Random thoughts, some cool ideas, some cool ways to communicate or think about ideas, some more math-y/machine learning stuff. If you have ways in which these can be improved for clarity or accessibility please let me know! I reserve the right to edit and update these at any point in the future.

Math-y, More Formal Posts

• Coordinal: A Postmortem.

  Most startups fail, this one did too. An account of the startup, the startdown, and lessons learned.

• Interpretable by Design - Constraint Sets with Disjoint Limit Points

  The simplex is a nice interpretable set with some useful properties.

• GPT-2 Sometimes Fails at IOI

  For Lisa, GPT-2 does not do IOI.

• Incorporating Preferences is Easy if You Discretize

  Continuous settings can be hard, but matching and moving arbitrary distributions is easier if you discretize.

• Interpretability As A Science

  Knowing the typical or base distribution of a measure is important for interpreting a specific instance of that measure!

Misc-y, Personal-y Posts

• A Bit For You

  Some musings on a single, but potentially most important, bit.

• Scratch/Progress on Site Building

  Scratch post for site imports, styling, etc. Skippable unless you're interested in the site's construction.

• Less Muse and Thicker Plots

  Some old but very pretty and satisfying visualizations for a pointless analysis. (JS, JQuery, HighCharts)