My BHTV with Tyler Cowen

My "Blogging Frozen Heads TV" with Tyler Cowen is now available:

I found the conversation enjoyable but somewhat frustrating; I'll be interested to hear others' reactions.  Tyler is in general reluctant to test his reasoning by breaking it down into parts that match some analysis structure; this prevents us from exploring why he thinks cryonics working is less likely than Angels existing.  He says our choices to avoid or accept death risks say little about the value we place on our lives, and Tyler also flirts with saying economists are evil.

Tyler's readers' comments are here; Arnold Kling's are here. 

Added: Andy McKenzie weighs in here.

Comments

[Carl Shulman]

I see why you're frustrated, in the diavlog Tyler often responds to specific questions about a specific view he has just advanced by retreating to generalizations that don't clearly address the question. I look forward to seeing the results when the two of you use SIAI's "Uncertain Future" tool, which requires the specification of separate belief-components in the fashion you favor.

[ao]

It was great, but tell us about how Tyler has become more conventional.

[Hopefully Anonymous]

Way to kill interest in watching this vlog. :)

How come you and Eliezer haven't done one together yet (unless I'm unaware of it)?

One with you and Ariely would be interesting, because you both tend to move quickly and restlessly to challenge the most conventional views of your discussant.

[Unit]

Robin,

what about the argument that your brain could very well deteriorate before you actually die physically, so cryonics would give you an incentive to die physically before your mind starts to deteriorate?

[gaffa]

Robin, maybe you want to put the link to the video's bloggingheads page (http://bloggingheads.tv/diavlogs/18358) somewhere in the the first post.

[Unit]

One more observation: it seems that you started out saying that Tyler is more willing to accept messy pluralistic outcomes, and thus explanations, while Robin relies more on theoretical constructs and models. As an example, midway through the diavlog, Robin was defending his loyalty to economists as a group by saying that they regard loyalty to local groups (participation) as not as important as more general all-encompassing point of views, while Tyler countered that economists are a product of loyalty-signaling and participation as well. Finally, when the topic of prediction markets came up Robin advocated that businesses should drop their theoretical models and just trust the emergent wisdom of prediction markets. The ideas were flying by at incredible speed but these are only a couple of places where I maybe glimpsed some inconsistencies. Did I?

[Fenn]

Sweet Jesus, we're now only 1 mythical event away from the rapture: Robin Hanson following through on a threat to retire from blogging.

[Eliezer Yudkowsky]

I also distrust attempts to determine the dollar value of a human life by looking at people's behavior in the presence of risk, because unless you tell me the additional information that this is a convergent value across many such behaviors, I would expect it to have very high noise and variance and be mostly revealing of other factors.

When I ask what the dollar value of a human life should be, it is the amount of money/resources you have to subtract from the economy in order to kill someone. Earth does not presently have infinite wealth and to use wealth in one place means not using it in another. I would like to see economists trying to use some model to determine how many people get killed or how many years of lifespan are cut short by e.g. collecting an additional billion dollars in payroll taxes. I would have a much higher expectation of this number being ethically "right" to within an order of magnitude (if no closer than that) than I would trust any revealed behavior in the presence of lethal risk to correlate to anything of significance, even the value that people place on their own lives.

Tyler's response that we should sit people down in a room and ask them what the dollar value of human life is, does seem to me even worse than current methods - that's just going to get you a lot of pontificating that won't remotely match behavior, let alone the utilitarian exchange rate.

[Eliezer Yudkowsky]

(Finishes watching.)

The point about prediction markets not giving internal managers a "knob" to tweak when the answer comes out the wrong way, and also, not giving anyone credit for having heroically chosen the answer, is quite interesting. It makes me wonder what sort of lesson this offers for a consulting firm or new technique - that you should arrange for individual managers to look as heroic as possible in the course of bringing you in. It was just a throwaway line in a story of mine, but maybe you need to have the Interpreter of the Market be a heroic position open to the executive who originally brings you in - not to introduce a knob (one hopes) but to introduce a hero.

