AI Doomerism

Sep 19, 2026

I recently came across Prathosh A.P.’s lecture on “Pedagogy in the Times of AI”. In the lecture, he shares his concerns about recent advancements in AI (with a premise that we have already achieved an artificial general intelligence (AGI)).

But is this concern justified? Here are some of my thoughts:

• It feels like nobody understands what AGI actually looks like yet. Given that our definition of AI constantly changes with each new breakthrough, I wouldn’t be surprised if it never gets defined.
• I do feel that a lot of the concerns are worrisome, but they are also higly theoretical (i.e., list of potential harms that AGI could bring). But I also feel like justifying AI policy and safety research based on theoretical concerns also misses the point. I.e., are there things that we could do now about our current problems so that we could avoid having future problems?

For example, an AI doomerist might say: “Why bother studying CS? AI is going to program everything anyways in the future.” But think carefully about this statement. It makes a lot of assumptions: “AI is already perfect”, “human programmers are unnecessary in the future”, etc. If these assumptions hold true now, when why do we see growth in senior engineering jobs?