A tribute to June Sophia Ku (1981-2026)
June had an outsized influence on me and Misaligned Markets, and I'm forever in her debt.
June Ku was my partner of nine years and a core reason why Misaligned Markets exists. Two years ago, she gave me a very generous Christmas gift in the form of seed money for both this blog and a related project. June also graciously designed this blog’s logo. More than that, though, June was a mentor who nourished my philosophical curiosity, which very few people have meaningfully done.
My discussions with June, sometimes heated but always earnest, taught me a lot about philosophy while testing my intuitions and providing me with a sharper vocabulary to describe the world. June had a knack for teaching; even when she strongly disagreed with me, she left me with a more complete understanding of the weaknesses of my view and the strengths of hers. Her goal was never to dominate a discussion, but to point your gaze somewhere you couldn’t yet see. The intellectual charity June held toward ideas that weren’t hers is rare, especially outside philosophical circles.
June’s generosity and intellect were her defining features, and her self-described life purpose, metaethics, sat at the intersection of both. Much of her academic career revolved around ethics and semantics; she was deeply driven by how to formally express the values we take for granted in everyday language. June eventually left academia, but remained deeply committed to metaethical research. In 2009, she moved to Berkeley to work on MetaEthical AI, a proposed AI alignment solution. June spent the last 17 years laser-focused on this work, driven by intellectual curiosity, compassion, moral obligation, and, sadly, dread. While I have criticized the culture that sprouted AI alignment, I have never doubted the sincerity of individuals like June, who should be viewed separately from the leadership of EA organizations, AI safety labs, and AI companies.
June’s alignment work is distinct in that she was one of the few researchers with a background in formal philosophy—specifically metaethics, metasemantics, and metaphilosophy. This was critical to her project, which uses SetlX to formalize philosophical arguments with computer code. This put her in a very small niche, where I think even those committed to AI Alignment lacked the requisite background to fully appreciate her work.
I know June would find it disingenuous of me to highlight her work. We debated whether alignment is necessary or even a coherent goal for AI systems, and I have questioned the validity of her specific approach. But despite my reservations I supported her work because I know how much it meant to her. I won’t pretend to understand the project in full—it is an ambitious program at the intersection of philosophy, math, and computer science. Nonetheless, I can recognize when something has value beyond its stated purpose. MetaEthical AI isn’t just an alignment program; it genuinely explores ways of expressing and validating philosophical descriptions with the precision of computer code rather than the vagueness of natural language. MetaEthical AI has also informed an unsupervised approach to model interpretability.
Misaligned Markets sits downstream of MetaEthical AI not in content but in that June’s interest in alignment was my bridge to topics like cybernetics and evolutionary dynamics. If you enjoy Misaligned Markets you are in part enjoying something that June had a hand in shaping. While this post is a tribute to her, I consider the entirety of this project part of her legacy. In that same vein, the forthcoming next season of The Last Enclosure will be dedicated to her.
June, I don’t know if words can express my gratitude and the depth of the debts I owe you. But thank you so much for being my mentor, interlocutor, and muse. I hope I can become as insightful a philosopher and teacher as you were.
If you want to know more about MetaEthical AI, you can visit the project website or watch June’s 2021 presentation to the AI Safety Reading Group: