Thoughts and ideas

After reading the paperDirect Manipulation vs. Interface Agents
I think we already left the era where we were happy we were reaching all the information faster, now it has become so much more important to get the information that exactly we need, information which is free from all the noises it is packed with. I want to give RSS as an example; although it changed the way I surf the web, helped me to reach the information faster, and it was a good thing (which I still believe it is more than most of the things) I have subscribed to just so much channels (thinking how easy to subscribe to a feed) that I cannot keep up with them anymore. I am constantly thinking how I can fill the bridge between this overkill information and my needs. I have tried basic solutions like foldering different kinds of feeds, using smart folders and vice versa but well they keep on coming and they are not going to stop I guess. I believe an application which uses similar approach as in the Bayesian filtering could be a good solution to stop this nonsense. An agent application where I can train it to sort the messages according to current and changeable user needs and where I can watch the visualization of channels that are relevant to me with different cues. I believe applying these (not only Bayesian but also different kinds of sort algorithms) to channels that I am feeding could be an answer to my needs..or maybe not.

Problem with such an agent is I am-just-not-ready-to-give-control-to-it. Rephrasing, I have to trust it with my whole heart that it is going to do the vital selections for me. I think the biggest problem lies in here. So to avoid such a problem, we must pull back our dependencies to a level where it is balanced between our trust and its selection. It is an interesting subject b/c my approach subscribing to shitload of channels becoming obsolete with that kind of approach isn’t it? Well I dunno. I guess I should build such an agent and test it…