The hash power to do this would cost more than the payouts would be worth. This means that the only potential motivation to do this is to crash the network. A state actor with quantum computing capability might attempt this in the future if possible, but keep in mind that quantum resistant algorithms can be put in place very quickly by changing the number of zeros in front of the nonce.
In order for a fork to succeed and double spend, it has to rehash all the way back to the point it wants to overwrite and maintain that fork against all the other nodes in the system. If one simply submits invalid transactions to the network and does it consistently, the other nodes will censor that cheating node and be blacklisted.
There's a series of rules for submitting transactions to the mempool, any one of which if returning false indicates an invalid transaction. Long histories help prevent this because if a hash changes in any of the Merkle trees whatsoever, we know tampering has happened and it will invalidate everything after that point in the chain. Other nodes detecting such tampering would simply upload their own copies of the chain to the network to correct it.
I have to learn to get to your level of comprehension of POW, but you also assume that POW is how currencies should be managed, and as you know there are rivaling strategies.
I consider Bitcoin inferior and obsolete.
I do not deny its evident value, but I hope it will crash, and not in order for me to buy it at a lower price, but because of its waste of energy and initial distribution.
My bigger problem is why even Byteball, a far more advanced currency, repeat this same waste.
That said, Satoshi was blockchain 1.0. There's ways to improve privacy.
My main concern if efficiency.
Privacy contradicts with transparency, and I try to avoid thinking about which is more important and if and how they can be combined.
RE: What You Should Know About the New Blockchain Phone