almost 8 years on 2009-05-06

VALUE OF INFORMATION VS. UTILITY CREATED BY INFORMATION


(diffusion from high value, low utility - to low utility, high value)

a few weeks ago at our brooklyn y+30 meetup a gentleman came up to me and said that he had been reading my blog, but took serious issue with my definition of the value of information...

i generally like to state that the value of information is based on improbability.  information that is perfectly probable has no value, information that is highly improbable is almost priceless. 

this generally translates to saying that the information's value = f (speed * accuracy).  perfectly fast perfectly un-trustworthy information is worthless, perfectly trustworthy information that is un-timely is worthless -- value exists in degrees of speed and trust. 

this gentleman posed a question that at the time i didn't really answer, but i now want to take a second to actually fully respond.  he asked, (paraphrasing) "i admit that your definition covers some types of information, but there is this other type of valuable information which gains value the more people know about it.  for instance, knowledge that "smoking is bad for you" - if one person knows it is a heck of a lot less valuable than if everyone knows."

i actually totally disagree, but the question brings up a set of very interesting points worth discussing.  if you were the only person in the world that knew that smoking is bad for you - and you had an effective way to distribute that information (speed * expected accuracy), you could make a lot of money off of it.  if everyone knew that smoking was bad for them, no one would make any profit, the information would be worthless.

in the smoking case, what does grow with distribution is total social utility.  no question, if only one person in the world knew that smoking was bad, that would yield far less total utility than if everyone knew. 

the key is to remember that utility and value are two separate (though highly related) elements.

at the surface, this mismatch might sound like a problem/or imply that we should think of ways to actively convert valuable information with low total utility into worthless information with very high total utility... 

the good news is that when you bang on this phenomenon, you realize that that is actually precisely what a good functioning information economy does!  in fact, the core principle which drives information is diffusion from very high concentration high value pools with very low total utility, to very low concentration low value footprints with very high total utility.

we rely on innovators, scientists, and sometimes even thinkers, to be incentivised to not only discover new bits of information, but then to be incentivised to share their findings and harvest the profits. 

this doesn't always function properly.  in fact, in a very interesting yale open course catalog lecture robert shiller (econ 252) provides an amazing example of how this breaks down.  apparently, reading through fables and historical accounts, it appears as though individuals discovered the concept of probability countless times throughout human history, but each time the knowledge wouldn't spread and died quickly up until the renaissance.

why?  shiller argues that because understanding probability concepts before the rest of the world caught on was so incredibly valuable that it was guarded very closely.  individuals would become very powerful based on their discovery in their lifetimes, but then would take the secret to the grave.  this is precisely the type of information-scape that we do not want to occur... 

so, we should firmly keep in mind with any ip regulation/patent law/etc. that the whole idea should be to maximize the creation of valuable information, and let that value propel an economic harvest from high to low, to propel utility of the information from low to high.  when and where that dynamic breaks down there will be trouble.


original swl blogposts and letters 2007-2010