I mean, I didn't go to MIT, but I just finished a PhD in CS from UPenn, and I'll attest that he's not wrong. Perhaps excessively pessimistic on it, but like, what he's expressed is not an uncommon sentiment. Like you'd hear the same thing from like a third of the grad students I personally know (across the entire spectrum of top-20 CS departments). I myself was fairly ambivalent on my own PhD, but ended up finishing it because

1) I had some strong publications early on in my grad career (first 2 years) so it made it much easier to get to the eventual finish line without much hassle.

2) I'm not a US citizen, and graduating means that the immigration pathway gets a lot easier (EB-1 green card application instead of EB-2, for instance, and since I'm from India that's a huge difference).

3) My advisor let me study/work on what the hell I wanted in my final 2 years (essentially some macroecon + finance research at the Wharton school, and spending a bunch of time reading up and working on large scale graph computation) so the opportunity cost was low to "finish up" the main line of thesis work, with the understanding that any future academic career was out of the question.