I remember this essay and liked it quite a bit. I do research in this area and, although there's an armchair aspect to the piece, there's also something insight

I remember this essay and liked it quite a bit. I do research in this area and, although there's an armchair aspect to the piece, there's also something insightful in it about the "memory versus computation" paradigm. The way the algorithmic complexity literature discusses modeling I think is missing something -- an intuitive treatment of the tradeoffs involved in different types of data representations -- and this schema I think is on to something.