Friday, May 27, 2016

Robin Hanson vs Brain Gain/Drain

Dear Robin Hanson,

How's it going?   So... you wrote a new book... congratulations... kudos.  But, I'm probably not going to buy it.  In my previous entry I wrote that I definitely plan on purchasing John Quiggin's book about opportunity cost when it's published.  He's a liberal writing a book about opportunity cost.  You're not a liberal and your book is about...?  What's your book about?

Here's a clue that I found in my twitter feed today...




To be completely honest, pretty much everything about that article is horrible and terrible.  You had a wonderful opportunity to shed lots of light on an extremely important and fascinating topic but, instead, you did the opposite.  You delonged it.

Here's what I wrote to DeLong...


**************************************


So it will potentially be bad for the economy if robots become smarter? Would it also be bad for the economy if you become smarter? Would it also be bad for the economy if I became smarter?

What if the US imported a bunch of geniuses? Would this also be bad for the economy? If so, then brain drain would be good for the US economy. But I’m pretty sure that you’re not going to argue that brain drain is good for any economy. Yet, here you are, arguing that we should be worried about robots getting smarter. Increasing the total amount of intelligence in an economy is the opposite of brain drain.

What if a bunch of aliens landed in a spaceship? Would your preference be for them to be dumber or smarter than us?

Let’s say that you were stranded on an island with a bunch of people. Would your preference be for them to be dumber or smarter than you?


**************************************


Spoiler alert... DeLong didn't reply.

Check out this photo that I recently took...




On the left is Charlemange and on the right is Tiger.  Tiger is the very old neighborhood tom cat that my girlfriend decided to feed.  As a result... we now have possums, skunks and raccoons that regularly visit.  It's pretty impressive given that we really don't live in the country.

One of the raccoons, Crookedear, is surprisingly tame...





So what does all this nature have to do with smarter robots?  Well... it has to do with communication.  I know what the animals want... food... but they don't know what I want.  What do I want?  I'd love to have less weeds!  And it would be awesome if I was able to communicate the intensity of this preference to the raccoons.  They would pull my weeds... and I would give them food.

Can you handle some more nature?  Yeah?  Here you go... photos of plants.  The photos were taken by a fellow in India... Anurag Sharma.  I told my friend Michelle, who recently joined Flickr, to check out Anurag's photos and *favorite* the ones she really likes.  This was her response...

I did it all. Wow wow wow!!!  lots to see.  

Needless to say she *favorited* quite a few of his photos.  As a result, now Anurag in India knows that Michelle in America likes what he's doing.

The problem is that Anurag can't even take all the *favorites* that Michelle gave to him and use them to buy anything... not even a single grain of rice.  But I'm pretty sure that Michelle would be willing to give Anurag more than a few grains of rice in exchange for all the enjoyment that he provided her.

The website Medium is currently exploring ways to help certain writers to get paid via advertising... Revenue on Medium.

Here was my response...


**************************************


Right now it’s really easy to like a story. Why not make it just as easy to value a story?





If you value this story between 0 cents and 1 cent… then you would click the first heart button. If you value this story at 1 cent… then you would click the second button. Doing so would transfer a penny from your digital wallet to my digital wallet.

Once valuing a story is as easy as liking it… then I’m sure lots of people will be happy to contribute.


**************************************


How diverse is demand?   The doorbell just rang.  Went outside and found a box with an orchid inside...




More nature!  It's a hybrid... Dendrobium (Haleahi Twist x canaliculatum) x canaliculatum

I recently purchased it on eBay from ella-vate.  The thing is... I already have this hybrid!  I purchased it a couple years ago from The Orchid Trail.   Do I really need another one?   Here's a picture of the two hybrids...




They are different!  The one I recently purchased has a keiki on it.  Maybe I can try and trade it with shavedmonkey for his variety.

So how diverse is demand?

