I am legitimately saddened by how many graduate students, postdocs, and university professors altered their research direction to encompass large language models because of the attention they've been receiving in other areas of life outside of academia. Whether it's to critique them, enhance them, use them, or something else, I view it as a sign that as a research discipline, computer science is unhealthy. We call them "research disciplines" because they're meant to be disciplined about this sort of buffeting.
Diversity of ideas is important. Sticking with a research program long enough to see it through is also important. Changing up what you're working on every time Silicon Valley ejects a new artifact that gets news coverage endangers both of those values.
And holy hell is the monotony boring. Computer science is an interesting, sprawling field with a lot going on! Let's keep it that way!
I'm aware that over the last year or so I've been a critic of #LLM hype so I too am reacting to it. Lately I've been considering changing that up.
Add comment