As you can see, ‘decide’ may be a somewhat strong word in this context; I’ve basically worked on what my advisors have suggested, and found it interesting enough not to quit. I suspect I could have worked on practically any problem with much the same results.
Yep, sunk cost is not always a fallacy.
There’s a better way to put that: switching costs are real. Sunk costs, properly identified, are fallacious.