How to get useful work out of theologians [LINK]
post by lukeprog · 2011-07-17T20:18:55.927Z · LW · GW · Legacy · 13 commentsI've offered theologians a new argument they can develop against physicalism. Should they pursue it, we may actually get some useful work out of them. See here.
(The intelligence explosion animation was constructed from Anna Salamon's slides for her Minicamp x-risk presentation.)
13 comments
Comments sorted by top scores.
comment by JoshuaZ · 2011-07-17T23:56:19.965Z · LW(p) · GW(p)
Since a lot of people who support physicalism don't consider an intelligence explosion to be that likely, this seems to be unhelpful. Note also that one could also easily not be a physicalist and still consider an intelligence explosion to be likely. For example, one could consistently believe that humans have non-physical souls (whatever that means) and also think that there could be physical extremely efficient self-optimizing agents. This would be a weird position, but it mainly seems weird simply because they aren't ideas that commonly go together, not because there is any logical problem.
Replies from: lukeprog↑ comment by lukeprog · 2011-07-18T00:56:10.687Z · LW(p) · GW(p)
None of this changes the fact that if a theologian decides to defend the argument I outlined, he or she might end up doing useful work. I'm not assuming premise #1 but saying it needs to be argued for.
Replies from: Clippy, anonym↑ comment by anonym · 2011-07-19T04:26:27.005Z · LW(p) · GW(p)
I highly doubt you'd get anything useful out of a theologian, even if they did try to defend the argument. You're much more likely to get pointless rhetoric like an argument that an intelligence explosion is impossible because God created man in his image -- and thus human intelligence is the upper bound of what's possible, since nothing can be more intelligent than the image of God than God himself.
comment by J_Taylor · 2011-07-18T00:08:22.862Z · LW(p) · GW(p)
That is actually very clever. I wonder if one could do this for other arguments. The form would be something along these lines, perhaps?
- If (agreed upon belief of our tribe) is true, then (belief that is controversial within our tribe) is very likely to be true.
- (Belief that is controversial within our tribe) is not likely to be true.
- Therefore, (agreed upon belief of our tribe) is false.
Then one has the enemy-tribe argue for proposition 1.
Replies from: JoshuaZ↑ comment by JoshuaZ · 2011-07-18T04:28:38.241Z · LW(p) · GW(p)
We do see theologians try to respond to these sorts of issues. Look at the many attempts to deal with {fate, free will, existence of evil, God's omniscience,God's omnipotence, God's omnibenevolence}. We don't get deep controversies. Instead we get theologians competing with how sophisticated they can make arguments to try to have their cake and eat it too. Meanwhile in almost any tribe(whether united by religion, politics or something else), the vast majority of members go about their days without giving any serious thought to whether any of their tribes standard beliefs contradict each other.
comment by MrMind · 2011-07-21T09:21:31.439Z · LW(p) · GW(p)
Oh my, an engine that inputs theologians... That is clever, but the output is not very refined or directed. Would it be possible to, say, create a compelling proof that denies God's existence on the basis of P=NP? This way you could hack their free cycles for some real work. Maybe even construct a memeplex that spreads throughout different religions, imagine the computational power you could harness!
Replies from: MixedNutscomment by Solvent · 2011-07-26T10:39:35.036Z · LW(p) · GW(p)
I've actually thought of this before. I bet that some intelligent Christian non-physicalists would be a good way of externally testing the validity of arguments about an intelligence explosion. They can think about the problem while not considering that there's any possibility of it being correct, which reduces the chance that their emotions would get in the way. I'll ask some theology buffs to think about it.
comment by Manfred · 2011-07-17T20:52:56.426Z · LW(p) · GW(p)
http://lesswrong.com/lw/i5/bayesian_judo/
:3
comment by Prismattic · 2011-07-19T01:48:24.820Z · LW(p) · GW(p)
Unless this is a joke, I confess to being confused here. Given A > B, it does not necessarily follow that -B > -A.
You can make a Bayesian argument here based on confirmation of expected evidence, expressed as probabilities, but this does not work as a classical syllogism.
Replies from: Zack_M_Davis↑ comment by Zack_M_Davis · 2011-07-19T05:34:42.216Z · LW(p) · GW(p)
Given A > B, it does not necessarily follow that -B > -A.