Also, I wasn't arguing in my post that this was a solution to 10%; just that in my world and sector there's lots of jobs despite this big unemployment number. And we need more technical literacy (not just programming skills as some commenters have misinterpreted) throughout our education system.
For example, every business class I took as recently as 2003 involved products in boxes and shipping.
It's pretty fucking hard to automate programming. Heck, we can't even solve voice and character recognition.
I think it will be still be a long time before robots replace humans, even at Wal-Mart-like jobs. There are a few places where it's appropriate (self checkout, for instance), but there are things that humans do without batting an eyes are very very hard or expensive for robots (spoken communications, visual and auditory classification, spatial navigation and obstacle negotiation).
Humans aren't really that expensive honestly. Considering that the rent on a few servers from EC2 can equal the hourly wage of a Wal-Mart employee, it's going to be a while before these equations start working out.
You're absolutely right about people not being very expensive though. Labor costs actually make manual manufacturing in China less expensive than in fully roboticized factories in Japan or the USA for many things.
Technology is very good at creating value, but not very good at creating jobs. All 72 tech companies in the S&P 500 combined employ about 2.3m people [1], whereas just Wal-mart employes about 2.1m people. Those numbers are worldwide employment, but for reference, the US had 15m unemployed in November.
Even the breakout, unimaginable success that is Google only employs about 28,000 people. And I'm sure that for every hire Google has made, they've displaced many more people that used to depend on print advertising. (To be clear, I'm not saying that technology is net bad for the economy. Technology creates a lot of value for the individuals that work and invest in it, and for the economy as a whole. However, technology is not a massive creator of jobs.)
So unless startups become a multiple of the size of the entire S&P 500 technology sector, they're not going to make a noticeable dent on unemployment.
[1] I compiled a spreadsheet (https://spreadsheets.google.com/ccc?key=0AqpDDf0Sr4EXdDJMZEt...) based on information available on Wikipedia[2] and through Google searches
[2] http://en.wikipedia.org/wiki/List_of_S%26P_500_companies and individual company pages