I have a hobby as a quake archivist.
It’s fun for two reasons: the nostalgia and the hunt (e.g. variable rewards).
Many files I seek to archive are lost. They maybe in the deep web somewhere, or they may be in peoples personal archives, but unless I can surface them, they are essentially “lost”. Lost to the archives at least.
I was thinking, soon the AI models will be good enough to generate list artefacts on demand.
For example, take a missing version of a given quake1 modification.
The context might be:
- all preserved versions around the missing version
- all derivations of the mod (e.g. extensions, diffs between versions, base code, etc.)
- all documentation about the mod
- all webpages and related content for the mod
- all webpages that mention the mod (e.g. reviews, etc.)
- all usenet discussions for the mod
- and on.
Then request a plausible instantiation of the missing artefact/s.
I could do this for the simplest mods now, with careful scaffolding for the LLM.
Evil peoples could poison my archives with such “finds” now.
I’m not sure what to think about all of this.
Is archiving useless then? Perhaps.