Discussion & Related

What I Learned by Analyzing My Own Research as Data

I asked an AI to analyze 140+ repos and 50+ papers as a dataset. The unifying thesis it found: compositional abstractions for computing under ignorance.

October 15, 2025 · 8 min read

Bernoulli Types: A Foundation for Approximate and Oblivious Computing

A unified type-theoretic foundation for probabilistic data structures, approximate computing, and oblivious computation with information-theoretic privacy guarantees.

October 14, 2025 · 4 min read

Encrypted Search and Oblivious Types

Oblivious types give encrypted search information-theoretic privacy against access pattern leakage. No ORAM, no computational hardness assumptions. Here's how.

October 14, 2025 · 4 min read

Discussion