Rise of the Mentat Search Brain
Posted on May 17, 2008
Filed Under The Startup
The idea is that currently search engines are powered by giant AI brains - computers programmed to select the best of the links -based on numbers of inputs then covertly shaped and rated by their in-house teams. This is good to a certain degree and G is in the everlasting process of mastering this method
However it be not the only market. If we reference beloved F. Herbert, then we can discuss the Mentat model - Mentats being of course humans pumped up with various aroma in order train their brain to replicate the processes of the computers. This of course due to the nuclear development of AI which made computers hyper-intelligent but without the moral framework (read soul) in order to exist in harmony with humans who they viewed as inferior.
Coming back to this present reality; we can look at the natural evolution of search. Del.ici.ous has the right idea but too many dots (sic) in their name to be mainstream. (read this more ways than one) But their model being that humans select what is their favorite sites and mark them - the more humans that mark a site - the higher it shows up in the delicious results for that keyword search.
Take a search of my current home of ‘Thailand’ for example
Google computer results are the usual suspect of a Gmaps at the top of where is Thailand- along with some nice photos - then the official tourism site, wikipedia, cia factbooks, Bbc
All good and sterile results considering that ‘Thailand’ is a highly generic term and the user could be searching for anything.
But compare to the user generated results of delicious. There at the top we have a backpackers guide to Thailand - then we have a hotel reservation service which has been spammed in there through exploitation of human bookmarking.
After this we have links to Thai recipes, some random record label that happened to be tagged as Thailand, some photos and a guidebook.
All in all comparing the first page results of G to Delicious we see a few things
- G is assuming that by searching for ‘Thailand’ we are searching for generic factual information regarding this term. As a result this is what we get
- Delicious is returning results of anything that humans tagged as ‘Thailand’ with minimal sorting and AI via other factors. Since the majority of Delicious users are younger and mobile - thats how we get a backpackers guide on the first result as opposed to maps and wiki facts.
The point being that with D - the results are shaped by the users input - so the results are only as good as the other users within the system.
This leaves us to the major cross-roads between D and G. Overly objective versus overly subjective.
So where is the solution?
Well the solution is literally the 40 billion dollar question for the person or team that can come up with the right answer. Both models are perfect, yet flawed. The winner will be that who is able to take the best of both of these worlds and with the correct twist create the ultimate search and organization experience.
Without saying too much at this point, I have founded and been working vigorously for the past 6 months on a concept that will present the solution to the described problem. Within the next month we will be beginning a limited beta testing in which to perfect our system. I will need the best and the brightest of my readers to help with the testing.
Some traits that we will need
- Those involved on the front lines of the industry - pushing the web towards its natural (r)evolution. Those that play with the new toys as they are put forth, embracing the best and discarding the rest. Those who want to see a better web and have the experience and knowledge to give constructive criticism in order for us to achieve this way.
- People that are do internet marketing for a living and frequently do the directory, article, link-building thing. Even those of you are into OMG!!! Spam. Why am I asking? Because we need help to perfect the system to only allow quality links in our system. And who better to ask than those who work on the weaknesses everyday. (if you help I may even throw in some permanent PR6 backlinks
) - Common non-savvy users - The type of person that turns on their computer 3 times a week and reads yahoo news and checks the auctions on ebay. We need to test out usability and cleanness of interface
- Anybody who would have a little bit of time to use and evaluate the system to make constructive criticism. We will need 50-100 users for our first run and would hope that each would devote at least 20 minutes to evaluate a first impression and give us basic feedback. (of course if it is compelling you will spend days)
We can say that we will be ready early-mid June. If you can offer your help please send an email to me at
neyma (at) webgenome.com
and we will be in touch.
Comments
Leave a Reply
