After some other reading, I’d like to adjust my previous position: I now think Szostak’s FI might be used to state a hypothesis about GE. This necessarily requires a specific function, knowledge of all the sequences which allow that function, and details of the environment relative to the question (because you need all that to calculate FI). I suggest that Sanford is prevaricating.
“Language or code” is irrelevant, it’s all chemistry.