The financial gains from starting a startup have a wildly varied distribution. The analysis won't be very practical if we assume any meaningful exit at any age. Not sure how a "weak exit" is considered the most common outcome.
A good exit used to seem boring: we already know the family will be financially very well off. That said, it would be interesting to move away from deterministic modeling and build a stochastic version, where we explicitly account for probabilities of success as a function of age, integrate RSU appreciation, and incorporate target net worth.
I suspect there’s a net worth threshold at which, probabilistically, pursuing a startup becomes more likely to get you there than staying in big tech.