Unfortunately I don't see that ever happening. The only way perlmonks is coming back is through optimization or more resources. Or, if AI training sets eventually become cheaper to download from aggregated collections than re-scraping them.

I've been meaning to offer to help with that, but haven't found time yet. But I think a good option would be if one of the host names (e.g. "perlmonks.org") was designated as the one to use when logged in, and the other aliases were directed to a second server which would serve cached anonymous-view content from a database which was a replication slave of the real one. If I could run that on a Digital Ocean droplet or Linode, I'd be happy to pay a few bucks a month to offload traffic from the real perlmonks.