Daily Note (5th August 2026)
Finally had a couple of science meetings today and got to work on some science! Started work on a rough model for converting an r-band selection (think SDSS) to a physical property based selection (roughly mass-SFR-Av) using basic FSPS models and some SFHs from LP1, that can be used to compare predictions for things like luminosity functions from mock surveys to real data.
Talked with FC and DD today about a fun project they're working on that searches paper plots in a fun twist on Pathfinder, on applying for jobs, and AI/Data Science on campus - see the Hawaiian Data Science Institute that Peter S. is also a part of, the Aloha Intelligence Initiative and Lava.
At astro-ph today, William DeRocco talked about organizing the Rogue Worlds conference series to think about free floating planets, and an overview about a really great overview on this based on a wonderful review by Malena Rice The Dynamics of Planetary Ejection (Rice et al. 2026) bringing together 20 years of simulations and addressed a variety of scenarios and ejection mechanisms, which is going to be crucial when thinking about finding these systems with Roman.
Attended a talk at 1pm on using AI in research, which broadly covered MCPs, RAG, cli-based scriptable workflows, reproducing a published result from its paper (e.g. in lean for math papers). Some talk about local models with ollama, opencode. Demo with a RAG tutorial.
I'll be using initials for students and ERCs when I mention them to protect their privacy (at least for now).↩