Fixing the bloat "problem" isn't on our agenda right now. We can achieve linear reductions in blockchain size through various schemes, but none of them make a particularly large impact. If anything we will implement a lightweight wallet that uses a bloom-filter-like approach to tell full nodes which set of viewkeys its interested in, and then left those nodes do the heavy lifting. In fact, we can offer a lightweight pruned blockchain right now by only maintaining a limited set of blocks as well as the txoset and the key image set, which will give you a fully operational Monero node that needs less than 1/10th of the current blockchain space. That tooling is premature, though, as there are other priorities (eg. headers-first sync up) before we can focus on that.

It boggles my mind that Evan didn't design the protocol around micropayment channels, where every full node could be incentivised to mix transactions or vote on InstantX transactions or whatever. It would make the anonymity set so much larger.