I know he tried Haskell. I agree with him on a lot of points he made there. But:

1. He investigated it long after he had already been successful using something more mainstream. If you are starting something up then I think it would be good advice to use what you know and get on with it. If that's Haskell then good for you. If its not then don't waste your time. Get on with the product.

2. Even then he didn't finish rebuilding his game because he was drawn back to his day to day work. The guy is just product focused.