He presents good arguments about why both types of languages can't currently break into the mainstream, but it unfortunately seems like these options form a partition of the possibilities (i.e. a functional language must fall into one of the two). If this assumption is correct, then by his analysis no functional language can break into the mainstream.
I'd love to see a discussion over whether or not that assumption is valid, but for now let's assume it's true. I don't think the languages are the problem. Instead, I think it's the general momentum and consensus of the community. Without books like Learn You a Haskell (http://learnyouahaskell.com/), Haskell would be an order of magnitude more difficult to begin learning. I'm still trying to figure it out, and there are a lot of examples of topics where a similarly good tutorial would ease my learning. Eventually, as the language gains popularity, more and more such work will exist. Furthermore, if popular companies demonstrates that it's a great choice, more people will want to pick it up. If all this stuff happens, our intuitions will grow and adapt to make such a language a natural choice. In summary, the future of functional languages depends more on our understanding and promoting them as a community, as opposed to the languages themselves.