No Data Lasts Forever
lilysthings.org
lilysthings.org
This is the crux of it. You can have your holographic quartz or whatever new technology is supposed to last a hundred million years, but unless someone is keeping it from being thrown away after your family has a yard sale after you die, it's gone. Getting people to care about your data is the hardest thing required to make data last forever.
If fees for adding data were continually adjusted to keep data size to an economically viable size.
I.e. there would be a continual auction of adding data, with the fees spread out among verifiably maintained repositories (say fastest generation of hash across all data from previous hash, every minute.)
And each repository could compete by price to service data requests.
The point being: long term distributed incentive to maintain and serve data.
(Scaling to vast amounts of data vs. sharding - allowing individual repositories to hold partial data - balanced to ensure distributed backups, but not more than necessary to keep each archive size manageable for many providers.)
——
Best done without creating any new coin type. I.e. avoid creating any non-aligned hype incentives.
Maybe this has already been done?
Maybe that will be the ultimate destination of all data: To train AIs.
In the future, data creators should be compensated for their contributions to build AIs.
Even the knowledge of today is not well-preserved. If Stack Overflow were to disappear tomorrow, how many other resources would there be for all the programming knowledge? There aren't that many great programming books. Thank god Github has been archived in the Arctic Code Vault. It could all be lost in a few decades.
Even cultural significant things like how to produce Saturn's F-1 Rocket Engine gets lost. The manufacturing processes, machining techniques, and supplier expertise is gone and wasn't written down. That was just ~60 years ago. The recipe for Fogbank, a material for nuclear warheads, was lost, and it took 5 years to reverse engineer it.