How do you convince NFT users to stop using on-chain Bitcoin?
For example by not sharing the data behind historical signatures (which, by the way, leads us to the same model I described above, but just being forced, instead of being voluntary). Bitcoin is a payment system, which means, that pubkey-based functionality should be priority number one. Other use cases are of course possible, and you can create complex scripts, but it should not be the main use case, but just some "fallback" mechanism, intended to resolve conflicts. Which means, that if you have P2TR address, then spending by key is the main functionality, which should cover the majority of all cases, and spending by script is a "rescue mode", which is needed, if something goes wrong. However, if that "emergency path" is deeply confirmed, then it can be safely downgraded into key path again.
Which also means, that if new nodes wouldn't require downloading everything during Initial Blockchain Download, and would downgrade historical data into something like OP_CHECKSIGFROMSTACK, or into other kinds of proofs, then projects like Ordinals could no longer rely on their "huge OP_NOPs" being served by Bitcoin nodes, and would be forced into storing those kinds of data by themselves.
Another thing is transaction joining. If regular payments would be joined, then they could compete with Ordinals and other misuse. Because if you have some "payment-only" transactions, then it is easy to combine signatures, because you just add public keys. However, if you have "data-based" transactions, then compression is harder.
Also, I am not the only one thinking in that way. Another example:
That isn't a use case for *Bitcoin* in that it's something Bitcoin doesn't actually accommodate on a fundamental basis: Bitcoin nodes don't need to store or provide access to historical blocks to operate. They only do today (to the extent they do, many don't) to aid new nodes coming up securely, but in the future that will be accomplished via other means because transferring terabytes of blockchain to process and throw away whenever someone starts a new node won't be sufficiently viable.
Also note that for example chapter seven from the whitepaper, called "Reclaiming Disk Space" is still not applied on the Initial Blockchain Download stage, which means, that Satoshi's conclusions, expressed as "storage should not be a problem even if the block headers must be kept in memory" are not fulfilled, because you still have to download more data, than "just block headers". And also note, that we can get there gradually, for example by applying "just synchronize signatures-only" first, and see how that approach would be received by the community, before going further.
Also, the current model, where you have to download everything, is one of the reasons, why you cannot increase the size of the block by too much. Because then, you can get beyond verification time, which means, that blocks would be produced faster, than you could verify them. However, if anyone thinks that having bigger blocks is a good idea, then improving Initial Blockchain Download is the first step to get it done (and also other simplifications are needed, to not allow people for uploading videos, which could remove all advantages of higher on-chain transactions per second ratio).