Only Professors
Jay Caspian Kang's piece in The New Yorker features Hollis Robbins, a professor and former dean who has been thinking through what AI does to the university from the inside. She has a provocation she's put to her faculty colleagues: write a memo answering the question "What specific knowledge do I possess that AGI does not?" Those who can't produce a compelling answer, she argues, have no defensible place in the institution.
Robbins' argument is that universities have spent decades standardizing themselves into interchangeability that leads to commoditization. The Common Application, transfer credits, learning outcomes written to be equivalent across institutions, all of it designed so that a student at one school learns roughly what a student at another learns, which means the parts are interchangeable, which means, as she puts it, that the faculty have been told they are not special. The vulnerability to AI follows directly: if it doesn't matter who teaches the class, it eventually won't matter if a human teaches it at all.
Her prescription is to lean into the opposite, become specialists at the edges of knowledge, bespoke, unteachable by a model because the model hasn't gotten there yet. Her own example is the African American sonnet tradition, on which she is one of perhaps three people on earth who know as much as she does.
To be at the edge of knowledge and do something there that AI cannot replicate, you first have to have walked through the center. The African American sonnet tradition is an edge because there is a tradition to be at the edge of, and to understand why Marcus Christian matters, you need to understand what he was responding to, which requires knowing enough of the canon that the deviation from it means something. The student who arrives with a different baseline because ChatGPT has already transferred the standard content is arriving without the deep familiarity that makes edge knowledge legible. The logic is that you cannot rush to the frontier if you haven't crossed the territory.
Robbins is describing a university that will be leaner and stranger and more differentiated, and I think she is probably right that this is where the surviving institutions end up. While reading this piece, I wondered who gets to attend that university in the model she is proposing. The model being: find the professor whose specific knowledge you want, pay for the credential of that person rather than the institution they are affiliated to, is the model that works if you already have the cultural capital to know which edges are worth standing at, and the network to find who stands there. Kang notes that this conversation happened over dinner in Austin with twenty-five-year-old billionaires. That is the profile of the person this model was designed for, whether or not that was the intention. This implies levels of privilege and access that a minority of students will have atleast given the current state of K-12 eduction.
Robbins' argument is that universities have spent decades standardizing themselves into interchangeability that leads to commoditization. The Common Application, transfer credits, learning outcomes written to be equivalent across institutions, all of it designed so that a student at one school learns roughly what a student at another learns, which means the parts are interchangeable, which means, as she puts it, that the faculty have been told they are not special. The vulnerability to AI follows directly: if it doesn't matter who teaches the class, it eventually won't matter if a human teaches it at all.
Her prescription is to lean into the opposite, become specialists at the edges of knowledge, bespoke, unteachable by a model because the model hasn't gotten there yet. Her own example is the African American sonnet tradition, on which she is one of perhaps three people on earth who know as much as she does.
To be at the edge of knowledge and do something there that AI cannot replicate, you first have to have walked through the center. The African American sonnet tradition is an edge because there is a tradition to be at the edge of, and to understand why Marcus Christian matters, you need to understand what he was responding to, which requires knowing enough of the canon that the deviation from it means something. The student who arrives with a different baseline because ChatGPT has already transferred the standard content is arriving without the deep familiarity that makes edge knowledge legible. The logic is that you cannot rush to the frontier if you haven't crossed the territory.
Robbins is describing a university that will be leaner and stranger and more differentiated, and I think she is probably right that this is where the surviving institutions end up. While reading this piece, I wondered who gets to attend that university in the model she is proposing. The model being: find the professor whose specific knowledge you want, pay for the credential of that person rather than the institution they are affiliated to, is the model that works if you already have the cultural capital to know which edges are worth standing at, and the network to find who stands there. Kang notes that this conversation happened over dinner in Austin with twenty-five-year-old billionaires. That is the profile of the person this model was designed for, whether or not that was the intention. This implies levels of privilege and access that a minority of students will have atleast given the current state of K-12 eduction.
Filed under:
Education