John Carmack's success is due to id Software incubated at Softdisk. The reality of that success, at least in the beginning, was an already established stream of revenue, a ready assembled team, and proper leadership. The brilliance and innovation of Carmack came later. They had day jobs, revenue, and product/market fit already established before spinning out of Softdisk to launch id. That isn't enough to account for the super stellar success they would later enjoy, but it fully explains launching a successful business. There isn't any magic or secret sauce there.

> The problem with this way of thinking is that existing markets are probably efficient enough

Oh, that's certainly not true. The challenge there is the expertise to see the current dysfunction versus the risk of solving for it. Solving for the dysfunction often comes at competition for the limited resources required for current practice of art and rejection from just about everybody. Knowing this there is a choice to be made:

1. Sustain the current practice with current income and no change in risk

2. Proceed with the solution knowing that failure is probable and that a startup begins with a financial valley of death (burns cash without the revenue to make up for it) while knowing the greatness of the idea is financially irrelevant.

I say the greatness of the idea is financially irrelevant, because disruption requires some degree of originality and originality scares the shit out of people. That means the idea will burn money churning in frequent rejection until it achieves some manner of market assistance.