First, they embrace the causation/constitution distinction that has been invoked by advocates of EC. That's a distinction that has been invoked by most advocates of EC, despite the fact that it does ultimately lead them to difficulties.
Second, A&M embrace the "mark of the cognitive" idea, championing (a version of?) Haugeland's MotC.
And, I think this second point is valuable, since my reading of Haugeland is that, at times, he opts for an objectionable form of operationalism, but at other times he opts for some other MotC. A&M describe a non-operationalist MotC. So, I think Haugeland's view is equivocal about what distinguishes the cognitive (or what is intelligent) from the non-cognitive.Third, A&M also suggest a reading of Haugeland's San Jose example wherein functional equivalence comes into play. I had not thought to read the text that way (although I don't think it ultimately changes anything).
Fourth, they accept intentionality as (part of) the mark of the cognitive.
Fifth, they accept a distinction between derived and non-derived representation.