thankfully we do have people trying to archive things. sadly not everything will make it into that. just to much new stuff all the time to keep up with. but if we can keep the important and mostly important stuff

It sucks that we already have internet lost media

Other historical artefacts like pottery, vellum writing, or stone tablets

I mean I could just smash or burn those things, and lots of important physical artifacts were smashed and burned over the years. I don't think that easy destructability is unique to data. As far as archaeology is concerned (and I'm no expert on the matter!), the fact that the artefacts are fragile is not an unprecedented challenge. What's scary IMO is the public perception that data, especially data on the cloud, is somehow immune from eventual destruction. This is the impulse that guides people (myself included) to be sloppy with archiving our data, specifically by placing trust in the corporations that administer cloud services to keep our data as if our of the kindness of their hearts.

One of the most interesting aspects of historic preservation of anything is that it's an extremely new concept. The modern view of it is about a single lifetime old, dating back to the early 20th century. Historic structures were nothing but old buildings and would be torn down with the materials repurposed as soon as there was a better use for the land or materials. Most historic buildings that date to the 19th century and earlier are standing not because people invested significant time and money into maintaining a historic structure as it originally was but because people were continuing to live, work, socialize or worship in the structure.

Preservation is entering a very interesting new phase right now particularly in transportation preservation as many of the vehicles in preservation have now spent significantly longer in preservation than they did in active service. There are locomotives that were preserved in the 50s and 60s who's early days of preservation are themselves a matter of their history. There are new-built replicas of locomotives from a hundred years earlier that are now a hundred years old. In railroad preservation there's also now the challenge of steam locomotives being so old and so costly to maintain that some museums are turning to building brand new locomotives based on original blueprints

To be realistic we need to pick and choose what to keep and expend effort/resources on those chosen things.

Without a technological breakthrough in data storage at some point there's got to be some kind of triage done. We all generate more information now than ever before, and this trend just keeps increasing. With things like A.I, XR, the metaverse or other similar concepts it'll also get exponentially more insane how much data we generate. It's not realistic at the moment, technologically or financially, to keep all of it in multiple geographically distributed copies, in a format that will last forever. For a lot of people or organizations it's not even feasible to keep one copy in some cases due to costs.

To do otherwise we would need a breakthrough that enables insanely cheap, infinitely scalable storage, that is immune to corruption (physical or digital) and optionally immutable to prevent modification. It would have to function in such a way that any reasonably advanced civilization can use the basic laws of physics to figure out how it works and consume the contents without any context of what the devices are. It would also have to work regardless of how fragmented it is, to use terms of today's technology if they only find one hard drive out of what used to be a pool of 100, it still needs to work on some level.

It's an interesting thought experiment and hopefully there's some ridiculously smart people working on it.