To expand on that a little, the EA summit was expensive enough that I expect the price to turn away most people not already fairly involved in the movement, and that seems like a lost opportunity.
On the other hand, you can start to do this with an order of magnitude less money.
I think there’s an inherent problem that if you want to do the best possible thing, you’ll never know that you’ve succeeded (you’ll often no that you didn’t), and therefore never get that satisfaction.
People earning to give seem to deal with this by setting goal amounts or income percentages. If they reach or exceed their goal, they’ve succeeded, even if they had some dollars that they could have donated but didn’t.
This kind of trick is harder to use for something like choosing a thesis topic, but you could try something like it. You could have a goal that your thesis should “substantially further the goals of effective altruism” or maybe more usefully that it should relate to some aspect of EA like “international development” or “human truthseeking” that seems useful to investigate further. You can consider candidate topics that fit your goal and choose freely using any criteria you want, whether they include “looks higher value” or “looks like more fun,” and as long as your topic meets your original goal you can count yourself as having succeeded.
But yeah, a supportive community also seems really important (-: