More thoughts on Occam's Paradox

I’ve been re-reading my Occam’s Paradox post as well as the comments and trackbacks (which are excellent – please click through them if you have a minute).  I fell a little short of really saying what I originally intended for the post, which was that I think that we have a tendency not only to make things more complicated than need be, but also to focus on too many things (and therefore the wrong ones).  As a result we try to assimilate too much data to make decisions (not recognizing the massive diminishing returns on this effort) and try to pay attention to too many things.  I wrote a post a while ago about trying to cram too much information into financial models that argued that more complex models are not necessarily better or more accurate.  I’m realizing now that I’m connecting the dots here that this is exactly the type of behavior I’m talking about in Occam’s Paradox. [By the way, there’s an entire post to be written about how VC’s play into this in their decision making (how many customers do we need to talk with, how many models should we run, etc before we make an investment decision?) – but I don't think I should go there in this post].

I came across two posts that I wanted to bring your attention to that speak to this general subject area and that are both worth taking a look at. The first is “Decision Making” by my fellow VC Blogger Will Price (be sure to click through this blog to the link for the Bastardi and Shafir paper – On the Pursuit and Misuse of Useless Information).  The other, Focus is the New Black, is by Paul Kedrosky.  The title speaks for itself – it’s a great read.

Please keep the comments (both public and private) coming - I'm enjoying this thread very much.

Comments

[Leonard Kish]

Seth,

I've been meaning to comment about this for some time, but I knew it the my comments would be long or it might take awhile. Or, as Blaise Pascal said, "Je n'ai fait celle-ci plus longue que parce que je n'ai pas eu le loisir de la faire plus courte. (translates approximately to "Sorry for the length of this letter, but I had not the time to make it short.) - which brings me to the point: sometimes simplifying takes far longer, and far more effort, than making things complex.

I decided to get on and post after serendipitously reading two related articles today, both by the science writer John Horgan (www.johnhorgan.org). The first was about common sense on the NY Times Op-ed page (http://www.nytimes.com/2005/08/12/opinion/12horgan.html) The second was on his website, a Discovery article about how whether the brain (my favorite subject) abstracts (my second favorite subject) the world down to the single cells to represent information. (I used to research on Parkinsonian rats similar to what they are doing to Danny in the article.)

The point he makes in both of these articles is that it can be an incredibly complex task in the human brain (and in computing) to distill things down to a single point of leverage where an effective decision can be made. How many years has it taken for us to go from 1's and 0's to 4GL?

What often simplifies are layers of abstraction. In hindsight, these layers may seem simple, but the process of getting there is very complicated.

We mean many things when we say "a simple solution". We live in a world that is very complex, and we have evolved very complex processes to deal with and continue to survive from DNA to the brain to language. These themselves are layers of abstraction. DNA abstracts the structure of proteins, neurons abstract information about the world, and language abstracts what we have in our brains and lets us put on paper or into bits.

My business and life philosophy is to look for areas that are very complicated for people. Then look for ways in which these systems can be abstracted into a simpler, more manageable form. (Even if it is just a checklist for investment criteria). Therein lies the root of value creation in any arena - and the soul of our technology to do this very same thing for the general purpose computing environment and personal storage management.

BTW, what synthesized much of this for me was the writing of the brilliant Peruvian economist, Hernando de Soto. Pick up his book "The Mystery of Capital". My abstraction of his theory: What differentiates the third world from the first world is the first world's foresight and luck in building a consistent system of abstraction of property - so that it can be managed, valued and traded with trust, security and simplicity across a broad range of transactions.

Sorry this post was so long, as I did not have time to make it short.

Cheers, Leonard

[Eric Darst]

I've been reading along on Occam's Paradox and, of course, had all sorts of tangents and redirections I would like to add - maybe I'm the culprit... To stay on track with the laws concept I would suggest that the Law of Entropy is involved - Given time, everything tends toward chaos. I've certainly been in meetings where this is the case. It can certainly be argued that chaos is the ultimate complexity.

But that's not why I'm here... I postulate that complexity is nothing more than propogation of thought with mutation borne out of individual experience. Maybe a meme towards an end (http://en.wikipedia.org/wiki/Meme). From this there are two directions it can go; extinction or a step in the evolutionary path. Distilling simplicity from this primordial goo is knowing what influences have little sensitivity and being damn lucky guessing which way the butterfly wings will flutter. Mutation, though, is the essence of creativity and must be encouraged.

Two years ago I asked to give a business plan pitch to a group of angels and other consultants as part of an incubator finance track. Prior to the meeting I was told my business plan wasn't "thick enough". I sure hope that complexity for complexity's sake is a dead branch on the evolutionary tree...