Thinking Aloud About Lectures
Two things used to happen in a classroom.
Knowledge used to be disseminated (by a human), and knowledge used to be acquired (by many humans). There would usually be the one prof whose job it was to disseminate knowledge, and there were usually many people (students) who would acquire knowledge. These two things (one prof, and many students) would constitute a classroom.
And so you may well be justified in thinking that a classroom was a one-to-many function.
Here’s my question du jour: in the age of AI, should we stick with this arrangement or change it. If you say stick with it, why? If you say change it, why?
But hang on for a second: let’s go back to the classroom and think about it some more.
Upon reflection, you realize that knowledge acquisition doesn’t happen only by listening passively to the professor. Learning also happens when a student engages with the teacher on a one-to-one basis, for example. But this can very quickly get tiresome for all the other students in class, as all of us have experienced at least once in our lives (“You remember that guy who just wouldn’t let go? Gawd!”)
But the list isn’t complete yet. In a classroom setting, learning also happens when students engage with each other in debate. A lively conversation among many or all of the students, moderated by the professor, is also a great way to learn. In this case, learning happens on a many-to-many basis.
Of course, not every many-to-many conversation has to be a debate. The examples given above are very far from being comprehensive. They are only meant to be indicative. But the point is that for the most part, we are left with three different ways in which a classroom setting can lend itself to learning:
One-to-many, where the prof speaks, and everybody listens.
One-to-one, where the prof is in conversation with one specific student
Many-to-many, where the entire class is engaged in spirited conversation with each other.
Should I Have Been Out of a Job by Now?
Let us assume, for the sake of argument, that AI becomes better than the professor in terms of at least possessing knowledge. I would argue that we are already well past that stage (and in my own specific case, I know that I lost the “who knows more” battle a long time ago).
Let us also assume, again for the sake of argument, that the AI becomes better than the professor at disseminating knowledge. This is not such a clear-cut argument, and for many reasons. My biggest advantage is simply the fact that I am human. But my innate human-ness (not humaneness, note – that is a different word with a different meaning) is also an obvious disadvantage. Being human doesn’t allow me to create infinite and infinitely different ways of providing the same explanation with limitless patience. But if not now, then in the very near future, it is perfectly possible to imagine that the one-to-many function we spoke about being entirely taken over by AI. And I can also imagine parts of one-to-one tutoring being taken over by AI.
Let me concretize that for you. Imagine a classroom, or a campus, where students assemble together at the same time. But they don’t learn from the same human prof on a one-to-many basis. They learn, instead, from their own personalized AI tutor. Such a classroom should have better learning outcomes than a large class taught by a human professor. If not now, then in the very near future.
Why? Because lectures and lecturers were optimum for an age in which explanations were hard to come by. But in the age of AI, explanations aren’t scarce. They are, in fact, infinite. And relative to infinite, Gen-AI provided explanations, it is human-to-human shared experiential learning that becomes scarce.
The Classroom is Dead, Long Live The Classroom
Does that mean that there is no room left for classrooms, and campuses?
Absolutely not. Not only is there room left for classes/campuses in such a world, but these become more important – much more so.
Why? Because there are two other kinds of learning that can take place in a classroom. Your AI tutor(s) can absolutely do a fantastic job of teaching you, but wouldn’t it be great to be able to talk about what you’ve learnt with folks on the same campus? If they’re learning the same thing as you, you can have a nuanced conversation about whatever it is that you’re learning. If they’re learning something else (no matter how closely or tangentially related), even better!
It would be great not just because of the joy to be derived from learning for its own sake, although that is its own reward many times over. But even more so, and this needs to be explicitly said in our times, there is immense benefit from meeting other members of our own species, preferably on a daily basis. Not only can AI not be a substitute for this in a tautological sense, rather the value of human companionship goes up in the age of AI. Campuses that allow for in person, peer based interactions over a sustained period of time become an incredibly valuable service. Priceless, even.
But in an academic sense, those one-on-one reflective conversations, with either a professor or with a student (or both), can help you better remember and better understand what you’ve just learnt.
Also, one quick point that matters: remember, here at EFE, we celebrate middle of the road approaches. I’m not advocating the death of all lectures and the firing of all lecturers. There is value in having some good ol’ fashioned lectures even now! But yes, I am very much suggesting that we should, at the margin, be experimenting much, much more.
The age of AI isn’t the death of the campus and the classroom. Quite the contrary, in fact. But yes, I argue that the age of AI should certainly imply the death of the one-to-many lecture. That was a solution that worked for its time, but it worked because we didn’t have a better technological solution to Bloom’s 2-Sigma problem. If we now have a better technology to help that problem, we shouldn’t let inertia stop us from vigorous, multi-faceted experimentation.
Here’s an excerpt from a lovely post by Ethan Mollick:
This returns us to the world of organizations. While individuals rapidly adopt AI, companies still struggle with the Garbage Can problem, spending months mapping their chaotic processes before deploying any AI system. But what if that’s backwards?
The Bitter Lesson suggests we might soon ignore how companies produce outputs and focus only on the outputs themselves. Define what a good sales report or customer interaction looks like, then train AI to produce it. The AI will find its own paths through the organizational chaos; paths that might be more efficient, if more opaque, than the semi-official routes humans evolved. In a world where the Bitter Lesson holds, the despair of the CEO with his head on the table is misplaced. Instead of untangling every broken process, he just needs to define success and let AI navigate the mess. In fact, Bitter Lesson might actually be sweet: all those undocumented workflows and informal networks that pervade organizations might not matter. What matters is knowing good output when you see it.
If this is true, the Garbage Can remains, but we no longer need to sort through it while competitive advantage itself gets redefined. The effort companies spent refining processes, building institutional knowledge, and creating competitive moats through operational excellence might matter less than they think. If AI agents can train on outputs alone, any organization that can define quality and provide enough examples might achieve similar results, whether they understand their own processes or not.
Ethan Mollick’s garbage can applies as much to the campus as it does to the company. We shouldn’t ask how to optimize existing learning workflows.
We should ask, instead, how to reimagine these workflows from scratch.
Trains aren’t an abundance technology, says Eli Dourado:
One thing I got a bit of crap for in the hallways of the Abundance conference is my not infrequent mockery of trains on Twitter.
I’m sorry, trains are not an abundance technology.
I think many people in the abundance scene like trains because:
1. America’s inability to build…— Eli Dourado (@elidourado) September 7, 2025
Neither are classroom lectures in their current form.
You may or may not agree (and I hope you don’t, because disagreements are awesome). But whatever your answer, please do tell me the why of it, and let’s talk!