The abstractions and tech which cryptocurrency works on is fragile in most of the same ways as the layers that came before it (credit cards/banks/etc). Try to look at all this holistically. How many people and resources do we dedicate on this planet to keep track of money, economy, mine the coins, cash the checks, swipe the cards, keep the lines working? All of these abstractions make trade faster and more liquid but…
So we need to understand why part of it sucks, why part of it is great, and develop better alternatives that are also great but suck less. For example, it is important to keep track of who owes what to whom, otherwise people can't trust each other to make promises on a large scale with strangers. It's less good that people can gain a lot of profit simply by manipulating numbers and moving financial instruments around. Part of this stems from all value judgements being collapsed down into a single number which of course can never be self-consistent [1], part of it stems from informational differences between actors and other real-world deviations away from what a "perfectly-liquid" market should look like. If financial activity was truly a "market optimisation" mechanism, we ought to be able to replace the entire thing with a neutral algorithm and get rid of bankers and personal profits here completely.
[1] you cannot totally-order a 2-vector (and in general a n-vector) so that the ordering obeys the neighbourhood principle