This means you will never build anything with zero marginal cost. That cuts out a lot of opportunities - and with software eating the world, more and more opportunities over time.

Only few companies can "win the world" by definition -- and besides work, it takes tons of luck, relinquishing control to VCs and stock markets, and boards, and all the right stars aligned. Plus the personal toll.

One is better off building something smaller, where work and talent is enough to maintain it.

There are more opportunities in those companies (e.g. a $1 to $50 million dollar business) of which there are tens of thousands, than to build the 1 in 100 behemoth.

You'll have more chances of success, and you'll live and sleep better. Because money, above a certain amount, also has a marginal utility curve and diminishing returns.

I agree the costs are high, but if you don't pay them, someone else will.

$1..50 million dollar businesses: I agree, great stuff. I'm just not sure that such businesses have realistic long-term moats against the Amazons of the world.

Software will never truly eat the world until there is a way to move matter and collect energy and resources through software alone, without having to create new infrastructure. Until that is solved there is plenty of opportunity everywhere.