John "Jake" Baumgarten·Jun 26The Most Valuable Codebases Live on the Right Side of the Edge of ChaosNote: The following is heavily AI-assisted, but it took me over a year of reading physical math and CS textbooks to be able to collaborate…
John "Jake" Baumgarten·Jun 3What Symbolic Dynamics and Category Theory Can Teach Us About AI-Driven Software EvolutionNote: The following is AI-generated, but it took me over a year of reading physical math textbooks to be able to frame the discussion with…
John "Jake" Baumgarten·Nov 8, 2025Visualizing Refactored Claude-Generated LangGraph CodeUsing AI to Learn About AI
John "Jake" Baumgarten·Mar 2, 2025Visualizing a Graph Attention LayerWhat does a layer in a graph neural network that employs self-attention look like? Like this:
John "Jake" Baumgarten·Feb 5, 2025PyTorch Graph Neural Network Training via DelegationBelow is a place-transition diagram of the code we are going to explore. In the diagram below the module VanillaGNN is delegating fit…
John "Jake" Baumgarten·Jan 5, 2025Making GNN Recommender Code “Say the Right Thing”When Is 47 Lines of Code Better than 5? When the 47 “Says the Right Thing”
John "Jake" Baumgarten·Nov 27, 2024Visualizing a PyTorch Wasserstein GAN Training LoopOur brains think like DAGs — directed (mostly) acyclic graphs. Computers think like stack machines. PyTorch processes via large scale…
John "Jake" Baumgarten·Oct 16, 2024Visualizing Cloud Native Go — Story 3BStory 3B — When You Give a Tech Talk, Developers Ask Tough Questions
John "Jake" Baumgarten·Jun 13, 2024Visualizing Neural Network Python — Story 3Story 3 — A Load, Train, Visualize Python Network
John "Jake" Baumgarten·Jun 13, 2024Visualizing Neural Network Python — Story 2Story 2 — Software Graphical Idioms for Deep Learning