Posted: Mar 10, 2012 8:03 pm
by stalidon
Well, this started when we were talking about model-dependent realism in the other thread. We could build a whole structure of meta-models for MD realism with dependency injection, in an abstract library, and feed them to an evolving neural network in digital space. If we then up the level of abstraction we'd probably be talking about what kind of 'concepts' the neural net form about it's reality when we through models at it.

That is, my only reason for wanting it to stay in the Philosophy section is that we probably will be able to talk about some intricate philosophical problems in a language that, despite technical, makes sense to someone with programming knowledge. I suspect professional philosophers in the p of mind department or the formal logic department talk at this level of technicality when dealing with their subjects too.

(My digressions into the state of the 'industry' are slightly off-topic here, my hope being that we abstract from the level of tools to the theoretical-scientific level)