Wolfram Computation Meets Knowledge

Wolfram Summer School

Alumni

Jonathan Miller

Science and Technology

Class of 2023

Bio

Jonathan Miller obtained his PhD in biology in 1988 as an experimentalist and a student of Aaron Klug at the MRC LMB, Cambridge, where at age 24 he discovered the first zinc finger. Miller suffered a PhD in theoretical condensed matter physics at Caltech in 1991 on equilibrium statistical mechanics of the ideal Euler fluid, resolving a problem posed by Onsager. He worked on chaotic Ising dynamics and the critical properties of the mean-field quantum spin glass at AT&T Bell Laboratories (11111) before moving to the University of Chicago as a post-doc of Leo P. Kadanoff. Miller studied superfluid helium at Caltech before joining NECI Princeton and later Princeton University. In 2002, he joined Baylor College of Medicine as an assistant professor, where he discovered novel microRNAs by computational methods, and identified a conserved sequence length distribution suggesting that standard methods at the time strongly underestimated selective pressure on short sequences. In 2008, the Nobel Laureate Sydney Brenner invited Miller to join a new English official language university in Japan as 15th PI; the campus was still a pile of sand when he arrived. There, Miller sought to remediate bigotry and misogyny imported – ironically – by Western academics. Observing that camouflaging cephalopods such as squid report to us their perception of their visual surroundings through real-time high-bandwidth channels, Miller began a program to deconstruct cephalopod perception to its informatic and neural origins. These days he is preoccupied with LLMs, which will leave no area of human endeavor unaltered. Miller attended the WSS2023 where he was stunned by the extraordinary brilliance, talent, dedication, patience, and thorough professionalism of its mentors, who reflect brightly on the Wolfram enterprise. Now is the age of GPT (LeCun estimates its lifespan as five years) and the Wolfram Language (lifetime no doubt much longer), and once this symbiotic construction figures out how to evolve on its own – as it most surely will – forces will be unleashed that Miller, a committed behind-the-scenes diversity advocate, hopes will not exacerbate the ever-accelerating divergence between the haves and have-nots. How can we bring the poorly-resourced and poorly-connected along together with us? Or will AI be to H. sapiens what H. Sapiens was to Neanderthal man, and exterminate us just the way we have annihilated the once rich diversity of life on Earth?

Project: In-Context Learning in LLM as an Emergent Phenomenon