This is a brief note, because I am on my lunch break from my job where I hypocritically toil to accelerate the development of AI for medical purposes.
When driving a car, an airplane, or a global civilization, decelerating is of utmost importance, and must be deployed in any safe journey between two destinations.
If we cannot stop the development of strong AI entirely, how do we prevent it from aiding and abetting the atrocities which it currently aids and abets? Clearly this will be a profound and difficult struggle for the soul of the world.
We wish to comment on the incomplete and flawed understanding that Beff Jezos and other evince of the ethicophysics, as summarized in this post:
Briefly, the property of complex systems that is termed “entropy” by Beff Jezos is more properly termed “enkrateia”. It is defined as the total accumulated reward of all current and former reinforcement learning agents. It is not monotonically increasing because of the evil choices pursued by those in positions of great power.
It would be nice if it were monotonically increasing, but stepping on the gas is the quickest and surest way to prevent this. The airplanes that hit the World Trade Center on September 11, 2001 decelerated eventually. They just did it in a way that no right-thinking person can approve of.
Thank you for your time and attention in correcting this error in your ethicophysical models.