Hello, SFIA audio listeners. In this month's Nebula exclusive, big alien theory,

we're asked at the reason alien civilizations might be rare is because most aliens are huge.

To hear it and every episode early and ad-free, plus hours of bonus content, check out

go.nebula.tv slash Isaac Arthur and use my code, Isaac Arthur, loomed in our future and seem ever closer,

but it may be that the dangers of the future can reach back into the past itself, and even

without a time machine.

Greetings and welcome to Science and Futurism with Isaac Arthur, for an explanation of one of the

more perplexing and

potentially unsettling concepts that emerge when we contemplate the future of artificial

intelligence, Roco's Basilisk. This thought experiment involving superintelligent AI,

a causal or retro causal events, mind uploading and simulation, and the possible future torment of folks

who oppose AI is one of the most controversial topics out there. Indeed, many think even

speaking about it is a bad idea, and today I'll be explaining why it is not, probably,

and why it remains a good thought experiment in spite of a fair number of flaws that

we'll outline today too.

It's a topic that's prone to getting overhype and clickbait discussion, mostly

either drumming up its fear factor or explaining why they don't think it works, and we'll cover

those, but we'll be deep diving at it and some similar scenarios today, and indeed, while it has a lot of

legitimate criticism, some of those are flawed too. At its core, Roko's basilisk posits posits

a future in which an omnipotent artificial intelligence, designed with the capability to enforce

retroactive consequences, backward in time, chooses to punish those who did not contribute to its creation.

It is a concept that stretches our understanding of causality and morality.

...