Picking one now, when the threat from quantum computers is very likely still decades away, seems very premature though. There is a good chance that whatever we picked today would be at best outdated and at worst insecure by the time it actually mattered.
for Bitcoin, yes. For protocols that involve encrypting network packets using public keys, it might make sense to pick now (and I believe this is the reason OpenSSH did so). I do not know why it doesn't make more sense to extend/redesign the protocol to encrypt using negotiated ephemeral encryption keys, but maybe that's something specific the protocol.
I still can't believe why the bitcoin protocol isn't using end-to-end encryption between nodes using self-signed certificates. That would prevent information leak that someone would harvest and attempt to break a specific ECDSA key.