Zeroconf is not built atop 'trust Wu, Ver and Wright'. It is built atop 'we don't have overwhelming tx backlog that makes doublespends hard to detect'.
without a backlog, it's still batshit crazy to accept 0-conf. people have double spent against both bitpay and coinbase; peter todd did it publicly once. if you close out the invoice before confirmation, there's nothing stopping the consumer from pushing a higher-fee transaction using the same inputs.
In the relevant protocol model, it is not 'batshit crazy'. It is another parameter one uses into the calculation between shrink and volume. in the same way that WalMart makes a calculation between how much merchandise walks out the door unpaid vs hiring a dedicated shoplifting observer for each and every customer in the store.
the "tx backlog" really seems like a red herring. 0-conf has an inherently weaker trust model. satoshi acknowledged this:
As you figured out, the root problem is we shouldn't be counting or spending transactions until they have at least 1 confirmation. 0/unconfirmed transactions are very much second class citizens. At most, they are advice that something has been received, but counting them as balance or spending them is premature.
Yes, it is indeed a weaker trust model. But one completely at the discretion of the affected party.
As opposed to SegWit's inherently weaker trust model that can affect everyone.
accepting 0-conf transactions doesn't mean trusting "Wu, Ver and Wright".....it means trusting the payor not to steal back his money from you.
Agreed.