This is almost equivalent to saying that the expected value of a startup increases with the scale you aim for -- clearly not true after some point.
It may be true that the number of $1 billion exits per year is only, say, 10 times fewer than the number of $50 million exits. It's an empirical question. The numbers matter.
Let's just raise the $1 billion number until it's a better bet to go for $50 million. At some point the argument that "both are just absurdly big numbers" breaks down.
Perhaps the reason people think that the numbers don't matter is because they believe that the skills required to build a $50 million company is no different than those required to build a $1 billion (or $100 billion) company. That I buy.