I was impressed by his methodical way of dissecting the design of information systems and how they interacted with the social context. The didactic brilliance was consistent with his (and David Chapman's) famous AAAI paper about the "Pengi" system https://aaai.org/Library/AAAI/1987/aaai87-048.php -- although I didn't realize at first that he was the Agre in "Agre and Chapman".
Yet, somehow, I missed the importance of whatever he said about privacy. Like others I was more interested in what new things we could build if we were all as sharp as him.
(The "Surveillance and Capture" paper mentioned by the Post seems to capture an important distinction between two modes of privacy invasion; even 20 years later I see attempts to discuss privacy concerns founder on a failure to reckon with this distinction. I will now read.)