Excellent question! My obviously non-compelling contributions to the recent AI-polling thread <http://friam.471366.n2.nabble.com/How-soon-until-AI-takes-over-polling-tp7599471p7599481.html> were intended to evoke ideas like those expressed here:
A Question of Responsibility by John Collins https://www.academia.edu/177687/A_Question_of_Responsibility E.g. "Chomsky’s general point in this passage, I take it, is that the empirical coverage of any theoretical discourse can be rendered as a commitment to a set of the relevant entities ([...]). Such ontological commitment to the sets of the relevant entities, however, is not required for the explanatory goals of the given sciences, unless, of course, the science is a branch of mathematics that is concerned with large sets and their properties, and there the identity of the entities is irrelevant." If we imagine the output of an ML inducer as a just-so-story (similar to Kepler's laws) and an identified mechanism (similar to Newton's laws), it argues for something like inference to the best explanation. Sloppy IttBE can easily lead to "conspiracy theory". But well-done IttBE is simply good science. The *difference* lies in the well-done. Enter orgs like Pro Publica, contrasted with your crazy Aunt poking around Facebook. I *think* EricC was trying to make a point like this in his last response in the deductive fidelity thread. I still owe a response to that. But the idea that believable rhetoric needs something like *coherence* ... not as formal or strong as consistency, but something like it. And the point I made in my 2nd AI-polling post is that it not only matters that your argument hang together. The mechanics of the logic matter. It's the *method* that makes the difference. On 11/12/20 8:44 AM, Steve Smith wrote: > Your reference to the little evidence, etc. leads me to another new word > I encountered: "Aeteology" in the tongue-in-cheek usage of your own > oft-lobbed "just so stories". Is there any difference between a "just > so story" and "a conspiracy?". Intent? Consequence? > Convergence/Divergence? -- ↙↙↙ uǝlƃ - .... . -..-. . -. -.. -..-. .. ... -..-. .... . .-. . FRIAM Applied Complexity Group listserv Zoom Fridays 9:30a-12p Mtn GMT-6 bit.ly/virtualfriam un/subscribe http://redfish.com/mailman/listinfo/friam_redfish.com archives: http://friam.471366.n2.nabble.com/ FRIAM-COMIC http://friam-comic.blogspot.com/
