I looked around at Elm a little bit recently. I use React/Redux daily at work and have been learning Haskell on the side. When I took another pass at Elm it was extremely easy to pick up. I was able to skim through the docs super fast and start writing real code in about 15 minutes.
I noticed myself typing Haskell code a bunch of the time and wondering why it didn't work though. For me, the hardest part in Elm, and Haskell as well, is figuring out how to structure data that is composition in nature or that involve Union types that are records. I can never quite seem to get the dereferencing right.
My interpretation of that sentence was that it's easy for developers familiar with other languages to pick up Elm, but less easy for people with little or no experience what-so-ever.
I think he was contrasting with languages like Haskell and Scala, where even people very experienced with other languages have a hard time picking them up.
You hire experienced developers. Then you teach them Elm.
I don't think he's saying that Elm is actually hard to learn (like Scala or Haskell).
Not really seeing this as a strong point. It feels like the same mindset that drove me away from Scala (and many others [0]!). Sure, it might be good for the company in short term, but what happens when you run out of experienced Elm developers? At least Elm fairs fairly better with having usable samples on their own (clean) site.
[0] https://groups.google.com/forum/#!topic/scala-internals/r2Gn...