It started as a "read-it-later" service that would extract the content and bookmark any page you wanted but afterwards the idea that it could be used as a distributed, curated, web of "clean" and tracker-free html documents. Basically, every web page you saved on your instance would be saved on your IPFS server as well and it wouldn't be hard to write an extension to check if any url you want to open has already a cleaned version on IPFS.
So, with more people installing/using this system, the more the different instances would collect pages and more people with the extension could go on without needing to visit any site that could actually track you. What pulled me off this (besides "regular" work and family) was the fact that IPFS does not have yet any sort of ACL for your pinned content. As it is now, your node will serve anyone that asks for content that you have. If you are pinning content from different websites it would be quick to make you a target for copyright lawsuits.
I still use my own instance, but at the moment it is just something that does more or less the same as Wallabag. I do wish I get some time to make it more useful for more people though.