number that is not even scarce once its made. And until it is foss, you are just denying the users their freedoms t…
So while "proprietary" (i.e. closed source or limiting who has access to it) may not be the precise description, a LOT of software is (and will be) "in-house" stuff. This is because while commodity solutions are great in many way, the are general tools, and will never map exactly to the current task and needs of a business.
Just like how the plumber is happy to buy common solutions to piping problems (e.g. standard sizes, common specialized parts that solve problems that show up all the time), a software engineer can pull from Free Software as his "standard parts". In the end, though, because every need is at least slightly different, a business will always need to hire somebody to put those parts together - just like they hire an in-house plumber or other specialist.
A consequence of this, though, is that all that specialized work is not going to be particularly useful outside the place where it was originally used. The occasional useful features can be pushed back upstream, but that still leaves a lot of glue around.
As time goes on, and the collection of "standard tools" grows, I expect the ratio of "glue" to "useful new feature" to increase.