Why AI Didn't Actually Make You Ship Faster — Gabriel Spencer-Harper, Meticulous
AI Engineer Code 2025 · 11:15
Automated frontend testing and regression detection
Meticulous builds automated frontend testing software for engineering teams developing web applications. It records browser interactions during everyday development and turns them into tests that evolve with the application. Teams review screenshot differences before merging changes, catching problems such as broken navigation, missing elements and changes in frontend calculations. In July 2026, Meticulous named Dropbox, Notion, Wiz and ElevenLabs among its customers.
Its record-and-replay architecture selects sessions that exercise different user flows, then runs them against both the proposed code and the base version. Recorded API responses replace live backend calls, preserving the original session’s data and allowing frontend tests to run without a backend. Meticulous combines this with distributed browser execution and a modified Chromium browser designed for deterministic replay. The technical focus is on making a growing collection of realistic interactions practical to rerun, without requiring developers to maintain each flow as a handwritten test. Coding agents also receive a feedback loop before submitting changes for human review.
The company’s technical perspective includes Quentin Spencer-Harper’s experience testing large frontends at Palantir, where he spent ten years helping lead engineering for its main frontend. His writing emphasizes test maintenance costs and using ordinary development interactions to build regression coverage. On July 14, 2026, Meticulous announced a $15 million raise to further develop its automated testing platform. The announcement framed change review and feedback as bottlenecks as AI makes code generation less expensive.
AI Engineer Code 2025 · 11:15
Affiliations reflect their AIE appearances, not necessarily current employment.