Somewhat Useful
-
Jan 13, 2026
A (growing) list of things that have made me thing about the role of AI in my job/life/planet
-
Dec 3, 2025
Neo4j has opinions about how data is handled...
-
Dec 3, 2025
Sankey diagrams illustrating implicit choices.
-
Aug 21, 2025
The type system only hints at some important information
-
Oct 12, 2024
The "bitter lesson" is that _complexity_ is ultimately the limiting factor.
-
Oct 10, 2024
Automation is hard because it is mostly about errors
-
Oct 2, 2024
Personal packages are useful
-
Jul 16, 2024
Visualizations to assist program undersanding
-
Apr 16, 2024
Brittle-like-a-fuse is a good thing for all systems
-
Mar 15, 2024
Most lesson material is a prop for teaching *something else* that is hard to directly approach.
-
Feb 6, 2024
The world is both bad and better, and that is a good thing.
-
Jan 29, 2024
A notion of edit distance using the standard library
-
Jan 5, 2024
Systems express opinions on what they make easy & hard. Vega makes it hard to abstract over names. This isn't necessarily bad, but keep it mind when you make your schemas.
-
Dec 20, 2023
"Static vs. Dynamic" is such a straight-jacket...
-
Nov 29, 2023
Failure helps you grow.
-
Aug 21, 2023
Part of the answer to "what does graphviz's `dot` algorithm actually do?"
-
Aug 7, 2023
Type and data-level programs need the same low-level concepts. Why do they look different?
-
Jul 31, 2023
Make parallel lines from a single set of endpoints using Vega transforms.