the other day i visited Amazon.com. i have not visited the site for some time so i was just browsing through it. I quickly noticed that amazon actually make good recommendation on some products based on your previous browsing and shopping experience. the recommendation are quite useful to me and i was able to follow the suggested links and get what i wanted.
this brings out a popular question: can a web portal be more intelligent than just serving contents to users? my answer is yes and i am a firm believer of that. i remembered many years ago when the .com hypes were high, i went to a personalization summit in san francisco. i did not remember much of what happened in that conference (other than over 50% companies presented in the conference were out of biz now), but i do remembered a presentation by a speaker who he talked about finding soul mate - imaging you were asked to pick 6 movies of your interest. if you can find someone who picks exactly the same movies as you did (or something similar), the chance is very good that whatever the movie he or she picks, you may also be interested in viewing. and vice versa. this is a quite simple example of intelligence in predicting what users want and present them with the results.
this is something we at ziplecture wants to do. once we grow our user base, there will be a rich set of information we can use to build a prediction model that will link users of the same interests, find people who might help you to solve your question, provide recommended reads and training lectures to users, etc.
