ThoughtStorms Wiki

Context: TheLogicOfTheFuture, CollapseThinking, Pessimism, ArtificialIntelligence

(ReadWith) SyntheticData

I tweeted :

If anyone wants a quick map of the future, it's this.

1) Today's AIs are as smart as they'll ever be, because they are trained on a large store of human expertise which is about to be destroyed.

2) AIs are going to destroy the business model for experts because it's obviously way more expensive to train and support human expertise than it is to hire the AIs.

3) But that means that dumb humans are going to use AIs to achieve what previously would have required expertise.

Instead rewards will go to those who game the system

4) People who got successful with AIs are going to use their success and power to try to push their ideas and values on the world, and they'll often be successful.

In doing so they'll spread increasing amounts of self-serving disinformation throughout our MemeticEcosystem.

5) Future AIs will need to continue learning from the wide pool of human knowledge, but they'll be increasingly disrupted by the unchecked and uncontrolled disinformation they find out there among humans

6) Over time the AIs will dumb down, tracking the decline in human wisdom.