Previously: math for AI and AI for math

Now just kinda trying to figure out if AI can be made safe