Timeline for How to keep humans pilots instead of AI in sci-fi future?
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 12, 2015 at 0:43 | comment | added | smithkm | To expand on @dsollen's comment Humans are very good at confusing randomness and patterns. We see non-existent patterns in randomness and when we try to be 'random' ourselves we tend to produce predictable patterns, and can't see that we are doing so. Generating secure randomness is something we've put a lot of effort into, and we're able to make computers really good at it, as long as we don't allow ourselves to intrude and screw it up. | |
| May 11, 2015 at 17:45 | comment | added | dsollen | I can see your first point, but I think AI will be far less predictable. It's trivially easy to add a random number generator to an AI. Given certain options it randomly chooses possible options. It could even randomly choose suboptimal choices, as game theory shows that this is an important part of making a Strategy that can't be counter-played. Randomness and avoiding others predicting your strategy are well known concepts in both game theory and AI; it's mostly a solved problem. A remotely well implemented AI is less predictable then humans. | |
| May 11, 2015 at 17:02 | comment | added | Jim Green | Professional poker players hate to play against amateurs. A pro can always guess what another pro is thinking, and knows that most pros know the probabilities. An amateur, on the other hand, can bluff and go all in with nothing... but can the pro take that risk? | |
| May 11, 2015 at 16:10 | comment | added | DaaaahWhoosh | One of the reasons I don't want a self-driving car is because it'll probably go the speed limit. I'd add to your answer the fact that when it comes to disobeying orders and getting the job done, humans are way better than AI. | |
| May 11, 2015 at 15:44 | history | answered | Mike.C.Ford | CC BY-SA 3.0 |