#AcademicRunPlaylist - 2/22/25

Well I’m back in the US! I’m so grateful that I was able to spend some time in Japan, and I’m looking forward to spending a lot more time there over the coming months as a number of new projects kick into gear. While I’ll miss having my tuna onigiri breakfast, at least I was able to listen to a bunch of talks on the flight home for my #AcademicRunPlaylist!

First was a great talk by Siddharth Karamcheti on language-driven learning for interactive robotics at UW https://www.youtube.com/watch?v=eIW4TMPDy0k

Next was an intriguing talk by Felix Yanwei Wang on using interactive task specification for more effective robotic systems at Ai2 https://www.youtube.com/watch?v=E4JCO4RCSps

Next was an incredible talk by Robert Katzschmann on creating lifelike robots at Stanford. Katzschmann clearly covers the challenges of designing and debugging biohybrid robots (seriously). Highly recommend https://www.youtube.com/watch?v=DxqBvurAE50

Next was an interesting talk by Max Argus on using VLA models for robotic manipulation at Ai2 https://www.youtube.com/watch?v=5vI_gA1ByzA

Next was “Digital Empires” by Anu Bradford. This is essentially two books - one is a great analysis of the dominant international technology regulatory regimes and their strengths and drawbacks, and the other is an editorial on these topics. The first is informative and thought-provoking, and I also liked Bradford's identification of the fragmented nature of the EU's financial markets as the primary cause of its anemic tech output

Full review: https://bookwyrm.social/user/bwaber/review/6719340/s/a-good-international-legal-review-paired-with-an-already-out-of-date-and-sparsely-researched-polemic#anchor-6719340

Next was “GDP” by Diane Coyle. People throw around the concept of GDP like it's an objective entity, but rarely do they peak under the hood and understand how and why it's calculated in a certain way and the implications of those very subjective, human choices. Coyle reviews this history nicely, pointing out that this metric is very much designed for a manufacturing-intensive world, that many parts of the economy are left out, and how challenging it is to come up with accurate, coherent metrics in an ever changing world. I only wish this book were longer. Highly recommend https://press.princeton.edu/books/paperback/9780691169859/gdp

Last was “Voices in the Code” by David Robinson. This is probably the most in-depth book on the socio-technical nature of software systems that I've ever read. After starting off with a high level overview of how and why algorithms are developed in high stakes situations, Robinson dives into the case of organ transplantation. Examining the history, scientific background, and the messy and continuing process of developing algorithms to parcel out organs is enlightening. This should be required reading for anyone developing or using algorithms in high stakes settings. Highly recommend https://www.russellsage.org/publications/voices-code