What percentage of people know how to build software they want themselves?

I suspect it's a relatively low number - hence developers making comparatively quite decent salaries and software companies having an excellent potential for massive returns.

I don't really see this point as sexist or anti-sorority (although it's true sororities are heavily stereotyped in the US, in my experience) - I think this question would be equally applicable as "find stylish young people in a bar and ask them what they'd want or need to advance socially with a wide audience online."

The idea is that sororities self-select for people who are interested in being social, and that women in sororities are often trendsetters and hence excellent marketing targets.

Lets not ask the doctors what they want; maybe they will get offended because we think they can't build themselves.

Do you really think "work for hire" is a great way to generate startup ideas? This is exactly what pg is pitching there.

The advice also goes strictly against what pg himself is advising in the same article: The place to start looking for ideas is things you need. There must be things you need. [14]

But let's look at the one example that pg gives in the article: Rajat Suri didn't offer to write any software restaurants might need - he learned what they needed, and then struck out on his own.

If anybody had suggested he should just work as free IT for restaurants, that would sound kind of ridiculous, no?

But please, try out his advice, see how it works for you.

I was referring to the part where you said women would get offended; it haves nothing to do with the actual effectiveness of the approach (asking them).