On 07/06/2015 11:10 AM, Marcus Daniels wrote:
At some point won’t these behaviors too be mastered by machine learning? Obviously, I’m not just taking on gaming here, I’m taking on the idea that people ought to master narrow “skill sets” at all. Ok, so a gamer can track 7 objects instead of 3. Machines could track hundreds or thousands. Better to design the machine, no?
Arbitrary google response: Age-related differences in multiple-object tracking. http://www.ncbi.nlm.nih.gov/pubmed/15746018 Let's say you wanted to, I don't know, _drive_ or maybe juggle ... or simple play flag football with your grandchildren. It seems like multiple object tracking exercise might help. It's a bit silly to suggest such skills are always "narrow". I had an interesting discussion the other day. A friend suggested she _needed_ a personal trainer in order to exercise, that without the trainer, she would neither be motivated nor know what/how to do various exercises. She used this "disability" of hers to argue that she doesn't get much out of yoga (the 1 or 2 times she tried it, heh). I can't really sympathize much with her position. The point of exercising is to consistently _try_ things ... to poke around and see how/if you could do it slightly differently. Having another person tell you what/how to do something is way less rewarding than learning how to do it yourself ... even if all we're talking about is twirling a coin between your fingers. (Sure, if you're really really good at something and you want to be much better, then you need a trainer to sqeeze out that hidden performance, but not at the amateur level.) For the exact same reason, running on forest trails (as opposed to treadmills or in circles on a rubber trac k) is actually a very "broad" skill. And it's a very handy one. Is it better to build a robot that can run on forest trails? No. That would be very cool. But having your robot run around the mountain isn't near as rewarding as doing it yourself. Is it better to build a robot to run in circles on a rubber track? Yes, absolutely. I see zero benefit from having humans do that, much less rewarding the fastest ones with medals. 8^) -- ⇔ glen ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com