I'm really not a big fan of the technical/nontechnical distinction. Here's a better idea: Be a useful co-founder. In your example, it would appear that our subject is not skilled, so the first thing they need to do is develop an edge that will allow them to be of use to a company.

http://techcrunch.com/2013/02/08/thoughts-from-a-former-busi...

I don't usually come across helpful posts on TC, but this one caught my eye. Basically, if you look like the stereotypical idea guy/gal, you need to put the startup dreams aside for a bit, work at an established company, gain some experience, and, here it comes, develop skills.

Now, I'd argue this guy (the TC guy) is spreading himself a bit too thin, but the idea of being the utility player on the team is pretty good. When combined with your work experience, you might be an attractive candidate. Go a step beyond that by following OP's advice and, you know, actually build something.

It doesn't need to be a home run, but it should speak for itself, and when you have something, anything really, that speaks for itself, you're in pretty good shape.