In my opinion, we should increase the difficulty of some hack like that and not facilitate it.
You're of course partly right, and I've already written that I would support a reduction of the byte limit for data in Taproot. Although I'm still hoping for this fad to wane "on its own", because I'm not sure what functionality you would cripple if you lower that limit.
But regarding illegal data, I'd be much more afraid of a "destructive" attack of a group which uses standard transactions (not technically "standard", as Taproot tx are also standard, but those who look like financial ones) with illegal data, building a protocol around that, just to drive the price of Bitcoin down. A short seller attack, or a government attack. Thus no, we can't rely on nobody doing that ever, with or without ordinals. What we must fight for is for legislation equiparing Bitcoin explicitly with ISPs in this aspect (everybody in his jurisdiction).
Do you have some proof to back [the possibility of storing data on Monero] up if you are 'sure' about the possibility?
There is at least
this Reddit thread and
this one (see answers from manicminer5 and from BrugelNauszmazcer, which would work even if tx_extra is eliminated, like some developers seem to plan, although it needs a "cooperative" uploader sharing a view key). You're partially right that data on Monero
can be stored privately so it's difficult to retrieve for an observer without additional information, but you can broadcast the needed keys in another P2P network (as for this you don't need double spend protection). The "destructive approach" would thus also work there.
I'm afraid that no modern cryptography can solve that problem - if somebody is able to read the data, even the uploader himself, then he can make them accessible for others, too, if this is his intent.