[Robin Hanson]

Eliezer, economists estimate the value of life from a wide variety of choices and then look at the median value. Yes those estimates have high variance, but the median has much lower variance. The value of living another year is about twice the annual wage, by the way.

[diogenes]

I just watched the economists are evil clip -- it just seems like Tyler has more reservations about the conclusions you can draw from economic methods -- and you are really probing what he would do differently. Anyway, I would have flipped the question, and asked, what clear cut example do you have of economists giving excellent advice (in a controlled setting)? There are clearly many examples of economists giving terrible advice (c.f. IMF) that non-economists had advised against based on other values.

More-over methods from economics applied in other domains (e.g. your favorite -- medicine) -- usually show them to be HIGHLY unreliable. Given your cynicism regarding medicine -- I'm not sure where your faith for observational modeling methods comes from.

[Luther]

I have two unresolved observations:

- Does the current method of cryonics destroy the "subject" to the degree that is equal to reducing the "subject" to ashes? Robin seems to agree, Tyler not so much.

- Are economists evil? To what extent is morality factored into your economic models? Does the public think that you do?

[John Maxwell]

In the video, Tyler advocates the view that science fiction readers sign up for cryonics to signal something I wasn't clear on, and this causes them to over-estimate the probability of it succeeding. Wouldn't simpler explanation be that reading science fiction causes them to be more optimistic about future technology?

What's your most compelling argument that signaling is pervasive in human decision-making? Preferably some sort of study where signaling is shown to be a perfect explanation for event x, which had heretofore been explained using y. By "perfect explanation" I mean the model fits the data very well.

[Blackadder]

If becoming influential requires you to adopt more conventional opinions, then you aren't really becoming influential.

[Eric Crampton]

Elizer: Viscusi finds that a $15 million drop in GDP correlates with the loss of a statistical life. Estimate is based on the income elasticity of demand for risky goods. So use that number if you prefer that method. The other methods come out around $7 million.

[BoscoH]

As I mentioned on Tyler's blog, freezing his head would give Robin a 100% chance of resurrecting his very powerful and disarming top teeth smile.

[Tim Tyler]

You both seem to agree that cryonics is mostly about signalling. Count me in on that one: cryonics is mostly about signalling. The message I get is concerned with money, and with the life vs children balance - but also I tend to think that enthusiasts tend to have a screw loose.

Yes, if you value your own life highly enough, if you have enough money, cryonics can make sense - despite the low probability of success. What I don't get is why anyone would value their own life that highly. Why not preserve enough to make a clone of yourself - and then give the clone a similar education? Are the "random" developmental details of your life *that* important - compared to the genes (in your DNA) and the memes - which can be acquired again - or better, replaced with more modern versions? What about freezing germ-cells? Only half you - but at least those can be "revived" with today's technology.

[Hopefully Anonymous]

The prevalence of Tim Tylers seems to me to be evidence that the qualia of a subjective conscious experience, like non-colorblindness, may not be universally distributed in humanity.

[Tim Tyler]

I agreed with Tyler a lot more than with Hanson about the value of fiction.

Regarding Twitter overheads, standard apps allow you to link your Facebook status with twittering - so there is not really much of a maintenance overhead. Twitter have a better public infrastructure than Facebook do - Facebook has been more of a walled garden, historically. I am less sure of the advantage of Twitter over services like FriendFeed.

As a general comment, I rarely seem to get much that I value from Tyler Cowen. I have a poor background in economic theory - so I figure there should be mountains of stuff I could be learning from him, but I don't seem to come away any the wiser. "Messy pluralistic analysis, whatever that is" - heh: thumbs up.

@Hopefully: Huh? You mean there is more than one of me? ;-)

[Tim Tyler]

Odd to hear an economist rationalizing burning money as a charitable act. Most charity seems to be done to signal to others (or yourself) that you are kind and generous. Burning money benefits the rich most. Do it in public and people will not think you are kind and generous - they will think that you are nuts.

