Tuesday, July 28, 2009

Diverse Teams? Really?

The Netflix team competition has been in the press for a while now, and deservedly so. I was somewhat doubtful it could be done. Specifically, I was doubtful there was any solid reason to think it could always be improved by 10%, and I think that is true, although I don't personally know when that level is reached, but we should be able to figure it out ahead of time. I don't see what proof there is for, say, 100% predictability, or even 99% predictability. People do change over time, sometimes.

But, in the NYTimes yesterday, a not-purposefully amusing look at the competition.

Here are some choice snippets:

  • Headline: Netflix Competitors Learn the Power of Teamwork

  • It has also changed conventional wisdom about the best way to build the automated systems.

  • The biggest lesson learned, according to members of the two top teams, was the power of collaboration.

  • The formula for success was to bring together people with complementary skills and combine different methods of problem-solving.
So, what they're saying is that, basically, a diverse team approach is best, and that this is a new finding.


If they were some sort of business sociologists, or organizational communication people, or weren't "statisticians, machine learning experts and computer engineers," then they would have known this years ago.

I am happy to see this combination of sociology, statistics, and computing covered in the mainstream media, but it is rather sad to see our understanding of people and work so misrepresented.
“The contest was almost a race to agglomerate as many teams as possible,” said David Weiss, a Ph.D. candidate in computer science at the University of Pennsylvania and a member of the Ensemble. “The surprise was that the collaborative approach works so well, that trying all the algorithms, coding them up and putting them together far exceeded our expectations.”
Maybe they shouldn't ask Ph.D. candidates in computer science about the psychology and sociology of human work? I don't think it is so much that collaboration worked well, but that a diverse team works well, and that is not new.

(OT, I think my college English advisor was named David Weiss.)

Wednesday, July 22, 2009

Digital Music

I was reading an article covering "casual pirates" over at the NYT and this sentence stood out: "paid-for downloads from services like Apple’s iTunes have fallen short of hopes..."

How strange, I have never heard anything like that before. Usually it's how Apple controls too much of the market, and how well it is doing.

But really, the music industry does not want digital downloads to work. The stakeholders are far too invested in physical plant for that. If digital downloads succeed, that is a failure, and if they fail, that is success.