This is also very true for startups! 2-people teams can iterate and learn so quickly that at some point they are able to outcompete (or come very close to) exis

This is also very true for startups! 2-people teams can iterate and learn so quickly that at some point they are able to outcompete (or come very close to) existing market leaders (e.g., Figma/Adobe, Linear/Asana, Pulley/Carta). So the y-intercept (or the starting point) turns out to not matter much if the slope (growth and learning) is high!