Pascal's Mugging (2009) [pdf]
nickbostrom.com
Pascal's Mugging (2009) [pdf]
1–10 of 71 posts
Re: Pascal's Mugging (2009) [pdf]
#2Related from 2011: https://news.ycombinator.com/item?id=2932157
Re: Pascal's Mugging (2009) [pdf]
#3This is likely trending because John Carmack's recent post saying he was moving to AI references Pascal's Mugging: https://www.facebook.com/100006735798590/posts/2547632585471...
His Facebook post was also discussed in detail on Hacker News here: https://news.ycombinator.com/item?id=21530860
Re: Pascal's Mugging (2009) [pdf]
#4This is likely trending because John Carmack's recent post saying he was moving to AI references Pascal's Mugging: https://www.facebook.com/100006735798590/posts/2547632585471... His Facebook post was also discussed in detail on Hacker News here: https://news.ycombinator.com/item?id=21530860
I really love how this is could be seen as a demonstration that HN (or any other similar interest community, or probably the internet in general) functions like a hive mind, a collective consciousness. I've read that post by Carmack and I wondered exactly that "wonder what Pascal's Mugging is, interesting".
Sometimes you go around and wonder various things, but don't look them up or do them in the moment, and then sometime later your subconscious mind serves you up with an answer, perhaps when you are more relaxed you just think up of the answer, or it happens to come up in a certain context, the subconscious lights it up there. It might be a word that you see randomly in a newspaper, or you think of a person that was related when you met them etc.
And here the subconscious did the same wonderous thing, except it wasn't even strictly my personal subconsciousness, it was the group subconscious that found the information and presented it.
Re: Pascal's Mugging (2009) [pdf]
#5This is likely trending because John Carmack's recent post saying he was moving to AI references Pascal's Mugging: https://www.facebook.com/100006735798590/posts/2547632585471... His Facebook post was also discussed in detail on Hacker News here: https://news.ycombinator.com/item?id=21530860
I really love how this is could be seen as a demonstration that HN (or any other similar interest community, or probably the internet in general) functions like a hive mind, a collective consciousness. I've read that post by Carmack and I wondered exactly that "wonder what Pascal's Mugging is, interesting". Sometimes you go around and wonder various things, but don't look them up or do them in the moment, and then so…
But how useful is it? Pascal's Mugging was submitted to HN and discussed 8 years ago. If the collective consciousness keeps needing reminders of what it once knew, this is probably still inferior to a single intelligent person who reads a lot and remembers it all.
Re: Pascal's Mugging (2009) [pdf]
#6Earlier quoted context omitted.
I really love how this is could be seen as a demonstration that HN (or any other similar interest community, or probably the internet in general) functions like a hive mind, a collective consciousness. I've read that post by Carmack and I wondered exactly that "wonder what Pascal's Mugging is, interesting". Sometimes you go around and wonder various things, but don't look them up or do them in the moment, and then so…
But how useful is it? Pascal's Mugging was submitted to HN and discussed 8 years ago. If the collective consciousness keeps needing reminders of what it once knew, this is probably still inferior to a single intelligent person who reads a lot and remembers it all.
As a somewhat intelligent person, I still also do need reminders often. The trick is to forget mostly the unneeded stuff and maximally retain the useful stuff, and accurately discern between the two. The goal is not to remember everything.
Also it would be strange to expect this group consciousness to never need reminders when it continuously has new people added to it, who are unfamiliar with old things.
I think consciousness is more about active living rehashing of information and reupdating it, reupdating the worldview to adjust to constantly changing environment - not so much about building one single model that would somehow know everything. Such models tend to be stale or abstract and philosophical to the point of uselessness.
Re: Pascal's Mugging (2009) [pdf]
#7This can be easily resolved by considering that the victim of the mugging has finite resources. This is a usual remedy to problems were expected value alone gives stupid results (such as Pascal's Mugging). Something similar happens in lotteries where even if the expected value of buying a ticket is positive it is still not rational for an ordinary person to buy a ticket.
If I have $400 dollars I can't afford to take 1:1000000 risks that cost $200 each. I will go bankrupt with an enormous likelihood whatever the payoff. There is a minimum cutoff involving cost/probability below which it does not make sense to take up the opportunity.
There are links to similar theoretical ideas from the Pascale's Mugging wiki page - although from the casinos perspective not the gambler's - https://en.wikipedia.org/wiki/St._Petersburg_paradox#Finite_... and then https://en.wikipedia.org/wiki/Gambler%27s_ruin for example.
Most people will not take an 99% risk of going bankrupt in a game that will consume all their resource reserves; expected value as a statistic does not meaningfully capture the risk. Positive expectation, losing strategy.
Re: Pascal's Mugging (2009) [pdf]
#8This is likely trending because John Carmack's recent post saying he was moving to AI references Pascal's Mugging: https://www.facebook.com/100006735798590/posts/2547632585471... His Facebook post was also discussed in detail on Hacker News here: https://news.ycombinator.com/item?id=21530860
I really love how this is could be seen as a demonstration that HN (or any other similar interest community, or probably the internet in general) functions like a hive mind, a collective consciousness. I've read that post by Carmack and I wondered exactly that "wonder what Pascal's Mugging is, interesting". Sometimes you go around and wonder various things, but don't look them up or do them in the moment, and then so…
I have noticed this pattern before: Someone mentions a topic deep inside a thread on HN and next thing you know that topic is on the front page.
Re: Pascal's Mugging (2009) [pdf]
#9This can be easily resolved by considering that the victim of the mugging has finite resources. This is a usual remedy to problems were expected value alone gives stupid results (such as Pascal's Mugging). Something similar happens in lotteries where even if the expected value of buying a ticket is positive it is still not rational for an ordinary person to buy a ticket. If I have $400 dollars I can't afford to take…
That's a good argument against seemingly plausible but risky approaches.
This particular situation, of someone just pulling "it could be true!" out of their arse, can also be solved by framing things as "the more utility you claim, the less likely it seems and disproportionately so".
IE, if the chance of getting X from the scoundrel is less than 1/(X^2*C), even integral of all the offers together winds up very small.
Re: Pascal's Mugging (2009) [pdf]
#10Earlier quoted context omitted.
I really love how this is could be seen as a demonstration that HN (or any other similar interest community, or probably the internet in general) functions like a hive mind, a collective consciousness. I've read that post by Carmack and I wondered exactly that "wonder what Pascal's Mugging is, interesting". Sometimes you go around and wonder various things, but don't look them up or do them in the moment, and then so…
But how useful is it? Pascal's Mugging was submitted to HN and discussed 8 years ago. If the collective consciousness keeps needing reminders of what it once knew, this is probably still inferior to a single intelligent person who reads a lot and remembers it all.
The population using HN takes in new members continuously, and the fraction who read everything posted eight years ago is very small.