XML. That’s why they cant read and process memes. Hmm.
> XML. That’s why they cant read and process memes. Hmm.
well, XML and derivates are just protocols... and a protocol is just a data stream encoding and a state machine with timers. In particular, a protocol is not AI.
I have never before considered it until now, but to identify memes, one would first need text recognition. Then, one would need to apply it to all photos and videos posted on the net. Then, I imagine that one would need some type of AI to recognize humor by applying face/form recognition to the accompanying photo or video, figuring out the humor, and then identifying whether it is “friend” or “foe” humor.
Then one would perhaps need a large object database in which to store both types of memes, for quick lookaside translation to augment the AI and to store labor intensive manual overrides to compensate for errors generated by the AI algorithms.
Finally one needs the power and ability to “zot” unfriendly memes from selected friendly and neutral sites. (The ability to flag unfriendly sites is already up and running in some form, as “parental controls.” or “deceptive site,” “pornographic content site,” etc. management.)
That is a very heavy amount of processing, even if performed in parallel. It is perhaps not impossible (and one can imagine Red Chinese working on this exact stuff right now) but I wonder how long it will be before we see it in the USA (of course, there are always worst case scenarios, but to be effective those have to posit the taking down of router hubs, volunteer ham radio stations, and... oops... :) )