October 3, 2005

let her infer what i imply

a handy mnemonic so that you don't go around accidentally inferring things that you meant to imply.

lol.

(/grammarpolice)

"we tight, right?" - in reference to parliamentary party structure, courtesy of dr. paul sum. (yeah, we tight.)

ahem.

I didn't get much done this weekend. Spent all of saturday sick as hell. Got so tired of the vikings getting their arses kicked that i actually fell asleep on the couch. There's a big pile of laundry on my bedroom floor stinking up my room, and all kinds of plastic bottles and paper garbage in my car. Furthermore, my backpack is disorganized, I don't have half the contact information I need to investigate my DS story, and I lost my list of passwords to the computers in the studio -- again. Meanwhile, I have a test in interpersonal communications tomorrow and another in comparative politics on wednesday. about two hundred phone calls to make, and I also need to pick up a couple of shifts this week at pops to alleviate mounting financial tension. I'd be ok, except I am nearly out of coffee.

The reason for the topic line is this article from DrunkenBlog. I'll let you read it yourself, but the gist is that the Next Big Thing will be relative to the human capacity for mindless inference. One does not consider that the law of gravity will pull the water down on top of you when you turn on the shower in the morning; you just turn the knob and get wet. Having to think about what would likely happen whenever you did anything would make life impossible; maybe that's why we get so frustrated and feel like our lives have been irrevocably changed when our cars don't start because it's -300 degrees outside and we still have to go to class. We expect that certain things around us will just work, and work in an intuitive way.

Imagine if your computer could make as many inferences about you as you make about the world. Computers are minimally capable of this - URL autocompletion comes to mind, as does T9 text input. But they are fundamentally weak, reliant on recent inputs/relatively simple mathematical algorithms, and not very intuitive; computers themselves are not intuitive. What, sayeth drunkenbatman, if this were not the case? It presents some very difficult and expensive technical problems, some of which are addressed in the article. I won't go into much detail here, just because I'm pressed for time and I don't want to try and say something that DB says much better.

Swami Dusty will prognosticate, though: how awesome would it be if you could travel from one terminal to another, from home to car to school to work and back, and have all your data follow you securely? What if computers were able to tell who you were without you having to type a million usernames and passwords? Better yet, what if computers could tell what you were doing (writing a paper, blogging, etc.) and could offer tailored services without your having to ask for them?

I think it's possible. I think that data must become much more portable and accessible before anything revolutionary happens in the way we process it; some of the biggest barriers to such a scenario are the hardware and software incompatibility that plagues the desktop PC/mobile/wireless world. If by some freak occurrence (or more likely, series of freak occurrences), interfaces, data formats, storage, and transmission protocols become standardized, there would be a market for some of the pure research needed to develop truly adaptive devices.

Then again, maybe not.

No comments: