1/3 Every founder faces a tradeoff between prioritizing exploration (build new stuff) and exploitation (build more of what's working).

This is the multi-arm bandit problem, where the answer is normally some combination of the two, but this misses something fundamental.

2/3 In most tech markets growth has a strong positive feedback loop, which means the more you focus the bigger the gains will be (to a point).

The conventional wisdom of balancing doesn't consider increasing returns from focused efforts.

3/ Thus the best approach once you've found PM-fit is to prioritize *taking the entire market* - focus on what's working (80%) and leave a smaller portion of your time for new stuff (20%).

Many caveats, obv, but a guiding principle for me.