Almost every time I’ve made a major decision about how to spend my time, I’ve later thought to myself, “Man, if I had done my homework better on this decision, I would probably have made a better call.” Not that I think most of my decisions were bad—I think many of them were probably a good call anyway—just that they would have been improved (in expectation) by more thoroughly trying to learn things like:
- what sorts of options I had
- what were my main uncertainties about their goodness
- information that would resolve those uncertainties
This phenomenon has persisted despite me being aware of, and trying to correct, this tendency not to do enough homework. For instance:
When deciding which colleges I wanted to go to, I ruled some out (notably tech schools, e.g. MIT and Caltech) based on the general principle that I wanted to have classmates with more diverse interests. I think I probably ruled them out based on insufficient evidence.
When I decided to go to Harvard, it was mainly on the basis of getting much more financial aid there than I did elsewhere. I didn’t go to the revisit weekends for any of the other places I got in. I think Harvard worked out extremely well for me, but that again I could have done a lot more evidence-gathering. I also didn’t sufficiently consider the option of taking a gap year.
When picking classes at Harvard, I browsed the course catalog a lot, but I didn’t make much of an effort to seek out older students with similar interests/background to me and get their opinions. I think this led to me missing out on a bunch of interesting classes (mostly humanities) and taking a few classes that were too easy or superficial (freshman math and computer science, lots of “general education” courses).
When I was applying for summer internships during my freshman year, I only applied to two (Fog Creek and Khan Academy) based on things randomly being salient to me. I should have been way more systematic about this.
I got an offer from Fog Creek midway through interviewing with Khan Academy and accepted largely on the basis that Khan was taking a long time to move me through the interview process and wasn’t asking me very many technical questions, so I inferred that Fog Creek would teach me more technical stuff. This inference was pretty clearly unsupported; Khan seems like a great place to go to learn technical stuff.
In college I tried to become a research assistant for a couple professors. Both times I got offered relatively boring projects and ended up either flatly turning them down or not doing very much on them. Again, I picked these professors based on not-very-strong reasoning like who I was aware of at the time; I should have made much more of an effort to find people doing interesting work.
During my sophomore year I was part of a team that won some money in an entrepreneurship competition at Harvard. I decided to spend my summer working on our project. This didn’t go very well because (a) nobody else from our team was working on it full-time; (b) we didn’t have a very good idea of what we were building; (c) two of the team members didn’t see themselves working on it full-time essentially ever; (d) we didn’t have any users to give us feedback. All of these things were trivially foreseeable if I had tried to spend some time thinking about why startups tend to fail.
Many of my decisions at Harvard were guided by my “academic taste:” what kinds of skills I thought would end up being important, what kinds of research projects I thought would be interesting, etc. Unfortunately my taste changed a lot, and I wasn’t usually very sure about it. I think if I had tried harder to develop strong views about what kind of things were important to work on, it would have improved a number of my academic decisions.
Again, most of these things worked out pretty well in the end. But in each case I think there would have been pretty high returns to getting more information about my options and doing more to resolve my uncertainty about them.