Ideas on Architecture & AI
Notes on building distributed systems as executable graphs, and on why AI agents need something better than a repository to understand them.
Go vs Rust vs C++: Why Our Service Benchmark Surprised Me
Go led our generated-service benchmark. Rust led the simplified native versions. What an HTTP-to-gRPC workload taught us about integration costs, buffers, and choosing a stack.
AI Wrote the Service. How Do We Review It—and What Does That Cost at Runtime?
Generated composition can make a system easier to understand, review, change, and observe. A measured comparison against direct Go, Python, C++, and Rust implementations asks what that convenience costs at runtime.
One Graph, Four Native Runtimes: Preserving Architecture Without Erasing the Language
A practical test of language-agnostic architecture: what stays shared, what remains native, and how one typed graph becomes working Go, Python, C++, and Rust services.
Why a Repository Is Not Enough for an AI Agent: Architecture as a Shared System Representation
AI can already write functions and fix bugs, but understanding a large system is a different problem. An engineering case for a formal architectural layer — between intent and source code — that both humans and AI agents can use.