@DeRonin_ I'll add one more for you - if you want end to end testing on what you did make, add the Shiplight plugin and get easy, full coverage
https://t.co/o22GVCKhCC
AI made writing checks cheap.
it did not make senior engineering time cheap.
the real cost shows up when experienced engineers lose mornings inside triage loops that should already have a clear path to resolution
the hidden org problem in AI-generated browser testing:
nobody explicitly owns the ambiguity loop
the failure happens
everyone sees red
and the team loses time deciding whether itโs app, test, infra, or config
the slowdown isnโt the issue itself
itโs unclear ownership of the path to resolution
AI didnโt make browser failures more expensive.
It made ambiguity more expensive.
The real tax isnโt a broken test.
Itโs the time engineering teams lose figuring out whether the failure is a real regression or just suite drift after the UI changed.
The teams actually getting leverage from AI coding arenโt the ones generating the most tests.
Theyโre the ones shortening the time between code being written and confidence being restored.
The moment verification becomes a separate maintenance workflow, the old QA bottleneck just reappears inside engineering.
The real unlock is keeping browser validation inside the coding loop itself
AI agents can generate 100 browser tests in minutes
the real tax starts 2 weeks later
selectors drift
UI shifts
flows get reworked
now an engineering manager owns a flaky suite nobody asked for
that maintenance burden is becoming the hidden cost of AI coding velocity
We eliminated 85% of the manual UI testing we used to do before every merge. Now we're giving the tool away for free.
Today we're launching Shiplight MCP. One line install, and your coding agent can verify its own work in a real browser before you ever look at it.
Write the code. Navigate the UI. Take screenshots. Catch regressions. All in one agent loop.
Works with Claude Code, Cursor, and Codex. Completely free and token-efficient.
The bottleneck in agentic development isn't writing code anymore. It's verifying it. We're closing that gap.
Quick Start ๐ https://t.co/VWhC71cweZ
Demo ๐ https://t.co/gabAhFNoqs