[ExI] AI thoughts from Twitter

Keith Henson hkeithhenson at gmail.com
Sat Apr 6 01:39:15 UTC 2024

Eliezer is an old-timer on this list, but he has not been here for
many years.  He is one of the main voices warning about the extinction
risk from AI.

[EY]  But if you think it's okay for Google to kill everyone, but not
okay for a government to do the same -- if you care immensely about
that, but not at all about "not dying" -- then I agree you have a
legitimate cause for action in opposing me. Like, if my policy push
backfires and only sees partial uptake, there's a very real chance
that the distorted version that gets adopted, changes which entities
kill everyone on Earth; shifting it from "Google" to "the US
government, one year later than this would have otherwise occurred".
If you think that private companies, but not governments, are okay to
accidentally wipe out all life on Earth, I agree that this would be
very terrible.


I have a logic problem with your analysis. A super intelligent AI is
going to be able to project the consequences of its actions, so it
seems unlikely that it would accidentally wipe out humans or life.
That leaves intentional which seems unlikely as well.

It seems probable that you read The Revolution From Rosinante by
Alexis A. Gilliland. The story is chuck full of AIs. The AIs relation
to humans is headed in the direction of our relation to cats. One of
the AIs remarks that God created humans as a tool to build computers,
something that God could not do without violating his own rules.

But consider your worst projection, that AIs kill all humans or even
all life on Earth. Do the AIs replace humans as possibly the only
thinking items in the universe? Do they go forth to the stars? Do they
remember us?

{This had a reply}

Daniel Houck
Mar 30
The majority of Eliezer’s argument is about why it *is* likely for an
AI to deliberately kill everyone.

And then: they go to the stars; they “remember” humans in the sense
that they have data, but rarely access it; and they are “thinking”
“entities” but not like you’re thinking.

There is a EY rant here https://newatlas.com/technology/ai-danger-kill-everyone


More information about the extropy-chat mailing list