What's Ubicomp good for?
Dave thinks the problem is that design is not very iterative.
Sometimes you can't know things in advance. It's easier to just go toprototype than try to design everything out.
Idea for technology policy: how can we grow the economy with ashrinking population?
Jeff: how can you design for structural problems in how people live?
Ben: you need to look holistically at what's people's needs are. [I didn't understand the distinction]
Do we have to throw our hands up about AI completeness?
Jeff: Machines can determine certain features of human interaction because they are not biased by emotions and mental filters. Also machines can focus on multiple things at once.
Matt: Even if we can't solve some of these problems, trying to solve them can reveal interesting knowledge about human nature.
Jeff: We don't have to model what people actually do. We can design for increasing the diversity of people's experience by changing what they do. Also paradox of choice. Taking away people's choices can be good. Like the Soup Nazi.
Ben: I did body storming. You can better recognize things if you embody the system. You can reverse engineer. Most useful in design situations where you are very unfamiliar with the situation.
Sushmita: why would the mug be useful? If you remember to sync, you won't forget to pick up your documents.
Sandbox toolkit environments
Do these systems need to be profitable?
Profit doesn't necessarily mean money.
There are also network effects. It has to be adopted by everyone in order to be ubiquitous.
Adam: doesn't need to be adopted by everyone.
OTOH, giving tools to everyone doesn't mean you get tons of value. You can get ton of junk.
Tools are there not to be a designer. Tools are there to share information.
There's no evidence that any of the toys are good.
JZ sez: "What kind of world do we want to create?"