I’m currently reading Katabasis by R.F. Kuang, a fantastic author who seems to have complex thoughts and feelings about institutions of higher learning. The last book of hers I read, and also very much enjoyed and routinely recommend to people, is entitled Babel, or the Necessity of Violence. Babel takes place in a version of Oxford University where the academics practice magic by inscribing translations of the same word in different languages on an object, and the very slight semantic differences between the words translate to magical effects — linguists thus possessing the greatest powers in the universe. Some of that world spirit carries over to Katabasis, where practitioners of “magick” in universities create pentagrams in chalk and imbue them with power through mathematical, logical, and linguistic conundrums and complexities.
Academia has long been a part of my personal story as well, though I am not so talented a writer as to turn it into fiction the way Kuang does. My dad was an English literature professor until he retired, and when I pursued a Ph.D. in computer science it was with the intention of becoming a computer science professor, a road I ended up not taking, choosing law school and public policy instead. Then 15 years ago I went on the so-called “meat market” to try to become a law professor; that didn’t work out, but I’ve forged a much better (for me) technology and policy executive career since then, sticking closer to the levers of power that influence real-world impact and outcomes.
I grew up thinking that all of the smartest people go to universities to teach and to do research, and while I now know very well that is not true, some of the smartest in fact do choose the Ivory Tower. I know and deeply respect many academics, including in law and computer science. And while I bear a tattoo of the Tower of Babel on my left forearm (coincidentally gotten the same year as Kuang’s Babel), in part to represent my own complex thoughts and feelings about academia, I’m now joining the ranks of the robes, at least in a part time capacity: effective July 1, I will be an affiliate assistant professor of computer science at the University of Washington, Tacoma.
These thoughts come to my mind more often than they used to as I work more and more in AI, where the connections between industry and academia are strong. Many of the people in the field I most respect, like Yann LeCun, seem to have it all—university positions and ample publications, and also a history of shaping products and policies in the real world. Being at the intellectual edge has an inherent draw to it, for me and many more. Kuang in Katabasis in several places illustrates that compellingly, e.g. page 265:
“It is as Heisenberg said, dear Alice. That modern physics has decided in favor of Plato, that the smallest units of matter are not physical objects in the ordinary sense, but forms, and ideas. And when you have complete mastery of these ideas, when you can hold them in your palm and twist and tease them at will—then you will have stepped closer to God.”
This dedication to intellectual truth and the rejection of human friction carries forward as the novel’s protagonist, Alice, journeys through Hell. A few pages later, she muses, “Why wouldn’t everyone strip away the parts of their selves that caused them pain?… You could achieve so much without the burdens of personhood.”
What turned me away from pursuing computer science academia was the distance of that intellectual edge. The last couple years of my Ph.D., I could count on my two hands the number of people in the world who were interested in the granular details of my research. That is the norm, not the exception, as I understand it. I found it demotivating, and since then the career I’ve built, mostly in public policy, has been quite the opposite in the participatory and collaborative nature of the work, and the immediate visibility of impact of its outcomes.
Today, AI is closing the gap between intellectual edge and immediate impact. Research papers are published on arXiv (a popular online repository to share free drafts of academic articles during the lengthy review and publication processes of journals) in parallel to startup companies founded by their authors receiving millions to turn the same ideas into products. While I first reached out to UW Tacoma because I wanted to teach again, I’m now (obviously) also looking into whether I have research contributions to make as well, downstream of the abstraction and resilience ideas in my book.
At the same time, AI is making it easier to generate research papers—not good ones, but nevertheless well-formatted and structured intellectual “slop”, indistinguishable at a glance from real work. Modern tools will happily take the barest of bones of an idea you give them and churn out a complete research paper for you, though whether its citations to other work are accurate (or even real) is an open question. And the sycophancy common in most tools translates here as well—the chatbot will eagerly tell you that the paper is publication worthy and a meaningful contribution to the field, when a human expert would be far more skeptical. Yet, the specialized nature of academic contributions narrows down to a bare handful, in some cases, the number of such experts who can easily spot the fake.
It all puts an interesting lens on to the nature of the pursuit of truth, as I see it. I’m using AI tools to code a proof of concept of architectures for assessing and improving resilience. I wouldn’t be able to build much of anything, with the limited time and resources I have available (not to mention my own rusty programming abilities), except with AI coding assistance. I do think of what I am doing as using AI to explore a deeper possible truth and advance some jagged direction of knowledge. I’m not trying to produce research papers out of it, yet, and in fact I don’t think I want to try that on my own; I’d rather find a co-author who knows those communities, who has those relationships, and who could help me identify which (if any) of the various things that could be written up as paper-shaped objects are actually worth producing. But I do hope to produce tools, working on my own, that have potential meaningful value in helping other people, by making the AI we interact with healthier and more resilient.
In Katabasis, the academic magicians—notably including Alice’s advisor Jacob Grimes—typically viewed themselves better than other humans, isolating the pursuit of truth above human concerns and considerations. They were all lifelong overachievers, gold star and straight A collectors, as so many who succeed in academia are (myself included, to be transparent).
But the takeaway I have from two decades of real world life after leaving full-time academia is that these metrics aren’t meaningful. The only true source of meaning is other people. If someone (not AI) finds your work interesting or useful, that makes it interesting or useful. Contrary to Jacob Grimes, there’s no “closer to God” to be found in the individual mastery of forms and ideas—only in how we can be good to each other. That’s what continues to motivate me at both micro and macro levels.