Right - the actual answer here is that the people developing Linux weren't bored hobbyists in their basement at all, but employees of big corporations. (Perhaps employees for big corporations working remotely in their basement.) The development structure was different, in that you had employees from one corporation reviewing code from employees of another mailing lists full of people collectively deciding priorities, customers writing patches for their own needs, etc. And that was incomprehensible to people who expected traditional forms of software development; they couldn't conceive of anything other than what they knew or hobbyists in basements.

No one is saying that we don't need talented scientists writing and reviewing papers or organizing journals. The claim is that the structure of doing this through Elsevier doesn't make sense any more.