This excludes any sort of capital good like an accounting package, or his own example of groupware.
You'd never get most of the good product ideas targeted at businesses without industry experience, which sort of contradicts his advice not to work before starting a company.
I don't understand -- how are accounting packages or groupware examples of goods that don't solve a problem for oneself?
If I need to manage my money and taxes, I can build software for that.
If I need to develop some kind of collaboration tool to help me work better with teams at school, or on an open-source project (both not inherently business domains), I can build software for that.
Business tools are merely extensions of these. Unless I missed it, PG never mentions good product ideas targeted at businesses, so I am unclear as to where there is a contradicition. His mantra is to go for a wide user base, which typically requires mainstream business ideas, i.e., business-to-consumer, not business-to-business.
This excludes any sort of capital good like an accounting package, or his own example of groupware.
You'd never get most of the good product ideas targeted at businesses without industry experience, which sort of contradicts his advice not to work before starting a company.