For researchers who've founded a company before -

How would you approach your PhD if you want to start a company during/after your PhD? How would this influence the choice of advisor you choose, schools you pick, etc?

For background: I'm applying for a PhD in ML (NLP to be specific) and am actively thinking about this question