The 2018 bug would not have gone unnoticed, in large part because Bitcoin is an open and transparent ledger. The 2018 bug was also never exploited because you’d need to control enough hash power to mine a block to attempt it, and miners understood that it would be detected so quickly that instead of maybe getting +50 BTC from double spending they would’ve actually gotten -12.5 BTC from losing the block reward.

I’d love if someone sent me documents on how to verify the monero supply for myself, but I’ve yet to come across them. While you’re right that we have nodes to validate instead of doing it all by hand, the transparency and simplicity offered by inputs and outputs is desirable. If I really wanted to, I could export the whole UTXO set to an excel sheet and check it separately there. I can also see in my node the aggregate inputs and outputs in each block, and would be able to detect 99% of problems quickly just due to that.

The desire to ossify only applies to proposed protocol changes. A hard fork to fix a critical vulnerability is not a protocol change. And yes, potentially millions of transactions would become unconfirmed depending on how fast this bug was caught. But if you think that’s comparable to allowing inflation, you still don’t understand Bitcoin. Merchants could simply request payment be resent, and many people would do that. Fixing inflation is something which would benefit billions of people perpetually. A hard fork would be a great solution to any inflation bugs, because 21M is not something to compromise on.