

Experienced CTO-level Rails architect
Información privadaDescripción
Kent Beck recently said:
“AI has reduced the value of 90% of what I do to zero, but the other 10% has gone up by 1000x.”
So what’s in that “other 10%”? It’s the part non-engineers rarely appreciate — because they don’t see it firsthand:
• Debugability
• Introspection
• Solid structure
• High-quality tests
Let’s be honest: debugability in AI-generated code is a disaster. We’re witnessing a flood of code written by non-engineers — vibe coders — that seems fine until it explodes months later. That’s when the real nightmares begin: bugs that require eight steps across different actors, cause baffling side effects in unrelated parts of the app, and defy every attempt at consistent reproduction. They’re unfixable, unscoped, and they never die.
Introspection — the ability to look clearly into your code, your data, your stack — is an art. And as far as I can tell, AI isn’t introspecting. So it doesn’t care if your data is obscured or lost behind poorly constructed abstractions. But you should care.
Solid architecture should be non-negotiable by now. Yet somehow, we still see apps with business logic scattered across layers, leaky abstractions, and no internal consistency. We once thought these lessons would become table stakes. They haven’t.
And then there are tests. What counts as a high-quality test? In my paper, “Automated Tests from 10,000 Feet,” I explore how effective testing is about intent and confidence. In healthy codebases, tests are written with the code — not bolted on later. But for decades, computer science programs have overemphasized unit testing, neglecting system testing. Now we have a generation rebelling against tests altogether, many of them openly hostile to the practice.
Why? Because we failed to demonstrate the real value of tests. Unit tests rarely catch regressions. But a lean, strategic suite of high-quality system tests? That’s pure gold. That’s what gives you the confidence to refactor, to ship, to move fast without breaking everything.
It’s time for a course correction.
Are you building something that needs a founding engineer?
I’m looking for 50% equity, with a $0 salary.
If you're serious, let’s talk.