> There are a couple of situations where outages are not normal or acceptable: 1. Dealing with other people's money 2. Monitoring/managing other people's health

Generally true, but there is a couple of exceptions to this rule: if everyone knows that the company is brand new and does not have an established reputation, then using that app requires a general acceptance of risk.

Robinhood was brand new, and outages should have been expected. The problem with Robinhood isn't the outage, it's that it was marketed to college students gambling with their parent's money, who know just enough about the stock market to be dangerous, but not enough to invest properly.

From what I've heard, the "teams" maintaining most of these aren't paid half as much as a mid-level FAANG team.

Luckily for everyone, those industries are so old, they have accidental redundancy built in (paper records for old doctors who can't be arsed to use a computer, etc.).

Ok.

Can you describe to me how you’d design RH to ensure 100% uptime?

“Failure is unacceptable” and “failure implies total incompetence” are different statements.