So really all I know is that they intended to "scale" by increasing the blocksize. ... Has anything changed? Are they proposing realistic alternatives to lightning for scaling? Are there new developments that I should be aware of?
0 confirmation transactions. No, seriously - that's the amazing and innovative feature bcash has implemented to try to compete with lightning.
What, did they get some sort of commitment from a federation of miners/pools that when they receive conflicting transactions in their mempool they will mine the one they encountered first rather than the one with the larger fee? That could be done with like 2 lines of code. If it's more interesting than that than I will go do some research.
You mean disabling replace by fee?
Does that mean all I have to do to double spend is send the funds to myself seconds before I send them to the victim?
If the protocol is first transaction gets blockspace, the victim's transaction won't find its way into a block. Double spend will fail, but legitimate RBF transactions could be impossible.
But if the victim is accepting 0 conf, then the attack will likely succeed unless the victims node is particularly well connected
(EDIT 2 the following for accuracy/factual accurateness.)
With 0 conf, full nodes have to cache transactions since the latest block and stop propagating newer conflicting ones. For the attack to work, the cheater's transaction must already have been submitted before the victim's after all. As long as the network doesn't selectively hide information from certain nodes, cheating this way will fail.
(EDIT 1: as long as 0 conf implies "earlier tx gets blockspace" rule.)
Any fault in my reasoning? Syke? Experts?