The US has always had a big regional divide in skill levels as well as respect for knowledge, with the out-of-the-way places having a problem with both.

The problem is political dynamics meant that the attitudes of the backward parts of the country have actually set policy for quite a while. That benefited American corporations in the short term - lots of CEO probably think for-profit, student-loan-driven education is a great idea. But it's harmed the US and even US corporations as these policies show their longer term results.

> The US has always had a big regional divide in skill levels as well as respect for knowledge, with the out-of-the-way places having a problem with both.

There are a lot of really good schools in flyover states. Looking at the top 10 engineering schools in the US, depending on your definition of "flyover," there's CMU, Michigan, Purdue, Illinois, and UT.

Yeah, even Alabama has high rated universities.

But it also has a broad problem of lack of quality of high school level education. Maybe a few good suburban school but the average is terrible.

Tuscaloosa is education center of Alabama, having the university of Alabama but secondary education is bad enough stop companies locating there.

"Around the late 1980s/early 1990s Saturn chose not to locate a plant in Tuscaloosa because of the performance of the public schools, and by 1993 the city leadership wanted Mercedes-Benz to build a plant there."

https://en.wikipedia.org/wiki/Tuscaloosa_City_Schools

I'm guessing Saturn's concern was specifically with workers with just a high school diploma?

What makes this amusing is Saturn was probably looking to open plants in the South because most of the states are right-to-work states that are more business friendly and have fewer worker protections. The less-government approach that attracted them there also drove them away because less government also meant worse schools and worse workers.

I have a BS in CS. 90% of the concurrency I know is self taught, 98% of the I/O I know is self taught , 85% of the networking I know is self taught.

I went to RPI, which is not an ivy school but it is a good engineering school.

Is some of this a failure to distinguish between software engineering and computer science? Some famous CS person quipped that "CS is no more about computers than astronomy is about telescopes." That's fine for studying applied information theory and whatnot, but it seems like maybe the time has come for a formal bifurcation of the fields.

Aren’t the fields already bifurcated? Software engineering degrees exist. The fact that many software engineers have a more theoretical degree in computer science isn’t that unusual; you see the same pattern with business and economics.

Computer science needs renaming - like you said, computer science is not about computers, and it's obviously not about science either (it's closer to maths).

For me it's that 85% of the time I use networking, you're using something that I learned by myself. But that 85% accounts for only 10-15% of the knowledge. So there are a lot of concepts that I learned in school (like DNS, link/physical layer, microarchitecture, etc) that I rarely use. But just because I don't use a lot of the stuff often doesn't mean I never use it. And even the things that I really never use (like programming an FPGA) aren't things that I regret learning.

Any good concurrency resources you recommend?

Maybe those were older people? Threads came about in 96 and became default in the linux kernel in 2.6 (released in 2003).

Sharing memory between processes was a mess before that. Java green threads (emulating threads in user space, which often blocked badly on IO) were horrifying.

Python threading implementations were even worse of a joke till fairly recently.

Or could it be that the amount of information needed to be competent in our fields is getting too high?

It doesn't help that we are expected to have T-shape skills. We need to be both deep in one area and broad in other areas at the same time.