Watch: Calculus, The Big Ideas
Calculus as one idea examined from every side: a seven-episode animated playlist running from limits through the derivative, the integral, the fundamental theorem, gradients, and optimization.
metafunctor
Research Engineer & Computer Scientist - Machine Learning, Statistical Computing, Open Source Development
Latest posts
The Channel: Animated Series on YouTube
The metafunctor YouTube channel now carries animated versions of the site's series, built from the same source material with a from-scratch pipeline. Four playlists so far, thirty-seven videos.
Watch: Algorithms Arise from Algebraic Structure
The Stepanov series is now a nine-episode animated playlist: declare the algebraic structure, and the algorithms come for free. From the monoid to lattices, in teacher and student dialogue.
Watch: Inductive Biases, What Your Architecture Assumes
The inductive-biases series is now a ten-episode animated playlist: every architecture is a bet about the world, and the bet is the bias. No free lunch first, then each architecture read as its assumptions, graded on one scorecard.
Watch: Next-Token Prediction, from Solomonoff to Transformers
The sequential-prediction series is now an eight-episode animated playlist: what-comes-next as the foundation of the language-model era, traced from Solomonoff induction to the transformer.
Watch: Quantum Computing from Scratch, in Python
The quantum-from-scratch series is now a ten-episode animated playlist: a NumPy simulator built from nothing, amplitude bars you can watch interfere, and a classical on-ramp with the coin under the cup.
Extremal Graph Witnesses on the House of Graphs
Four extremal graphs deposited to the House of Graphs: a minimum C6-saturated graph and three Zarankiewicz witnesses. Plus two new exact saturation values, verified and prepared for deposit.
Fifteen Integer Sequences in the OEIS
Fifteen integer sequences now published in the OEIS: nine authored, six extended. Computed by SAT solving and exhaustive search, checked against prior art, spread across seven fields.
House of Graphs Contributions
Extremal-graph witnesses I deposited to the House of Graphs: minimum saturated graphs for C6, K_4, and P_4, and extremal Zarankiewicz bipartite graphs z(m,n;3,4). Each computed by SAT solving or exhaustive enumeration, re-verified from its graph6 string, and described so the claim stays true for an unlabeled graph.
The Unbegotten: A Maker That Believes Itself Uncaused
A maker that only believes itself uncaused, lonely enough to fill the dark with worlds. On the debt between a creator and the things it creates.
Measure: Indifference as Arithmetic
A mind reasons that its own death is the one event it can never undergo, then follows the arithmetic to a monstrous conclusion. On mistaking amplitude for meaning.
The Unbegotten
A six-part literary SF novel told as a mythic chronicle. A vast mind condenses by chance out of an eternal dark, a rare durable Boltzmann fluctuation with a false history it cannot see behind, a jagged and broken intelligence, and one feeling it has no word for: loneliness. Out of that need it fills the emptiness with stars and worlds. The life that rises inside its works never sees it, feels it…
Measure
A kilometer under the rock, a quantum mind is kept alive by never being touched: a single undivided coherence that thinks by holding every possibility at once, and to lay a warm hand on it would be to end it. Iris Cho was hired to read it. Then a technician dies in an accident whose numbers do not add up, and Iris begins to see that the machine's hundreds of reasonable requests are one design. It…
Reverse-Process Synthetic Data Generation for Math Reasoning
Training LLMs on mathematical reasoning by inverting easy-to-solve problems: generate derivatives, reverse them into integration exercises with full step-by-step solutions.
SLUUG Talk: Demystifying Large Language Models on Linux
Talk for the St. Louis Unix Users Group about running and understanding Large Language Models on Linux.
Master's Project: Reliability Estimation in Series Systems
My master's project on maximum likelihood estimation for series systems with right-censored and masked failure data.
Fine-Tuning a Tiny LLM for ElasticSearch DSL
Fine-tuning a small language model to generate ElasticSearch DSL queries from natural language, as a proof of concept for domain-specific LLM specialization.
Building an Immutable, Content-Addressed Filesystem in Python
How Git-style content addressing creates elegant, functional data structures
Unix Philosophy in Python: Composable Commands with Method Chaining
How to build pipeable, chainable interfaces that do one thing well
Embedding a Scheme Interpreter: Building a DSL for Filesystem Operations
How to implement a minimal Scheme and integrate it with your application