Tag
The article presents a framework for deciding how much autonomy to give AI agents based on two factors: ease of checking the output and ease of undoing errors. It introduces four levels of delegation, from agent as assistant to full self-driving mode, and illustrates with a decision tree.
PostHog's engineer used multiple long-running Claude Code sessions to rewrite their SQL parser, achieving a 70x speedup over the previous ANTLR-based parser, while barely looking at the code themselves.
PostHog used an AI agent based on Karpathy's autoresearch to find a three-year-old bug in their ClickHouse query engine that prevented proper primary key usage for timestamp filters. Fixing it improved performance by 11% and reduced scanned granules by 62%.
BossHogg is an agent-first command-line interface (CLI) tool designed for managing PostHog analytics and feature flags.