= 27d567bc2d48d9f40e9ccc20ea370b15)
GDC3 comments on Confidence levels inside and outside an argument - 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 (174)
I speculate there's at least two problems with the creationism odds calculation. First, it looks like the person doing the calculation was working with maybe 60,000 protein molecules rather than zillions of protein molecules.
The second problem I'm having trouble putting precisely in words, concerning the use of the uniform distribution as a prior. Sometimes the use of the uniform distribution as a prior seems to me to be entirely justified. An example of this is where there is a well-constructed model as to subsequent outcomes.
Other times, when the model for subsequent outcomes is sketchy, the uniform distribution is used as a prior simply as a default. Or, as in this case, it's clearly not an appropriate prior. In this case, the person is probably assuming that all combinations of proteins are equally likely (I suspect this assumption is false.)
Isn't the problem more like: they are ignoring the huge number of bits of evidence that say that cells in fact exist. They aren't comparing between hypotheses that say cells exist. They are comparing the uniform prior for cells existing to a the prior for only random proteins existing. They sound more like they are trying to argue that all our experiences cannot be enough evidence that there are cells, which seems weird.
This is a misinterpretation. The argument goes like this:
True statement: There is lots of evidence or cells. P(Evidence|Cells)/P(Evidence|~Cells)>>1.
False statement: Without intelligent design, cells could only be produced by random chance. P(Cells|~God) is very very small.
Debatable statement: P(Cells|God) is large.
Conclusion: We update massively in favor of God and against ~God, because of, not in opposition to, the massive evidence in favor of the existence of cells.
This is valid Bayesian updating, it's just that the false statement is false.
Upvoted for successfully correcting my confusion about this example and helping me get updating a little better.
Edit: wow, this was a really old comment reply. How did I just notice it...