the only 'thing' that segwit is useful for is being a gateway format for alternative networks like LN and sidechains, but as we all know LN and sidechains are not bitcoin
Segwit's script versioning also allows the introduction of new signature schemes relatively easily via soft fork. It makes implementing Schnorr signatures much easier. Schnorr signature aggregation can provide significant space and fee savings.
Which is brilliant. But, let's be honest, based on his posting history, it's not like franky1 is going to be thrilled about the prospect of making implementing features via soft forks easier, now, is it?
In b4 he posts "another Core plot to bypass consensus blah, blah, blah".
which is a trojan horse. imagine the blockstream/barrysilbert team (many exchanges, merchants and devs that made up the majority of the NYA agreement) decided to add a new script that allowed any input to be added even if the signature didnt link to the input (eg i put doomads utxo's into a tx to me, without the signature needing to prove im doomad's utxo owner).
there are many many other dangerous implementations that can be added and by having its as 'soft fork' (no consensus required) means that there is no way to prevent it.