= 27d567bc2d48d9f40e9ccc20ea370b15)
gwern comments on How Likely Is Cryonics To Work? - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (122)
This is not the first time I've seen an attempt at calculating it; I will note that it's interesting that the more steps you add, the lower your probability gets. It's a strange kind of analysis that can only go downwards...
Could this just be that people overestimate how likely things are until they really start getting into all the things that could go wrong?
That is true, but this style of analysis is predicated on a sequence of steps, each one of which must succeed, and hence the more steps you have, the lower the end result probability must be; if you were just correcting for overestimation by analysis, then there ought to be analyses or points where one realizes one has been too pessimistic and increases the probability.
But that can never happen with this kind of analysis: the small result is built into the conjunctions. If one realizes one was wrong in giving the probability of a particular factor, well, one can just 'fix' that by breaking it into some more substeps with <1 probability!
And if you look, there are illegitimate conjunctions in the OP. For example, 'cryonics company is out of capacity', besides being way too high (notice that '86% chance 'something goes wrong in getting you frozen' fails a basic outside view test - suspensions do go wrong fairly frequently, but not anywhere close to 86% of the time!), is a false conjunction; if you're paid up for Alcor, why can't CI handle your case or vice versa? Cryonics companies have taken patients off each others' hands before.
Some of my points seem unlikely enough now that I'll probably remove them. For example:
One could also break a step S with X% probability into several steps that combine to less than X% probability if after looking at all the ways S could happen you decided that they combined to be less likely than you originally thought S was.
If we have actual data on this, this should replace most of my first section. Can we get it? Though the people in a position to collect the data have a lot to lose by admitting poor numbers.
Also, some of my 86% for the first section includes things that we can't tell yet if they worked out all right:
And then there are things that are not currently a problem but could become one:
Actual data on the fraction of the time someone signed up for cryonics is actually suspended in what we think was the correct way would be really helpful, though.
http://www.alcor.org/cases.html seems like a good starting point.
Ugh. Some of that makes for very horrifying reading. One of the ones labeled worst case scenario makes me want to track down the people who did the autopsy and punch them.
I did say 'fairly frequently'.... Nor does long involvement necessarily save one; Mike Darwin was rather angry at Ben Best over how he botched Curtis Henderson.
No, in this disjunction of conjunctions, the more details of any kind you add, the less likely a favorable outcome looks. If we expect reality to be unbiased, we should also expect some ratio of favorable to unfavorable details, which, ceteris paribus, should be maintained as we go to higher granularities of detail.
In other words, "motivated stopping" and "motivated continuation" should not, together, be a sufficient explanation for the results of an analysis.
Say I went into this thinking my chance of being frozen correctly was 95%. Now, with more details on what has to go right for this to happen, I think 86% is a better estimate. Details don't have to make things less favorable. They just usually do because we are optimistic.
EDIT: that "correctly" above should have been an "incorrectly".
...but then you think and research for longer, and find out in even more detail what could go wrong, and your estimate drops to 80%. If you can predict which direction your belief will move in the future, something went wrong somewhere.
I'm sorry, I meant to write "incorrectly".