"Something big is happening" but maybe not in higher ed
Matt Schumer’s “Something big is happening” article went massively viral earlier in February. Schumer draws parallels between early Spring 2020 and today with AI, especially Agentic AI being the COVID analog. He argues that we are at the opening stages of a remaking of the economy and social order. The evidence for this is that huge swaths of people who write code saw that skill outsourced to AI models seemingly overnight. As Schumer and others argue this is not an isolated industry. Since everything we do runs on computer code this is the capacity that unlocks the ability of AI to directly disrupt normal job functions across the economy. The whole piece is worth reading. If you pay close attention to news from the AI labs there will not be anything revelatory here and it sounds like hype to a lot of observers. What this essay does do is crystalize an argument lots of other smart people had made part of before and provide a narrative structure to signal events like the market sell-off of software as service stocks.
It is way beyond the scope of what Nik and I do to weigh in on the accuracy of this essay. We don’t have the kind of insider view of what is going on in the labs or in the labor market to pretend we know what is going on there. What we do see is that this is not happening in higher ed. Last year we recorded a podcast on the slow rate of change in higher education and I want to return to a few of those themes here.
Access to cutting edge platforms is limited and slow moving. Most of the buzz in the last few months has been around agentic coding. We benefit from an enterprise ChatGPT relationship, but even in this ecosystem simply getting access to the technical parts is challenging. I currently cannot get CodeX to communicate with Github inside our ecosystem. I started pursuing permission to use Claude Code for the campus a month ago and we do not have access. This is not a critique of the people in charge of those approvals, they are doing their jobs. This is a recognition that the pace of change in the technology does not match our institutional processes.
I have a few automated tasks set up in ChatGPT. If you really stretch it you could call them research agents. Every Monday I get a set of reports about funding streams, a specific federal government regulation I am following etc. When I tell people this they look at me like I am an alien, 6-12 months behind the cutting edge of agent deployment.
Literacy on this topic is incredibly low. I can count the number of people on one hand who I know are experimenting with these tools in my circle which includes most of the AI early adopters at least at our own University. The thought of trying to onboard faculty to a next generation change is daunting–especially when we don’t have approval to use the tools.
I want to share an anecdote that might be informative for folks outside higher ed. Last summer I met with our Provost and Vice President of Academic Affairs to discuss some of the early noise about job markets and AI. She agreed to let me have an audience with the Chairs and Deans in August. From there I pitched integration of AI tools at the program level and had several programs jump at the chance. One of the early groups was full of smart and diligent educators who met throughout Fall to revise curriculum and propose a new course. Somehow, they launched it as a special topics course this Spring and have requested approval to make it a permanent course which could happen as early as next Fall. Ask anyone in higher education— this is lightning speed for us. From inception to permanent course in one calendar year is as fast as things can possibly change here. How do we map that onto the reality Schumer describes where two major labs release major coding updates within an hour of each other?
Something big may indeed be happening, but it is going to take us way more time than it should to catch up. I worry it will be too late.