Pen. Paper. Timer. Exam Time Again.
Turns out, a tier two city in India and Harvard were always heading to the same room.
Somewhere in a tier two, or to be completely honest Tier 3 city, in India, my nieces and nephews are sitting an exam right now. Or preparing for one. Paper, pen, clock on the wall, an invigilator walking slowly between the rows. No devices. No internet. No assistance of any kind.
I used to think their world was the one that needed to catch up.
I was a product of that same system in the nineties. Every important threshold in my life came with a timed paper and a pen in my hand. Getting into Shri Ram College of Delhi University required a score so close to perfect that a single wrong answer could cost you your seat.
At JNU, studying economics, I struggled with statistics. The problem was the theorems. Proofs that stretched from one end of the blackboard to the other, step by careful step, and then disappeared when the class ended. Come exam day, I could not recall them to save my life. I carried a quiet suspicion that the system was testing the wrong thing.
Then came Stanford.
In my policy writing class, a professor of practice taught me something I had never been taught in any exam hall. Policymakers don’t have time to read everything. So you start with a hook. The argument comes first. The evidence follows. The memo has to earn its reader in the first two lines or it doesn’t get read at all. Nobody had ever told me that writing was a design problem.
In my data science classes, I was working with actual data to build models, including one to predict wine preferences. The point was never the wine. The point was learning to ask a question of data, build something that answered it, and explain what you found to someone who needed to act on it. Not recall. Not reproduction. Application.
I looked back at those JNU blackboard theorems and felt certain I had left the inferior model behind.
I started hearing it first on podcasts. Then in faculty forums and academic newsletters. The conversation has reached the highest levels. At Davos earlier this year, Gita Gopinath, one of the world’s most respected economists, now back at Harvard after her tenure at the IMF, mentioned that she is giving oral exams for the first time in her teaching career. The reason was simple: it is the only way to know if a student actually understands the material. Columbia’s Andrea Prencipe, in a recent Andersen Institute discussion, made the same point more bluntly. The take-home exam and the problem set are no longer reliable instruments. The only thing left is a direct, analog conversation.
So classrooms are reverting. Oral exams. Pen and paper in rooms with no devices. The invigilator is back.
Which brings me to the question I can’t stop sitting with.
Were we ever testing intelligence? Or were we always just testing whatever technology couldn’t yet replicate?
Rote memorization looked like rigor until critical thinking became the gold standard. Project-based learning looked like wisdom until AI made projects something a machine could complete before breakfast. Every format that felt like progress turned out to be, at least partly, a response to the tools available at the time.
Now the tools have broken the format we thought was finally the right one. And we are walking back to the room, the clock, and the pen.
My nieces and nephews never left that room. They are still there doing the thing I once looked down on. I got on a plane, went to some of the best institutions in the world, revised my understanding of what learning meant, and arrived, thirty years later, at the same four walls they never left.
I don’t have a verdict on what that means. I’m not sure the question has one.