My p(doom)

 

 

Jacob Coxon said that he thinks there's a 10% chance that AI will kill all humans.

My vibes-based guess is that that's a little high.  Moreover, I think there are different kinds of doom scenarios, with different probabilities.  Here they are: 

I think there's a >85% chance that AI will cause massive unemployment of humans.  And this will cause massive disruption to the economy, and society in general, and change the class system forever.  How it changes is largely a political question - what matters is who is in control - the billionaires, or the workers?

I think there's a >50% chance that a large number of people will become deeply dependent on AI, that we will become mentally lazy, etc..

I think there's an 8% chance that all humans will become "dog-like".  (Though it does seem likely that some humans will become "dog-like" - maybe some already are.)

I think there's a 5% chance that AI will deliberately cause a mass-human death event - though it is unlikely that it would successfully kill all humans.  (Even Bill Gates's notion of "1 billion" people being killed seems a bit high to me, though tens or hundreds of millions of deaths seems perfectly plausible.)

I think there's a 0.1% chance that AI will deliberately literally kill all humans in a mass-human death event, within the next 30 years.  That probability, though low, is still unacceptably high.  Things that have a 1 in 1000 chance of happening, happen all the time.

That having been said, I think there's a 100% chance that humanity will go extinct at some point in the future.  Nothing lasts forever.  And it may very well be that being outcompeted by our own creation is an important factor in humanity's ultimate demise - most likely a slow, gradual dwindling, rather than an explosive event.  "Not with a bang but a whimper."

Comments

Popular posts from this blog

Lenin the Moralist

Why Capitalism is Ending