Nai Chng makes the point that at the early stages of a startup, if you can't help with the coding you're... less than useful. While I certainly agree that learning to code should be as essential for non-technical founders as learning business essentials is for technical founders, I think you can take it a bit too far:
... at the very early stages of a startup, what are you going to do if you can't code? There are only that much research/admin/strategy/cust. dev/misc work a non-technical person can do. (...) Right at the beginning, you got nothing but a gut feeling about something and what needs doing is the building of a prototype to validate those feelings.
That is only a problem if you see a startup as a series of things to build. You shouldn't build a prototype to validate your gut feeling - you should build a prototype when and only when it is necessary to test the next hypothesis on your startup hypotheses. Most of the time, the way to get that next critical answer won't involve much, if any, coding.
So, I don't think it's true that you need to turn into a hacker. At some points, there will be more programming to do, and at others there will be less programming to do. All founders should be multi-skiled and willing to get their hands dirty doing anything the business requires, but they don't need to all be hackers. There are plenty of examples of successful businesses started by a combination of both hackers and non-hackers.
If you read this far, you should follow me on twitter here.