It's clear we're looking at this from two different viewpoints, and that is what makes forums so awesome.
I agree if he has some C# experience, continuing that may be the best option from a learning perspective. By all accounts Visual Studio is a solid IDE, and if his development machine is Windows to begin with, then it's going to all integrate reasonable well.
However, I was also pointing out the cost of doing business as a Windows shop. This may not apply to him, because I don't know what his career goals are. While Windows hosting is fairly inexpensive, scaling it typically is not, and hosting it internally is not (relative to income and use case, this last bit may not be true).
If the desire to freelance, or to build your own company, the cost of establishing yourself as a Windows shop is going to be higher out of the gate. Where when you first start a company, money tends to be tight, this could be of concern if it applies.
If the desire is to finish training and join a start up, chances are other tech stacks would likely serve him better. If the desire is to join a established company, .NET is more likely to be in place.
I agree that intranets are not accounted for in my link. That is a valid option, if such a thing would be of interest. I personally think the more interesting problem spaces exist on the larger web, and not on a intranet. But that's a matter of opinion, and it could be argued either way in the right conditions.
To expand a bit on my legacy point (and I agree this is not limited to Windows specifically). Anything that has major backwards compatibility breaks should be discussed. Angular Pre 1.3 / Post 1.3, Python 2.3 / Python 3, IIS 5 to 6 (I believe?), even jQuery 1.9 to jQuery 2.0 could qualify. The work is totally doable, but having to deal with rewriting old code to keep it working because the design, server, stars shifted, etc kinda sucks. It's absolutely worth pointing this out whenever the topic of legacy code comes up, that it may affect the chosen language / platform being discussed. In this particular case, it's of value because they would be apt to learn on the current platform, which may not apply when doing legacy work, knowing that is even a thing to be aware of, I feel is helpful.
I am seeing this as akin to someone who goes to college to study say, Latin. If that's your dream, more power to you, but knowing the effects of said choice should come into play before making it. While nothing prevents you from learning other tech stacks at a later point, what you focus on to learn will likely heavily impact what jobs you can take, and what the actual demand for your skill set is at least in the short term, which makes it worth seeing both sides, to which I thank you @ns81 for helping to illustrate.