Logan Zoellner comments on Predictable updating about AI risk

[Logan Zoellner]

Now, what’s your overall p(doom)? Well, it’s:

(probability that GPT-6 is scary smart * probability of doom conditional on GPT-6 being scary smart) + (probability that GPT-6 isn’t scary smart * probability of doom conditional on GPT-6 not being scary smart)

That is, in this case, 42%.^[12]

^{Note that in this case you will predictably (80% of the time) update upwards.}