Skip to main content


Are humans more powerful than rats?


in reply to Ben Weinstein-Raun

(sorry for the melodramatic / non-friendly depiction of a rat; I asked ChatGPT for an image of a rat ruling over humans and this is what it gave me)
in reply to Ben Weinstein-Raun

Weirdly, people don't seem that interested in this question. I guess I didn't clearly explain the motivation.

I spent most of today writing and journaling about general-purpose strategic power. Overall I think there are lots of tricky questions around strategic power, and they seem highly relevant for predicting AI trajectories. This piece was basically an attempt to consider a simplified case, where it's clear to me that humans have more strategic power than rats, but it's not obvious what tests one should use to demonstrate this, since it's not the case that we've eliminated rat problems.

in reply to Ben Weinstein-Raun

Reminds me of of the "you are bugs" scene in the three body problem.

> And as we go about changing the world to suit our preferences, the rats will remain unconsulted. It seems clear to me that rats will only get what they want, when what they want happens to be nearly-costless to humans.

This seems like it's making progress towards a formalization, though I think it still struggles.

If you imagine that covid virons were agents, then it seems to me that although there's a sense in which we're much more powerful than them, and you know, humanity could, if "it" wanted, defeat them, they can kinda get what they want without enormous costs to humans. And yet humans are still much more powerful than covid virons.

in reply to JP Addison

I'm not sure I understand the last paragraph; my guess is you're saying that covid virions are imposing large costs on humans to get what they want, and yet seem less powerful than humans; is that right?