It’s not obvious and depends on your personal fit for each and especially the quality of your startup ideas. I suggest making a spreadsheet with careful expected-value estimates, and picking the option with the higher expected value.
To help estimate the value of a very talented researcher, you could draft descriptions of your median and 90th percentile expectations for skills you would have in the future, and asking someone at a top organization how they would value someone with those skills.
To help you estimate the probability of success for trying to become a billionaire,
soft upper bounds come from the best startup opportunities observed so far; Sam Bankman-Fried thinks he had a <30% chance to succeed at building FTX, and most ideas are >3x lower probability than this
It’s not obvious and depends on your personal fit for each and especially the quality of your startup ideas. I suggest making a spreadsheet with careful expected-value estimates, and picking the option with the higher expected value.
To help estimate the value of a very talented researcher, you could draft descriptions of your median and 90th percentile expectations for skills you would have in the future, and asking someone at a top organization how they would value someone with those skills.
To help you estimate the probability of success for trying to become a billionaire,
soft lower bounds come from reference classes; you can beat these significantly because you’re actively trying to become a billionaire: e.g. see https://applieddivinitystudies.com/billionaire/
soft upper bounds come from the best startup opportunities observed so far; Sam Bankman-Fried thinks he had a <30% chance to succeed at building FTX, and most ideas are >3x lower probability than this