Can Git Solve Blockchain Scalability?

This (maybe) addresses the technical side of the problem while ignoring the economic side of the problem: you can't have a discussion about changing the blockchain without discussing block rewards and transaction fees.

Does each branch of the tree get the entire block reward? If so, then you've going to have an explosion of branches.

If not, and each mined block earns, say 1/Nth of the reward for N branches, then you most likely go back to a single chain of single-block merges, because each miner is going to try to force N back toward 1.

You also need to address the incentives around mining the merge block. If these incentives are not drawn from transaction fees, then you are forced into an inflationary currency.

I wrote the original post.

It's an excellent comment, thank you. I didn't think through the economic consequences. My first reaction is there need to be some betting in place. For example, there could be a fixed cost for a miner to issue a block, and at the same time there could be a variable reward for the same block that is tied to 'good' behavior (not too much branching and not too little) that gets known only when a merge block is issued.

Inflationary vs transaction fees: makes sense, thank you.