I've had some thoughts on, essentially, doing more of what historically worked; a mix of "archival quality materials" and "incentives for enthusiasts". If we only focus on accumulating data like IA does, it is valuable, but we soak up a lot of spam in the process, and that creates some overwhelming costs.

The materials aspect generally means pushing for lower fidelity, uncomplicated formats, but this runs up against what I call the "terrarium problem": to preserve a precious rare flower exactly as is, you can't just take a picture, you have to package up the entire jungle. Like, we have emulators for old computing platforms, and they work, but someone has to maintain them, and if you wanted to write something new for those platforms, you are most likely dealing with a "rest of the software ecosystem" that is decades out of date. So I believe there's an element to that of encoding valuable information in such a way that it can be meaningful without requiring the jungle - e.g. viewing text outside of its original presentation. That tracks with humanity's oldest stories and how they contain some facts that survived generations of retellings.

The incentives part is tricky. I am crypto and NFT adjacent, and use this identity to participate in that unabashedly. But my view on what it's good for has shifted from the market framing towards examination of historical art markets, curation and communal memory. Having a story be retold is our primary way of preserving it - and putting information on-chain(like, actually on-chain. The state of the art in this can secure a few megabytes) creates a long-term incentive for the chain to "retell its stories" as a way of justifying its valuation. It's the same reason as why museums are more than "boring old stuff".

When you go to a museum you're experiencing a combination of incentives: the circumstances that built the collection, the business behind exhibiting it to the public, and the careers of the staff and curators. A blockchain's data is a huge collection - essentially a museum in the making, with the market element as a social construct that incentivizes preservation. So I believe archival is a thing blockchains could be very good at, given the right framing. If you like something and want it to stay around, that's a medium that will be happy to take payment to do so.