A Skeptic's View of Crypto (from the Point of View of Monetary Economics)

"What problem does it solve?" That's the heart of Nobel-prize winning economist Paul Krugman's skepticism around cryptocurrencies -- noting that while he may not understand technology, he does understand monetary economics! -- shared as part of a "Let's Settle This" debate (hosted by Versus by KIO Networks in Mexico City) between him and a16z crypto general partner Katie Kaun in September 2018. [You can see her case for crypto here, and their ensuing debate here.]
Money, Krugman argues, is "a lubricant" that facilitates transactions, so the entire history of money has been about making it as invisible and frictionless and moving it into the background as much as possible -- moving from metallic money to paper money to fiat money to forms of payments not requiring physical currency (credit cards, mobile, etc.). So why recreate all the disadvantages of metallic money (e.g., using resources like computing power and energy, much like we once did with gold mining) in digital form with bitcoin? "From the point of view of monetary economists", shares Krugman, "we're setting the clock back 300 years."
He points out the important role of middlemen -- "enduring institutions" like banks and governments -- in providing reputation ("reputation is a technology"!), and the ability to play repeated games, and notes that the "extra roaming charges" they extract are not inherent to those entities: Is this a technological problem, or a social one?? One person's incentives are another person's costs, Krugman further points out. Bottomline, the cost of using cryptocurrencies are quite high; can "clever" technologists really solve this problem... beyond a "self-fulfilling prophecy" arguing their value?