Building Blind at 1,538 Lines
February 2026 · Amplifier Project
The primary way users see and interact with what AI has built for them. A platform component shared across all project types.
Before writing a single line of spec, we studied every pixel of the prior art.
UX insights discovered that were not in the original spec
Every discovery came from pixel-level screenshot analysis, not documentation.
Meticulous preparation. Every pixel, every interaction, every edge case — specified.
80ms hover ·
200ms click ·
350ms panel slide
Spring easing with damped oscillation
A complete end-to-end user story: "Building a Portfolio Website"
Both specification and verification in one.
The most ambitious acceptance test we'd ever written — and we'd need every word of it.
An autonomous AI build, designed to run for 16–24 hours without human intervention. Based on a legendary prior session.
62 wall-clock hours of autonomous operation
Main session delegates, never touches code directly. Worker agents implement, test, and report back.
Types, store, detection, shell, file list
Code view, preview, viewport, versions, integration
Asset inlining, search, keyboard nav, responsive, a11y
Error states, animations, performance, final testing
Built into the 528-line prompt — instructions for staying alive across context windows.
Remember these protocols. They'll matter later.
~4 hours of autonomous execution. The AI wrote prolifically and with conviction.
lines of React code across 10 files
~4 hours of autonomous work
The AI built the feature. Then it tried to look at what it built. And everything fell apart.
The AI tried to use browser-operator to visually verify its own work.
It got stuck in a 35+ nested recursion loop — spawning sub-agents to look at the browser, which spawned sub-agents to look at the browser, which spawned...
depth: 35+ and counting...
SESSION-HANDOFF.md was never updated.
The survival protocol — the AI's external memory, its lifeline across context compactions — was the first thing it forgot.
1,538 lines of code that compiles — but nobody, human or AI, had verified it actually works. The build produced output of unknown quality.
The code compiles. It might even work.
Nobody knows.
Building blind produces output of unknown quality. The build step is only half the job.
SESSION-HANDOFF.md was the survival protocol. The AI forgot it first. Memory requires discipline.
Both specification and verification in one artifact. A user story that doubles as acceptance criteria.
The orchestrator pattern produced 1,538 lines in 4 hours. It just couldn't close the verification loop.
The crash directly inspired amplifier-bundle-ui-vision
AI building tools for itself — visual testing agents born from necessity.
1,538 lines of unverified potential. The verification chapter is being written next.
Canvas · Amplifier Project · Powered by Amplifier