Topics
Four layers, from the math up to the plumbing. Pick a layer, pick a topic.
// foundations
Foundations
The ideas everything else is built on.
Machine Learning
The learning loop: models, loss, gradients, generalization, and how not to fool yourself.
1 article
Deep Learning
Layers, backprop, transformers and attention: the machinery behind modern models.
1 article
Models
Base, instruct, reasoning, open-weight, hosted: what the names mean and how to choose.
1 article
GPUs
Memory, bandwidth, batching and precision: enough hardware to make good decisions.
1 article
// trends
Trends
What the field is arguing about this month.
Prompt Engineering
Writing the specification a model can actually follow, and testing it like code.
1 article
Context Engineering
Deciding what goes in the window, in what order, and what gets thrown out.
1 article
Harness Engineering
The code around the model: tool execution, permissions, sandboxing, recovery.
1 article
Loop Engineering
Designing the agent loop: state, stop conditions, budgets and failure modes.
1 article
// patterns
Patterns
Architectures that keep showing up in real systems.
// tooling
Tooling
The plumbing: servers, workflows, observability.