2 ms·
Fundamentally, if humanity is given data set A, it will eventually generate dataset A' of 'higher quality' than A. Colloquially that might be called 'progress'
by greysphere 4y ago
Fundamentally, if humanity is given data set A, it will eventually generate dataset A' of 'higher quality' than A. Colloquially that might be called 'progress' or maybe even 'understanding' or 'learning'. That's different than regurgitation.
I don't have strict definitions for any of the quoted terms because I don't think we really understand them yet.
It seems very possible that human generative content shares much in common with 'stochastic parroting'. But I'd hypothesize there's a filtering process, sometimes at the individual level, sometimes at a group or even civilization level that (so-far-over-long-enough-timescales) tends to cull some information that has low 'value' and propagates other data that has higher 'value'.
When an AI system can do the same, that's a big inflection point as you can give it energy and time and it will self-improve indefinitely.
It's pretty clear humanity has this capability. It's not clear whether AI does (or ever will). Or, perhaps it might already be there, and we just don't understand enough to know for sure.
- swid 4y ago> Fundamentally, if humanity is given data set A, it will eventually generate dataset A' of 'higher quality' than A. This same phenomenon happens with evolution, which is obviously not ‘intelligent’. Each human can basically be randomly exploring a problem space near things they have seen and understand, but they have a way of knowing when they hit on something new and interesting. Even a dumb AI could randomly try a bunch of things… as long as it doesn’t try the same thing twice and had a lot of time, who knows what it might come up with. Validating the solution is usually easier than finding it.
- greysphere 4y agoOne theory I remember reading about some where posits life as a scheme for maximizing energy use. I'm not sure how this relates to information quality. I'd be interested in a reference.