PGP/GPG provides primitives for managing some sort of PKI but beyond that, I don't follow your argument.
For the sake of concreteness, here's for example how an auction can run on the blockchain:
At time T an asset A owned by an anonymous seller is locked into a smart contract for H hours at minimum price M. Participants are free to enter their bids until time T+H. At the end of the period, the smart contracts examines all bids - if all bids are less than M, no-one gets asset A and ONLY then is asset A released back to original seller. If at least one bid is above M, the highest bid is picked and the asset is automatically transferred to the bidder's address. Everyone else is re-funded automatically. The seller is the only one able to extract the cash from the contract.
Anyone can check the contract code and ascertain:
- the seller can't withdraw their asset and run away with it before auction ends
- the seller cannot reject a valid bid (>M) if the auction is already started
- everyone but the winner will be re-funded their money back
- winner (if any) is the only one who can retrieve the asset after auction is over
... and any other interesting properties. Regardless of whether Alice is from Argentina or from Zimbabwe.
How would you do this on top of PGP/GPG?
Note that source code bugs and hacks happen and are part of the process of working out how to do this stuff safely. There's a recent example of someone messing up a contract and $36M is stuck* forever because the refund functionality had a bug.