Scott Allison, founder of Team.ly starts off by looking back at an earlier decision not to learn to code:
I want to reflect on my experience as a non-technical founder and reassess my original decision - almost two years ago - to stick to what I'm good at, and not waste time learning to code.
..and comes to the conclusion that this may not have been the best decision.
Running a business is a repeatable skill; if you've done one well, you should be able to do another well. But a startup isn't a business, yet; it's an idea which may or may not work, and all the traditional business skills you've got don't count. Yes, you have good working knowledge of accounting, an idea of how to bring in sales leads, run a marketing and PR campaign, recruit people and build a great company culture, but that won't help you build a product that customers want. The problem with non-technical founders is that they don't know what they don't know, until they're in way too deep.
Coding is a subset of the startup skill of knowing how to make things. I don't think coding specifically is the most essential (unless you want to build a web startup, of course), but certainly, knowing how to create technology, be it robots, strands of RNA, electronic devices or good old programming, is a fundamentally useful skill as an entrepreneur.
If you read this far, you should follow me on twitter here.