Jason Calacanis And The No Free Lunch Theorem

Mahalo Alpha is here to help you search.

Jason Calacanis has recently launched Mahalo, which is a new search vehicle. It is based on human evaluation of websites and has been the subject of some criticism. For example as Michael Gray (Graywolf) recounts four noted SEO experts gave Jason a difficult time in a recent podcast. The others were Brian of Scoreboard Media, Allen Stern of Center Networks, and Andy Beard.

The SEO experts seem to have greater faith in computer-based search engines. They’re all based on algorithms rather than human beings. For example Google boasts about the strength of its PageRank algorithm. Even ASK seems to accept the power of this argument and has been featuring its algorithm in recent publicity. That would certainly seem to be the majority view at the moment.

Jason Calacanis might find some scientific comfort among those attending the Seventh Metaheuristics International Conference taking place in Montreal, June 25-29, 2007. Optimizers will be meeting to discuss the best methodologies for optimization. Apparently one of the fundamentals is the No Free Lunch Theorem. This was first stated by David H. Wolpert and William G. Macready in 1995. Wikipedia has the following explanation:

Some computational problems are solved by searching for good solutions in a space of candidate solutions. A description of how to repeatedly select candidate solutions for evaluation is called a search algorithm. On a particular problem, different search algorithms may obtain different results, but over all problems, they are indistinguishable. It follows that if an algorithm achieves superior results on some problems, it must pay with inferiority on other problems. In this sense there is no free lunch for search algorithms.

That may be something that Google and ASK would prefer is left hidden in academia. However search is more about marketing than technology. Perhaps Jason Calacanis can use the No Free Lunch Theorem in his efforts to further publicize Mahalo.

2 thoughts on “Jason Calacanis And The No Free Lunch Theorem”

  1. I won't take umbrage with the views of Michael Grey and others, I happen to agree with a lot of what they have espoused.

    But it should be pointed out that, in the black and white hypothesis they appear to be testing, they are stakeholders with a lot to lose should humans win the battle over algorithms 🙂

    How would a SEO consultant optimize placement controlled by a set of human guides? Hire lobbyists?

    And does that make Mahalo/ChaCha the Neo to Googles Mr. Smith?

  2. Thanks for those interesting musings, Jeff. I fear that if DMOZ couldn’t make it then Mahalo has even less of a chance, so we likely won’t have to face the dilemmas you describe.

Comments are closed.