"A bet on AMZN today is largely a bet on AWS"
Haha you gotta be joking. It's just the opposite. AWS is a business at scale with fat margins. Amazon retail isn't at scale, not even close.
He definitely moved the goal posts, but remember that AWS is most of Amazon Incs _profit_ today.
Also, AWS is a small-ish part of Amazons _revenue_ but given the relative growth rates that is less and less true every year.
The reason why AMZN stock is so high currently has a lot to do with AWS.
2) Which is another key point: the prospects of the company at IPO are very different from the company today. A bet on AMZN today is largely a bet on AWS, which is very different than a bet on AMZN in 1997. If Twilio continues to succeed, it will be because they can use this investment to grow into a stronger long-term position.
I don't know Twilio well enough to take a bet either way. I'm just pointing out that most smart people (you, even, perhaps?) didn't think Amazon would be that gorilla, either. And that was the correct bet, which just happened to turn out wrong.
So if forced to bet, I bet with statistics (against Twilio). But only because building that gorilla is insanely hard (and depends on luck) and not for any of the reasons you've cited.