I guess one thing that could use efficiently all this memory could be serving some RAM-guzzling search engine like meilisearch .

I have some projects for “one day” about such things. Notably #HumanLinks https://humanlinks.org would be about searching all pages with public links on the Fediverse, allowing search only for trusted instances or using your own friends instances for rankings.
HumanLinks.org

After migrating I don’t mind much losing 20k posts of me just babbling endlessly but there are quite a few posts I tagged with projects names and I would like to keep them as references for ideas etc.

Things like #atomique #MarxDawn #IneverDoAnything #HumanLinks posted with old account ( @[email protected] )

Maybe I should just write a script and keep them as text files in a folder ?

Did anyone build a search engine including only websites without any JavaScript ? (Not a proxy that strips JS — rendering the site unuseable — sites that don’t contain any as is).

That would be wonderful for #retrocomputing .

#HumanLinks #smolweb