tldr; some sub-sites are justified, others are just a pain.
I believe so, especially when you consider that revenue is generated via ads, and the more visits to a single SE site the better.
I tried to get the good folks over at RPi.SE to see that their questions fit happily inside of a linux SE and a Embedded electronics site such as that proposal I just linked to, but as you can see it wasn't happily received. From what I gather they can't see past the fact that their site is in beta and the other one isn't (although that will change soon).
Also, The top answer here indicates that those proposing sites aren't familiar with the intended purpose of the sites, to attract experts and then n00bs, not keep the site for n00bs. Even for a site for rookie programmers such as Arduino and/or RPi, it would hugely benefit them to have people with higher level knowledge in those fields visiting that site, and that probably won't happen.
A snippet of my answer from the above question:
In general though, it seems a lot of cliques just want their own 'board', but don't really understand what goes into making a site, and also don't understand why common interests should be merged into a single site. SO is really a shining example: There is pretty much every programming language under the sun in scope on SO, both compiled and interpreted languages. There isn't a separate site for python programmers, or windows programmers vs linux programmers, scripters, or arduino programmers, etc. If it's a programming question it goes on SO. C or C++ has a large enough following to warrant its own site, but that doesn't mean it should, for way too many reasons to list. Suffice it to say that
- Common interests should stick together - Programming concepts are fundamentally the same across all languages, even if syntax is different. Thus, someone with expertise in a different language could answer your question, but may never even see your question if it's on some niche site.
- A broader question scope draws more traffic - more traffic equals more Q&A and more search engine results, which equals better ad $, which equals the possibility of a sustained site and not a killed beta.