Wells Fargo Is in Trouble Again

I think often about the time I wrote:

And lots of computer engineers tweeted and emailed to be like “no, actually, it is a hard problem of computer science to have a big database of who has what, and to update it instantly and reliably to reflect transactions from many different sources.” And I was like, sure, fine, I guess. I still feel like I was entitled to be right: A bank is, at its heart, a computer for keeping track of who has money, and for updating its ledger as people send and receive money. And at a high level you and I could describe how we’d expect that computer to work — “if I deposit $100 in an ATM, the bank will increase the number in my account by $100,” that sort of thing — and we will be disappointed if it doesn’t work that way, if the bank loses track of who has the money or how much they have, or if it doesn’t update its ledger promptly or process transactions in the right order. If the bank messes up and says “look I am sorry but keeping track of money is a hard job and you can’t expect us to do it with 100% accuracy,” we will say things like “yes we can” and “that is literally exactly what we expect of you” and “if keeping track of the money is too hard for you then maybe you should not be a bank” and “now you have to pay an enormous fine.” And yet, sure, empirically, banks do sometimes mess it up. It’s not as easy as it sounds.