In our conversation, which has been edited for length and clarity, we spoke about AI becoming as divisive as DEI, the prevalence of scholarly slop, and how Bluesky poisons academic discourse. Evan Goldstein: You’ve described the Substack series as “intentionally provocative.” What — or who — were you trying to provoke? Alexander Kustov: I started using Claude Code and other agentic tools, and I was impressed with the possibilities and functionalities. Yet I still hear a lot of negativity about AI from colleagues who haven’t used those tools. The vast majority of folks in academia still think AI is that chatbot interface from a couple of years ago that hallucinates and cannot really do anything. But that’s not what I saw on my computer. I’d written short items about AI, so I thought to try to see whether Claude could assemble that and create a coherent, substantive post. It did a pretty good job. I just posted it. Then I realized it was a big thing.
Kustov: If we look at the best agentic models right now, they can do most quantitative social-science research tasks better than most professors globally. All of the kind of caveats here are important. What made a lot of people upset about this statement is that they interpret it in a very America-centric way. People were thinking about their colleagues at Oxford and R1 research universities. But I was born in Soviet Russia and went to undergrad in Europe. I have seen a lot of human slop. It should not be a controversial statement to say that agentic tools can do better. The value of qualitative research is going up because that’s something that AI cannot do well — ethnography and actually interviewing people in person, especially in hard-to-reach places. If you are gathering new data that is not yet publicly available, that’s the highest value that you can have as a professor and as a researcher. But I just came from an International Studies Association conference. I saw so much poor research that, if anything, I’m more convinced that my original point is correct. Too many people with Ph.D.s, with tenure, are producing work that is not contributing to human knowledge.
Meanwhile, the Chronicle has decided to interview Dr Kustov, who repeats his claim that AI agents do "qualitative social science research better than most professors globally"
Imagine having tenure and stability, and choosing to shit on his colleagues in a time of great instability for academia.
