F-Bombs and Fair Points: On Ronny Chieng's Harvard Speech
There was another controversial graduation speech, but this one was a little different.
Ronny Chieng, comedian and Daily Show host, took the stage at Harvard's Class Day last week and delivered a profanity-laced tirade against artificial intelligence that brought the crowd to its feet. The headline was the f-bombs. The actual argument deserved more attention than it got.
Chieng wasn't just playing to the room. Underneath the performance was a coherent position: that AI used as a substitute for thinking creates cognitive debt, that the process of learning and creating is the point, not the product, and that handing that process off to a machine is a form of self-robbery. He even carved out exceptions for medicine and physics. That's not crowd-pleasing red meat. That's a real argument.
He's right about cognitive surrender. People who use AI to avoid thinking are genuinely harming themselves, and the research backs him up. But here's where I part ways.
Chieng is a Harvard-educated comedian in an industry that AI cannot replicate. He has never needed a tool to compensate for a skill gap because his career depends on the irreplaceable thing: being Ronny Chieng. That's a privileged position from which to tell everyone else to destroy the machines.
For a lot of people, AI fills genuine gaps. I have tried to learn coding. Repeatedly. It didn't take. AI lets me build things I could not otherwise build. That's not cognitive surrender, it's access. The person who can't afford a lawyer but can use AI to understand their lease, the first-generation student who uses it to learn academic writing conventions, the small business owner who can't hire a copywriter, these aren't victims of cognitive debt. They're people with a tool they wouldn't otherwise have.
That distinction matters because the access gap is not theoretical. Low-income Americans routinely face civil legal problems without enough help, a reality documented by the Legal Services Corporation's Justice Gap report. Small businesses are already using AI because they do not always have the staff, money, or time that larger firms do, which is exactly the kind of practical constraint described in the NFIB's 2025 technology survey. And in coding, the difference between outsourcing and scaffolding is already becoming a research question, with recent work on epistemic debt in novice programming showing both the danger and the possible middle path.
The problem Chieng is identifying is real. The solution, destroy it, is the luxury of someone who doesn't need it.