Why AI Didn't Actually Make You Ship Faster — Gabriel Spencer-Harper, Meticulous

As agents write code faster than people can review it, assertion-based tests fall behind. This describes a way to run visual regression checks on every pull request without hand-written tests.

• Spencer-Harper argues the space of possible frontend regressions is too large for hand-written assertions, so verification, not code generation, is now the bottleneck.
• Meticulous records real user flows and replays them on every pull request, showing reviewers visual before and after diffs instead of relying on assertion tests.
• Three techniques make replay reliable, per the talk: fully mocked network traffic, a deterministic browser that removes flaky tests, and coverage-guided selection of which recorded flows to replay.
• Because every step is screenshotted, he argues code coverage finally means code that was actually checked, not just code that ran.

• videoLessons from Generating 12 Trillion Synthetic Tokens — Bogdan Gaza, DatologyAI
• videoStop Rationing Tokens: Let the Harness Pick the Model — Kimchi by Cast AI
• videoWhy 99% Accurate Browser Agents Still Fail — Derek Meegan, Browserbase
• videoYour LLM App Returned 200 OK. It Was Still Wrong. — Marina Petzel, Datadog

• videoThe Death of the Code Review: What the Data Actually Says — Laurie Voss, Arize AIAI Engineer
• videoHow to Get Your Org to Adopt Coding Agents (Without Shipping Garbage) — Eyal Blum, FigmaAI Engineer
• videoHow to Kill the Code Review — Ankit Jain, AviatorAI Engineer

Checking sign-in…

Loading comments…