Comments on: Disagreement Case Study: Hanson and Hughes

By: Robin Hanson

Robin Hanson — Tue, 12 Jun 2007 23:22:59 +0000

Peter, I’ll look out for a chance to disagree about the repugnant conclusion.

By: Peter McCluskey

Peter McCluskey — Tue, 12 Jun 2007 21:53:37 +0000

Robin, I’m a bit disappointed that you only analyzed the initial disagreement, which seems to result from a relatively common communications problem.
I’d be more impressed by an analysis why people disagree with you over the “repugnant conclusion” of utilitarianism, which seems likely to cause some important and persistent policy disagreements.
My intuition causes me to disagree with most people and say that what we should try to maximize is some combination of total welfare and average welfare, weighted according to whatever set of preferences evolution happens to have built us to value. This doesn’t satisfy me the way a less arbitrary rule created by a wise deity would, but until I can find such an external source of a rule for what to maximize, it’s hard for me to see why I should abandon the values evolution appears to have given me.

Matthew, here are some play-money prediction market contracts on SENS milestones:
http://www.ideosphere.com/fx-bin/Claim?claim=MMPost
http://www.ideosphere.com/fx-bin/Claim?claim=MMrev
http://www.ideosphere.com/fx-bin/Claim?claim=mtDNA
But long-term contracts on that exchange (and probably all play-money markets) tend to be biased to be too close to 50%.

By: joeo

joeo — Tue, 12 Jun 2007 20:16:11 +0000

Robin’s paper does seem like a case where protectionism could be warranted. It isn’t clear to me that human beings are better off when they are unable to do any productive work.

Treating uploads as having a different moral status than people may be reasonable. Killing a person is the complete destruction of that person. Killing an upload means losing the value of the hardware plus whatever memories since the last backup.

By: TGGP

TGGP — Tue, 12 Jun 2007 01:21:11 +0000

I agree with Brin on the inevitability of the erosion of privacy (and I've never understood why that was supposed to be so horrible) and I share his dislike of romanticism (though I don't buy into his dichotomy that he uses to replace the existing flawed dichotomy), but his frequent rambling about aristocracy indicate that he has bought into an all-too-prevalent myth that is quite silly for people of the present to care for.

By: Robin Hanson

Robin Hanson — Mon, 11 Jun 2007 23:02:03 +0000

Matthew, given the legal and other barriers to real money long term markets on such topics, they will probably not happen anytime soon unless some patron pays to create them.

By: Eric Falkenstein

Eric Falkenstein — Mon, 11 Jun 2007 23:01:04 +0000

I remember reading a book by Michael Moore and wanting to critique it for a debate with my brother. I quickly found that by page 15 or 16 the author already asserted or assumed so many points I disagreed with that my critique would have exploded. What’s nice about academic debates is they at least are focused on a few differing assumptions–most practical debates are so unfocused that it is impossible to begin to address them.

By: Matthew C

Matthew C — Mon, 11 Jun 2007 22:58:35 +0000

Robin et. al.,

I’m curious. Are you aware of any prediction markets that might offer contracts on AI, brain simulations, downloads, SENS, etc?

I notice Tradesports has offered a market for various US election outcomes, for whether we would find Saddam Hussein and Osama Bin Laden by a particular date, etc. I have to imagine that many of the bettors were US citizens. Given that, it seems quite feasible that someone outside the US could offer up a betting market for “Singularity” predictions, and perhaps many other questions of scientific interest.

By: Carl Shulman

Carl Shulman — Mon, 11 Jun 2007 19:51:15 +0000

Robin,

That’s why I specified a fixed number of such requests per communique, and a limited response set.

By: Robin Hanson

Robin Hanson — Mon, 11 Jun 2007 19:40:41 +0000

Carl, the norm of expecting each side to reply to every comment by the other side quickly results in an exponential growth of replies. If you want to force the other side to respond to a particular point, the best way is to make that point the only point you make in a particular step of the exchange.

By: Carl Shulman

Carl Shulman — Mon, 11 Jun 2007 19:27:09 +0000

“a dismal, elitist utopia … the division of society into a mass of well-fed plebes and a superpowerful elite … the enormous population of uploads would be forced to work at very low subsistence wages – the cost of their electricity and disk space – ruled over by a very few of the most successful of the uploads.”

The problem is much worse than just “reading between the lines,” as contextual signals are used to produce interpretations that don’t simply supplement the actual text, but are actually inconsistent with it.

The “very few…successful uploads” discussed in the paper are the reproductively successful, i.e. the “enormous population.” Whether we count uploads as individual copies or copy-families, the successful ones can’t simultaneously constitute the majority and the minority, and James seems never to provide an explanation of how the criticism could apply to your paper.

I appreciated David Brin’s meta-comment: if disputants selectively respond to points, they can do so in a biased fashion, avoiding two-way engagement on the most important issues. For long letters a 10-item list of propositions could be attached, and recipients would reply with agreement, disagreement or an explanation of why neither made sense for each numbered point. While perhaps not stylistically ideal, this could make failure to engage more transparent.