Blockchain is also built on the assumption that storage would become cheaper and cheaper (for the ever growing blockchain size), connection would be faster and faster (to handle the bigger TX volume).
And with example of ETH, layer2 to take the load off the baseline and similar with HIVE.
At some point, do we need to make the decision to permanent prune “useless” old data? We may be far away, but if the next bull market or the ones after, if we suddenly gained massive adoption, the current parameters may no longer work.
Very good observations and yes, that's the premises from which blockchains start. That there will be evolutions both at the software and at the hardware level to make it sustainable for a longer period.
Keeping the base layer light and moving more and more on the 2nd layers and sidechains will also help.
There was a discussion on the former chain to prune the blockchain. The decision was to not go ahead with it at that time, but the time may come when such a decision may be on the table again.
Would you be “sad” if your old records that were supposed to live on the blockchain forever got pruned?
Hard to say. The way it was the discussion on the former chain, I wouldn't like it, because some important elements may be lost.
But the first candidates for pruning that I wouldn't mind if they were gone would be content (posts/comments/custom jsons) made by or for applications that are no longer around. Especially if those posts become unusable (for example, dlive, dsound). On the custom jsons side, we may have something like Drug Wars and other games that either failed or ended.
EDIT: On second thought, I don't think real pruning is possible. Because blocks are linked between them by hashes. And you can't remove blocks from mid-chain or you will break it. If you don't remove blocks, the size remains the same.