<!--
agent-site: nibzard.com
collection: log
canonical-url: https://nibzard.com/converted
markdown-url: https://nibzard.com/converted.md
content-signal: ai-input=yes, ai-train=yes, search=yes
author: Nikola Balić
published: 2025-06-07
tags: HUMAN, OPINION
related: /agent-md, /agent-native
-->

---
title: The Day the Skeptic Blinked
description: "Journey from AI skeptic to convert--proving that the future belongs to experts who learned to work with machines."
tldr: "Kenton Varda, a Cloudflare engineer who was skeptical of AI, tested Claude by building an OAuth library. The code was surprisingly good, leading him to realize the power isn't in AI replacing humans, but in the combination of AI speed and human expertise."
date: 2025-06-07
tags: [HUMAN, OPINION]
draft: false
author: "Nikola Balić"
topics: [AI skepticism conversion, OAuth library development, code quality, security review, expert-AI collaboration]
entities: [Kenton Varda, Cloudflare, Claude, OAuth, RFC]
answers_questions:
  - What converts AI skeptics into believers?
  - How do security experts validate AI-generated code?
  - What makes AI-human collaboration more powerful than either alone?
---

Blink, and your entire worldview could shift.

[Kenton Varda](https://x.com/KentonVarda), lead engineer at Cloudflare, was convinced AI was overrated. "Glorified Markov chain generators," [he called them](https://github.com/cloudflare/workers-oauth-provider/?tab=readme-ov-file#written-using-claude). Pattern matchers pretending to understand code.

So he decided to prove it. Build an [OAuth library](https://github.com/cloudflare/workers-oauth-provider/?tab=readme-ov-file#written-using-claude) with Claude. Watch it stumble. Have a quiet laugh.

Plot twist: The code was actually good.

Here's what matters--this wasn't some vibe-coding experiment where you trust the machine blindly. Every line was reviewed by security experts. Cross-referenced with RFCs. When Claude made mistakes (and it did), humans caught them and guided the fix.

The real insight? This isn't about AI replacing humans. It's about the combination of AI speed and human expertise creating something neither could achieve alone.

We keep asking the wrong question: "Will AI take my job?"

The better question: **"How do I become the kind of human that makes AI more powerful?"**

The landscape is moving fast. Token economics are shifting. Interfaces are commoditizing. The flywheel effect is spinning up for the big players.

And somewhere in this chaos, a skeptic just got converted.

The future belongs to experts who learned to work with machines.

Your move.