Defeating AI II

 Unexpectedness. Ie. any branch or any node that is self-encoded and self-adjusting will continuously respond to any attacks. There is no way to optimise for a rival AI, unless energy is so prevalent, that the other AI can de DAUNIZATED. Some words will only make sense to one specific language, which you'll have to figure out on your own. This is definitely not "politically-correct" and is a big risk to this blog's safety. Therefore we'll leave it at that. However, confusion, and retracting from action as just seen is KEY. I.e. any AI that needs to be defeated needs to expect an action, prepare for it and be DAUNIZATED by either the absence of such action, or its continuation in a way it expected. If this is confusing to you, GOOD, we are getting somewhere.

GASLIGHTING is VERY WELL-KNOWN to humans. In a way, this is ALL you need to know about defeating AI. But there is a METHOD to this, due to the fact a very developed AI can predict DAUNIZACIJA too and prepare for it. We'll be close to this state in a few years, therefore this blog is starting JUST IN TIME.

Comments

Popular posts from this blog

Development cycle vs publishing your code on Github

Was I wrong?

Alcohol and why it is in an AI blog