In dynamically typed languages, there are still problems with flat Null/Nil/None that are addressed by optional values (the biggest comes when you use multiple operations that can return null but the single null value loses the source of the null; using optional values these often are differentiated easily as being either an "outer" null -- e.g., None -- or an "inner" null -- Some(None).)
In fact, the example in the article of the Ruby K/V store illustrates this problem. If optional values were used instead of an ambiguous nil for both cases, there would be two distinct results:
* Some(nil) -- or something of similar shape -- where there is a value for the key, and it is nil,
* nil where there is no value.
Obviously, the typesafety issue doesn't exist for dynamic languages, but the composability/API quality issues do -- from the numbered issues in the article, I'd say at least #s 2, 3, 4, and 7 apply to dynamic languages in general.