thanks, even as time goes by some topics remain actual. As to the economic ( and any other ) models. Models used to be simple enough to do the job . It worked well in physics, but not so well in social science ( economics included). Now we are moving to the next level - AI and neuronal networks are best in working with patterns and extra-large dynamic data sets - on the contrary humans are not good with data ( comparatively ) and therefore try to generalizations, experiment, applying Aristotle style determinism and reductionism. Then we apply more methods at discussing which model is more appropriate. But there is no need to make sets of such specialized and simplified models any longer. Each individual social (economic as a subset) agent can now be represented in the virtual model environment as a digital tween ( done with epidemiological studies and not only ) - model inside of larger models may be few time over . Adding RT behavioral and GIS data into each of such gidi tween submodels, and then run the whole thing fitting more real-time data as it goes - crazy fantasy? not any longer. AI worked like that with no need for text and word-based time-consuming arguments, just provide sufficient data and computational capacity models will selfadjust. Ashby Law of requisite variety made this case some time ago but now technical means ( computational power + AI + close to real-time data gathering ) are getting there too. In order to deal properly with the diversity of problems the world pose, one needs to have a model or\and repertoire of responses which is (at least) as nuanced as the problems one is facing, or try to model - only greater variety can predict given variety - so to say - ))) who on earth may\will read this crap of mine ))))
