3 ms·
Beware if someone assigns a p(doom). There is zero chance they know and 100% it came out of their backside. The only plausible p doom is a range of >0 and <10
by m3kw9 3y ago
Beware if someone assigns a p(doom). There is zero chance they know and 100% it came out of their backside. The only plausible p doom is a range of >0 and <100
They really need to listen to themselves when they say there is a 50% chance we all die from AI
- mitthrowaway2 3y agoThat's not how probability works. Probability already reflects your uncertainty. There's no "knowing the correct probability" except in very specific and rare situations.
- esafak 3y agoHe's a frequentist and you're a Bayesian, so you're both right.
- m3kw9 3y agoIt looks like you are agreeing with me in that for someone to assign a P to such an uncertain event
- mitthrowaway2 3y ago... I don't think you finished your sentence, but what I'm saying is that people should be (and subconsciously are) always assigning a probability to every event, or at least every event that rises to the level of their attention, regardless of how little concrete information they have about its likelihood of occurrence. If they have very little information, or a lot of information, that will be reflected in the numerical value of that probability. If an event is very uncertain, it means that the probability should be very high in entropy.
- Engineering-MD 3y agoWhen dealing with black swan events, prediction is always difficult. However, some scenarios can be thought to be more plausible than others. I always interpret these as relative figures of plausibility for comparison rather than actual probabilities.
- m3kw9 3y agoAnd how does that help anything other than to forward their agenda by making a high p(doom) that was pulled straight from their wazoo