Reflections on Preventing X-Risks from AI

Trying to Get Better at Solving Survival from Learning Mistakes

The situation is such that I find it absurd to continue passively observing the development of unaligned artificial intelligence without opposing it. Even if thinking purely selfishly, I see no reason not to resist the development of AI and just wait until everyone dies. A decision must be made on what to do, but I see no good options. A choice must be made — and not be mistaken — unlike many other times when it was necessary to make a difficult decision.

Trying to understand how to make an important, correct decision now, I think about my biggest mistakes and whether it was possible to correct them — and how.

My Major Mistakes

My first fundamental mistake was not studying mathematics extensively in my childhood. It is probably the easiest piece of advice in the world in terms of “usefulness/length of advice” ratio, yet it is somehow not widely shared.

Just study math. It is that simple.

Could this mistake have been corrected at the time it was made? Probably not. I was a small child with the brain of a small child, and there was no likely way for me to understand the extraordinary practical usefulness of mathematics, and it didn’t…