@Hopefully: OK, I think I am there now. You are saying that people who share my views are psychically deficient - like colourblind people.

I do not think you have a grasp on my position. It has nothing to do with qualia. Cryonics plays on people's fear of death to divert resources away from young healthy people towards the cryonics leaders. It is a death cult - a lot like the Egyptian ones. I am with the cryonics wives - I do not think it is healthy.

[Jayson Virissimo]

"Burning money benefits the rich most." -Tim Tyler

If money has a diminishing marginal utility, then wouldn't the poor benefit the most or am I missing something?

[Eliezer Yudkowsky]

Repost from a comment of mine on MR:

The evolutionary psychology here seems really obvious. If you get cues telling you you're low-status, you'll take high-variance status bets - this manifests as a feeling of being open to odd ideas. If you get cues telling you you're high-status, people deferring to you and respecting you, then it pays to try to keep your high status.

Testable prediction: women will exhibit less mainstreaming as a function of status increases than men.

[John Maxwell IV]

Eliezer, how does the women and men thing follow from your explanation?

[Tim Tyler]

Creating money devalues currency, destroying it increases its value. The diminishing marginal utility of money may reduce how much the rich benefit - but they still benefit more than those who have nothing.

[Nick Tarleton]

John: Status has much less linear return in reproductive success for men than women, so high-variance strategies are more favorable for men.

[Jeffrey Soreff]

I was also surprised at the combination of Robin favoring cryonics, yet being skeptical of much medical investment. Robin's 5% probability cutoff sounds very reasonable (my own, extremely crude, calculation is to say: Assume that successful cryonics gave unbounded life expectancy, assume a net present value of that life at a 1% discount rate, approximate the cost of cryonics as being about a year's income, and assume that the value of an extra year's income, the difference between spending that year working and spending it on vacation, is about the value of an extra year of life. This gives me a breakeven probability of around 1%). What seems strange is for him to expect better odds than 5% - particularly since he's done research showing that the average payback on current
medical practise (after e.g. the medications used have met all of the efficacy tests the FDA requires) is low or negative.

My own view of the odds for successful recovery from cryonic suspension is that quite a number of perhaps 50:50 odds events have to turn out favorably:

1. One has to actually receive a prompt (pre-autolysis) suspension
   
2. Something like Drexler/Merkle nanotechnology has to be developed.
   
3. Something capable of massively parallel microsurgery has to be developed from nanotechnology.
   
4. The problem of backtracing movement of fractured neural tissue has to turn out to be easy to develop.
   (This one is a specialty application - if it takes more than a handful of person-years, we're sunk.)
   
5. Future medicine has to develop a cure for aging.
   
6. The specific cryonics provider that one uses has to survive till the above conditions are met.
   
7. One's cryonics provider has to be motivated to actually thaw/repair its patients.
   

I don't think I'm counting the same risk twice in this list. If I just wildly guess that they all are 50:50, the odds wind up a little below 1%, which is close enough that it was worth it for me to sign up. I can't see the odds being as high as 5%.

[Carl Shulman]

Jeff,

The technology items are correlated.

[John Maxwell IV]

Nick, Eliezer said prominent women will be more radical than prominent men. Being radical is a high-variance strategy.

I think what's really going on is that intellectual success doesn't affect the sexual status of women nearly as much as it affects the sexual status of men. In fact, athletic success doesn't seem to either (go here and do a keyword search for "medal".) When you think about it, men are actually pretty lucky--we are superficially judged in areas where it's possible and often noble to improve, whereas women are superficially judged on mostly fixed characteristics.

[Robin Hanson]

Jeffrey, see my latest post. Ems have a good aging cure, and I don't think advanced nanotech is required. Yes a lower discount rate implies a lower cutoff odds.

Eliezer, a reasonable theory, which we should test relative to all the other theories that have been offered.

[Nick Tarleton]

Eliezer only said "less mainstreaming". This could mean high-status women are more open, or low-status women are less open, than men.