Prompt Caching in Coding Agents: How prompt caching works in various harnesses.
Coding agents resend full context every turn and providers cache by exact prefix match. See how each harness shapes that prefix and why the economics differ.
Applied AI engineering: retrieval systems, text chunking, embeddings, document pipelines, prompt caching, and coding agent workflows for developers.
Coding agents resend full context every turn and providers cache by exact prefix match. See how each harness shapes that prefix and why the economics differ.
How I turned repeated AI coding mistakes into Oxlint rules, measured the existing mess, and made touched files clean up their own warnings.
How we used AI, profiling, adversarial tests, and repeatable Go benchmarks to cut hot-path work and make filtered log writes up to 77x faster.
A simple CLIProxyAPI setup to use Codex subscription models from Claude Code and Conductor, with model aliases, config files, and a debug checklist.
Use Factory Droid BYOK with Claude Max and Codex subscriptions through CLIProxyAPI, a local OAuth-to-API-key proxy: brew install, models, and fixes.
SayBackend changelog for January 2026: theme management, Mermaid diagram support, and infrastructure improvements.
Updated for Zustand v5: persist state in the URL hash so users can copy, share, and restore React app state across sessions and devices.
How to break down text for RAG systems: from character-based splits to adaptive chunking, and which strategies improve retrieval accuracy.
How to deploy a Next.js app to production without Vercel: standalone output, Docker, and a Node server that runs on any VM, container host, or cloud.
Build the best Dockerfile for Golang: multi-stage builds, dependency caching, scratch base images, and .dockerignore tips for small, fast Go containers.