Would you say that a lot of smaller coins tout their quickness and cheapness as a benefit but it is simply because of their low usage? I have heard similar about bitcoin cash.

Its difficult to say. Having low usage would make most reasonable coins cheap and quick; but does not give us any information of how they would behave at scale.

In the case of bitcoin cash in particular; it would probably still be fine at bitcoin scale. This is because its main difference is removing an artificial limit within Bitcoin. As this limit was set without an empirical basis (and bitcoin showed no signs of degradation as it approached the limit), you would probably be able to scale the block size up some without causing problems.

How much you can scale is still at question. Here [0] is a talk about this very question. Using a small testnet (~6 miner nodes and 12 clients) they were able to achieve 500tx/s with a 1GB block with relativly minor optimizations of the standard Bitcoin implementation (the final bottle neck here is propagation delay reaching 10 minutes). This is, in my opinion, an upper bound on what the Bitcoin protocol can handle.

[0] https://www.youtube.com/watch?v=LDF8bOEqXt4&t=4079

Not the parent but I would say yes, definitely. It’s easy to make a cryptocurrency where there is less demand for transactions than can fit in in a small block. The real challenge is when the demand seriously ramps up, which only bitcoin (core) and ethereum have seen.

I believe a drop-in replacement called Bulletproofs is being developed to reduce the transaction size. It's been a while since I looked into what else is being worked on.

Smaller transaction sizes with adaptive block sizes seems like a good scaling solution. Is there something I'm missing?

The fee in terms of XMR actually reduces when there's more transactions. However if its value in terms of USD rises, then the fee in terms of USD still might go up with more adoption. Perhaps that's something the developers can adjust in their regularly-scheduled hardforks.

I could totally be wrong, but I'm pretty sure I read that Bulletproofs only reduce transaction size by about 40%. While that's still a nice decrease, it won't get the project to the scalability that it needs. It may buy them a years worth of time at the pace the transactions are growing before they'll need a better solution.

I just skimmed through their subreddit, and it appears to be between an 80% to 90% reduction.

That still makes it a little larger than the average Bitcoin transaction, but the mentality there seems to be that privacy has its price. While next-to-free is preferable, I don't disagree.

Thanks for the clarification! And yes, I agree as well that privacy does have a price. However, with transactions growing tremendously with the increased popularity and speculation of crypto-assets, there will be a point where enough is enough. I hope that a further iteration for greater efficiency will be found in the future.

Yes so we could imagine a new upper bound of 100 tx/s

This isnt future proof but should alleviate some pressures while internet and widespread computational infrastructure improves

Sidechains are on the research roadmap you can see it here: https://getmonero.org/design-goals/