Pooter Hall
We live in the time where “Woz” (Steve Wozniak) meets “Wiz” (Harry Potter).  In the dim recesses of our past our ancestors would call on the spirit world for relief and or support, given that other options were missing entirely or limited in their effectiveness.  Sometimes these were combined with more mundane things like herbs and changes in lifestyle that would, in themselves, solve problems.  At least for awhile.

The assumption in all this was that there is some agency (God, the devil, the goddess of fertility, and so forth) that is sentient, powerful, and available.  I will avoid getting into theological discussions; the point here is that it is now possible for us to create our own agents.  In doing so, we are in our own ‘wizard’s school’.

Voice recognition has been effective for at least the last five years, so it’s not really a surprise that someone could intone a chant or spell and make something actually happen, such as turn on a light or light a fire.  After all, we have the ‘Clapper’.

The MIT Media Lab is working on the ‘emotionally sensitive computer’, which reads the expressions on the face of the user to find out whether the direction it’s taking is appropriate.  If the interface is hidden (rather than sitting in front of a computer, a camera is embedded in some part of an ordinary room, and can decipher your objective without anyone else hearing or seeing you do anything) then it appears that you control something mysterious and unplumbable, when in fact it’s probably an embedded microcontroller and some software.  A way of understanding how much computing power is needed to do this is to look at the reverse, which is to render faces with varying expressions in real time, such as one sees on the Sony PS2 or the X-Box.  These cost from $300 to $400.  If it can be rendered, the rendering can be compared to a real world image, and an expression can be associated with a command.

Much of this requires the maintenance of a vast amount of context.  A classic scenario is: you pull into the driveway, get out of the car, the computer recognizes that you’re home and automatically unlocks and opens the door.  What it does next has a lot to do with where you’ve been.  If you’ve been working, maybe you want quiet music and a stiff drink.  If you’re home from the theater with the spouse, all you want to do is go straight to bed.  If you have a bunch of boards and hammer, it’s time for ‘home improvement’.  Maybe you have a bag of groceries under each arm.

One way the computer knows where you’ve been is to download history from the GPS module you installed in the car.  Up till now, these have been used for tracking commercial vehicles or your teenager; now you are spying on yourself.  What the computer knows is that if you’ve spent all day at the mall, you’re broke and exhausted.  If it’s Friday, there’d better be cold beer in the fridge.

The usefulness of this in computational terms is scheduling and resource allocation.  You might have a weakness for chocolate, the computer translates this into a $2 a day habit.  It also figures out you drive twenty miles to work and back, and this will require significant service work three months from now.  If this and the other factors are plotted on a graph, they show you slowly dying in a sea of debt.  All you have to do is earn or conserve another $1000 per month (after taxes) and you will retire a millionaire.

What you wanted was something that would clean house and make dinner, what you got was a system for reminding you to quit spending money.  There are already lots of people around doing that; why do you need any more from a machine?