These primers are self-contained, interactive documents that explain concepts through visualization and narrative. They're designed to be read, explored, and shared.
Each primer follows the same editorial format: warm typography, paper-like textures, and interactive elements that illuminate rather than decorate. Think of them as small books you can read in your browser.
Why AI Keeps Surprising You
An introduction to scaling laws, emergent capabilities, and why your intuitions about AI limitations are probably stale. The foundational piece for understanding where AI is headed.
Primer 02Visualizing an LLM: How Language Models Think [DRAFT]
From random numbers to coherent text. A visual exploration of layers, vectors, and the geometry of language models. The foundation for understanding mechanistic interpretability.
The Case for Super Agents
Why general-purpose agents outperform specialized ones as models improve. Explores fitness landscapes, context decay, and the asymmetric bet on capability growth.
Primer 04Context Windows: Many Trees Make a Forest [DRAFT, UNREVIEWED]
On context windows, working memory, and why decomposition beats loading everything into context. The limits of scale and how to work within them.
Storytelling with Data
Why the future of data work is neither dashboards nor internal apps, but something more expressive. A case for data applications that tell stories.
Primer 06The AI Data Stack
What happens when AI agents need to query your data warehouse? A framework for building data infrastructure that AI can actually use reliably.
Growing Services with AI
Why building is the wrong metaphor. Services should grow toward goals, with AI agents tending conditions rather than constructing step by step. Grow, don't build.
Primer 08AI-Native Infrastructure
What infrastructure looks like when AI is a first-class participant rather than an afterthought. Designing systems that agents can understand, operate, and improve.