RE: Swarm Intelligence in Honey Bees and Attractor Patterns in Humans
Of course they replace people if they are doing something more efficiently than people, but what scares me is the law of unintended consequences (there always are some). For example lets take a bot which picks content which will become popular. The bot gets really good at it, it earns good curation rewards and people don't so much, therefore people stop doing it so much; the normal evolution of collective thought and social taste therefore becomes subject to some serious inertia, because the bot is always picking what we would have liked yesterday rather than what we like to day, and because that then gets pushed to the fore and yesterday's thinking drowns out any new thoughts we may have had today. What's more, the bot misses edge-cases which it doesn't know about, so all exotic, unusual or unique thoughts which would have been surfaced by human curators end up being buried in a sea of innocuous and slightly stale memes.
@dcsignals so you mean exactly like the system we have right now? People post a ton of content but that content is generally only what was hot yesterday. But a bunch of minnows and dolphins swim to it in hopes of a curation reward whether they like it or not.
At least an AI could be given the ability to take actual user preferences into consideration. Right now it's catfish, TitsNTravel (also usually a catfish), and posts about steemit that make the top.
We've only had one or two good days where there was a decent mix of content up on the top of trending.
The inertia in the system is a product of the way the reward system is. Which is kinda the point @dana-edwards is trying to make. A thing gets popular once and there is this echo effect as everyone piles in hoping to whale bait. This makes less attractive and compelling content appear much more important than it actually is.
Also no AI has ever put anyone out of a job. Companies fire people, not AI. It's the company and the people behind it who make these decisions but honestly...
