Blog Archive

2025

• Tokenizer decasing
• kwargs.pop is probably a code smell
• Using overload to handle tagged union return types
• Protocols to make untyped code behave
• Rethinking evaluation and relative performance
• Exposing string types to maximize user happiness

2024

• String casing in python
• Correctly typing cached functions
• NewType in python
• TypeVars and Unions in python
• Enums with superclasses
• Enums and refactoring
• Correctly typing recursive hierarchies in Python
• Solving a ForwardRef error in pydantic

2023

• The burden of proof for code reviews
• Turning a byte pair encoded string back to its surface form

2020

• Fast way to find the rank of an item in pytorch

2019

• Using itertools.product with dictionaries
• Using itertools.product instead of nested for loops
• Parsing a spacy json to a tree
• Recursively dealing with trees
• Getting spacy to only split on spaces

2017

• Poster: A Self-Organizing Model of the Bilingual Reading System
• Talk: Lexical representations in theories of word reading