Yeah I know. So I've been following MW development since before Grin launched, and there's some interesting things about that. The original MW paper specified an interactive protocol with no programmability. It also required absolutely no historical data preservation whatsoever. Andrew Poelstra massaged the cryptography a bit and brought us the MW we know today, which does have programmability, at the cost of needing to preserve a range proof for every transaction forever, which in MW is called a transaction kernel. So you get multisig and all that good stuff (but no scripting, but there are ways to script with it nonetheless) but you lose that very powerful feature of needing only the UTXO set for full security.

I think that feature is the killer feature of MW, it is the trait that allows it to scale to the limit of the trilemma, you don't even need a block size if the chain doesn't grow forever, the block size is limited by a function of block time and network latency and not by transaction size. But we have to have multisig and that kind of stuff to do anything more interesting than plain cash, so here we are. I hope we can figure out how to build something that has both and I can see no fundamental reason why it is not possible.