Most internet startup wannabes don't realize that "consumer internet" is actually a pretty small little niche which is disproportionally well represented in the media and very vocal/annoying with its self-promoting blogs, rock stars, ninjas and all that childish crap. However, it's only a niche.
There are many more industries where you're selling to other engineers, where technical elegance and performance matters more than you'd think: industrial automation/test, real time data acquisition, finance and banking, aerospace, specialized scientific libraries, specialized compilers (for chips you've never heard of), heavily customized Linux kernels for embedded systems, custom file systems, the list goes on forever. Companies like Sun, Apple, Microsoft, IBM, Texas Instruments, Oracle and even Dell - all of them do a lot of this highly technical, lower level work, often in C/C++ and each of these companies has an ecosystem of startups revolving around them.
Your responses are the best illustration of SV echo-chamber effect. Keep on thinking that becoming an AJAXified CRUD htmlizator of MySQL tables is your only way to make money with a CS degree.
I can barely see out of my pigeonhole, but I'll try to reply anyway :)
Neither I nor the article ever suggested that performance is not important, or that low-level work is not important. The parent gave two examples - language development (almost entirely not a commercial activity anymore), and the development of the Linux kernel (GPL). Neither was supported his point, which was that performance for its own sake is the key in those markets, and I was responding to that with generalities about the 'commodity nature' of those examples.
You're outlining markets where performance is an important feature - to products sold to well defined paying (and massive) customers. Thus you are supporting the article, my summary of it, and disagreeing with parent, even as you lament the 'social media rockstar' phenomenon.