One would think that man could find enough variation in the orchid family, as it occurs in nature, to more than satiate his taste for variety. Yet man's appetite for variety is never appeased. He has produced over two times as many hybrids, in the past 100 years that he has been engaged in orchid breeding, as nature has created species in her eons of evolutionary effort. - Calaway H. Dodson, Robert J. Gillespie, The Botany of Orchids

There are around 30,000 species of orchids.  That's a huge amount of variety!  Yet, clearly it's not nearly enough variety!

Are you going to *favorite* any of Anurag's photos on Flickr?  Are you going to buy any orchid hybrids from ella-vate?  Well... she doesn't have any more for sale but The Orchid Trail still does!  I'm going to go out on a limb and guess that these activities don't match your preferences.

When robots are as smart as we are... are they all going to be the same?  Probably not.  Which means that they probably aren't all going to do the same things with society's limited resources.  Some robots are going to do more valuable things with society's limited resources than other robots.  As a result, some robots are going to have more money/influence than other robots.

The heart of the issue is the clear communication of demand.  People don't understand how important demand clarity is.  If they did, then we would be able to choose where our taxes go.  So the problem isn't smarter robots... just like the problem isn't more or less immigration.  The problem is that people don't understand the importance of facilitating trade.

We're assuming that robots are going to become more intelligent.  But do you have an explanation for why humans are more intelligent than other animals?  I do!  Here's my comment on  Evolution Q&A: Why did only humans become intelligent? by Greg Stevens


**************************************


You really did the theory of evolution justice!   I, on the other hand, really suck at doing theories justice.  Watch...

So... humans are exceptionally intelligent.   What is exceptional intelligence good for?  It's good for solving exceptionally hard problems.   But why did early humans, out of all the animals, need to solve exceptionally hard problems?    It's because out of all the animals, early humans had the greatest ability to (simultaneously) allocate the widest variety of resources.   This exceptional ability was the result of having hands, arms and... walking upright.

With quadrupeds... all four limbs are primarily dedicated to allocating a single resource... the animal itself.   But this specialization is a continuum that ranges from horses to raccoons to chimps.    Horses obviously have four legs.   All their limbs are quite specialized to allocating only the horse itself.  None of the horse's limbs are remotely capable of allocating other resources.    What about raccoons?  Do they have four legs?   Well, their front limbs are reasonably capable of allocating other resources.   Chimps definitely do not have four legs.   They have two legs and feet and two arms and hands.  They are quite capable of allocating other resources with their arms and hands.

As front limbs become less dedicated to only allocating the animal itself and more generalized to  allocating other resources... there's an increase in the total variety of resources that can be (simultaneously) allocated.  This creates a more difficult/complex allocation problem.... which requires more brain power/storage to optimally solve.   Well... a distinct advantage is given to exceptionally intelligent individuals.

Since you're fond of using lions as an example... let's compare them to zebras.  It would seem that the front limbs of the lion aren't as specialized to self-allocation as the front limbs of the zebra are.  Lions certainly use their front limbs to allocate themselves... but they also use their front limbs to allocate their prey.   But perhaps the biggest difference is that the mouths of lions are quite capable of carrying/allocating resources (food, cubs, other?).  Do zebras use their mouths to carry anything?   Not so much?  Therefore, lions are faced with more complex (allocation) problems than zebras.... and we should suspect that lions are more intelligent as a result.

So.... for lack of a better word... more "resourceful" body types put greater selection pressure on intelligence.   Humans are the most intelligent animals because our body types are the most "resourceful".


**************************************


Can we deliberately create truly intelligent beings without truly understanding the origins of our own intelligence?  If you have a better explanation for human intelligence then I'm all ears.

If we do manage to somehow create truly intelligent robots... and we do somehow end up with a bunch of unemployed humans.... then the superficial problem will be smarter robots.   But the substantial problem will be that we failed to effectively clarify the diversity of demand.  Except, this problem is hardly a new problem!

Based on the available evidence, it really doesn't seem like your book strikes anywhere near the root of the problem.  Which is why I don't plan on buying it.  And maybe you couldn't care less why I don't plan on buying your book.  No worries, don't mind me, this is just another prayer to Seldon.

No comments:

Post a Comment