It's August of 2026 and the Mbed platform is dead.
You're forgiven if you don't know what Mbed is. It can be summarized this way: Mbed was a cloud service to help engineers make stuff on little chips. Basically, think of it like Google Docs for microchips. It launched in 2009, around the time that I was starting to teach at TMU and I considered it as a tool for teaching my engineering students how to program their chips.
But after kicking the tires on Mbed, I moved on to more practical tools that could live on university and student computers. To me, "on device" or local applications were more viable for teaching than cloud-based ones. I didn't see the benefit of programming in the cloud and worried that this particular cloud-based service would one day disappear.
And disappear it did.
The fall of Mbed tells us that there are limits to cloud-based systems. Not all services need to be online. We don't need to pay to outsource everything. And, just like I don't think that we should put all of our eggs in the cloud basket, I think that relying on Sam Altman's magic online eight ball is a mistake.
What does Mbed teach us about AI?
HigherEd managers have been telling faculty and staff to get on board the AI train for the past couple of years. Put AI in the classroom and make students "AI literate", they said. They told us it was safe, that we could trust the performative"guardrails". Behind the scenes, because most were experiencing generative AI like a traditional Google search (through a web browser), they were also probably assuming that AI would be infinitely cheap (or free!) and that they could cut HR costs as a result. This way of thinking was found outside of HigherEd and resulted in the "tokenmaxxing" that defined early 2026. By mid 2026, companies realized that tokenmaxxing was hitting them where it hurts: the bottom line. This is the reason why universities don't actually provide employees or students with direct access to frontier models -- it would cost them too much.
So, what does this mean for AI in higher education? Well, for one, the shortcomings of current AI systems should give us the space to rethink the breathless enthusiasm of AI boosters. We don't need AI in everything. It doesn't belong in all of our university services and classrooms. Discussion about AI needs to provide space to explore and implement the "AI doesn't belong here" and "AI is not for me" perspectives if, for no other reason than the budget's bottom line.
And for the limited spaces where a critically-reflected AI approach might make sense? Well, it means a neutered approach, even for the biggest AI boosters, because they simply can't find the vast sums of money that would be needed to grant students or staff unfettered access to AI.
This is a two part blog post in which I argue that we need to curb our AI enthusiasm. This first post tries to draw lessons from an obscure online engineering tool. The second attempts to do the same thing with 1980s-era cook books and microwave ovens.
James Andrew Smith is a Professional Engineer and Associate Professor in the Electrical Engineering and Computer Science Department of York University’s Lassonde School, with degrees in Electrical and Mechanical Engineering from the University of Alberta and McGill University. Previously a program director in biomedical engineering, his research background spans robotics, locomotion, human birth, music and engineering education. While on sabbatical in 2018-19 with his wife and kids he lived in Strasbourg, France and he taught at the INSA Strasbourg and Hochschule Karlsruhe and wrote about his personal and professional perspectives. James is a proponent of using social media to advocate for justice, equity, diversity and inclusion as well as evidence-based applications of research in the public sphere. You can find him on Twitter. You can find him on BlueSky. Originally from Québec City, he now lives in Toronto, Canada.