Claus [02.15.07 10:24 AM]

Interesting.
I guess this is economically intuitive in the following sense: If you want a lot of market activity you want to break down your bundle of value to increase the substitutability of your goods. The simpler the goods, the more substitutable they are, so the more trade they can get. With complex bundles no goods can be compared so no market is really created.
Of course the flip side is that the more substitutable the goods the more competition they will find themselves in. That works well for the consumer but less well for the producer, which is why Windows and MS Word, and Mac OS X and the iPod are all these monolithic structures that bundle more and more value. Apple and Microsoft are doing their best to create unsubstitutable goods.

(I'm not an economist, so the above may formally be nonsense but I think it could be rephrased in meaningful economic-ese as something half way correct)

  Dave McClure [02.15.07 11:27 AM]

in addition to modularity, i think people may overlook the basic rqmts for # of participants and txn / event volume in prediction markets.

not that i'm an expert, but it appears the more successful efforts have fundamental advantages in terms of larger audiences and regular / significant events to measure.

example: it's easier to model prediction markets on major league sports teams and presidential candidates than on minor leagues & local elections, or around baseball with a 162 game season & tons of statistical elements, than say a once-every-two-or-four-years Olympic events.

again, this also mirrors open source -- if you don't have lots of participants, and lots of regular interaction events, open source models become a lot less useful.

my .02,

- dave mcclure
http://500hats.typepad.com/

  Peter Semmelhack [02.15.07 11:33 AM]

I think the primary value of modularity is that it introduces the one thing that is largely missing in most markets today - user control. Not unlimited control, but systematic control (think Legos vs. Playdoh). And I think you're going to see a large number of companies form around that opportunity, across the entire spectrum of products - digital and physical. Nike already let's you design your own sneakers, Toyota Scions are designed to be customized by their users, etc. ATMs, Amazon.com, blogs, Tivo, Home Depot, Make magazine - the list is large - have all proved the benefits of putting the power of choice and control in the hands of the end users. Eric von Hippel has been studying the power of this for a while. I think the time has come for it to go mainstream.

  Kind And Thougtful [02.15.07 01:48 PM]

Thank you for the interesting post.

I recently discovered your blog. It has immediately become one of my favorites.

I was wondering if you could please comment on the new free online encyclopedia Citizendium. As you know it is a non-profit endeavour.

Math, science and engineering are critcally important to this country's future. I think that Wikipedia and Citizendium will play an important role in educating our students.

I would very much appreciate any comments you or your readers have.

Thanks!

  Karim Tahawi [02.15.07 04:39 PM]

The modularity versus aggregation point is interesting. Modularity is important but in truth, we all deal with abstraction daily. Abstraction and therefore modularity are relative. What I think is complex may not be to you.

Regarding aggregation, it is true that big problems get big audiences but for a market to succeed, small audiences work very well (and most certainly better than the expert model) over time.

For markets to work, it is more critical to have a problem that is understood, to have a diverse set of people involved, and to have participation incentives. It may be that a handbag market failed because it was neither understood nor reached enough eyeballs.

Keep on eye on us (http://www.my-currency.com) because we are putting modularity and aggregation to the test (once we reach beta).