There was a great point raised by Guy on Bitcoin Audible (Guy's Take #54). He makes an argument not on the code having any bugs, but a possible game theory situation where commitments that could really disrupt block space economics.
Here's my best summery of the idea, please listen to the episode, he does a much better job:
It involved the possibility of exchanges wanting to do withdrawls to one address with commitments to each customer. Of course people will eventually have to unroll (and pay fees) (burden placed on customer, not exchange) to truly get their bitcoin. People said this could get used to get through periods of high congestion, however it will always save exchanges money, so why wouldn't they, even in periods of low congestion. If all things remaining equal, fee market-wise, early "unrollers" have to pay more fees, pushing off unrolling but, as blockspace becomes more scarce, people will start to panic by unrolling all their utxos. Soon all the accumulated "blockspace debt" will be attempted to be paid off, spiking fees crazy high. John Carvalo called this a "tx grenade." In addition, it also is a great way to spam network without paying much fees.
Once again, he explains it alot better. If 119 goes through, awareness needs to be raised that although its good to get your coins off exchanges, commited utxos are only truly yours once you pay your fee. Which can be really low like now, or really high in the future, exacerbated by people FOMOing into unrolling.
This is surprising. Certainly, exchanges wouldn't have the incentive to pay for fees even in low congestion moments. This could be done with a simple modification on their terms and services and leave this responsability to their clients. Definitely, in the long term, Bitcoin is destined to base its security on user transaction fees.
Thank you for your comment!