The AI Revolution Is Creating Two Types of Workers
Which type will you be when the dust settles?
Two weeks ago, I had a conversation that stopped me in my tracks.
A fellow CTO reached out to share his company's "AI transformation." Every developer on his team was using AI coding tools exclusively. He was evangelical about it - sharing metrics, success stories, the works.
Then he dropped the real number: his most senior developer, the most prolific AI user on the team, was only accepting 30% of the AI-generated code suggestions.
Let that sink in. The person who knew how to use these tools best was rejecting 70% of what AI recommended. Why? Because he had the depth of knowledge to recognize what would actually work and what would create problems down the road.
This isn't a story about AI failure. It's a story about the irreplaceable value of human expertise.
The Agile Déjà Vu
We've been here before. Remember when everyone became an "Agile coach" overnight? Suddenly every consultant had Agile expertise, every company was "doing Agile," and the market got flooded with surface-level knowledge masquerading as deep experience.
The same thing is happening with AI. Everyone's an expert now. Everyone has access to the same tools, the same knowledge base, the same instant answers.
But here's what's getting lost in the rush: the craft of actually knowing your domain deeply enough to guide, evaluate, and improve what AI produces.
Where AI Excels, Where Humans Remain Essential
I use AI daily for tasks that used to drain my time and energy. YouTube tags for our podcast episodes? AI nails it every time. First drafts of technical documentation? AI gives me a solid starting point. Brainstorming session agendas? AI helps me think through angles I might miss.
These are tasks where AI can process patterns from massive datasets and give me results that would take hours to generate manually.
But here's what AI can't do: it can't bring the weight of lived experience to complex decisions. It can't recognize the subtle patterns that come from having built systems that failed spectacularly, or having navigated team dynamics through multiple economic cycles, or having learned hard lessons about what works in practice versus what looks good on paper.
When Bob and I discuss leadership challenges on our podcast, we're not just sharing theoretical frameworks. We're drawing from decades of being in the room when things went sideways, making decisions with incomplete information, and learning from both our successes and our spectacular failures.
That's human intelligence. That's the craft.
The 30% Rule
That senior developer accepting only 30% of AI suggestions? He's practicing the craft of leveraging AI rather than being leveraged by it.
He understands the codebase architecture deeply enough to know when a suggested solution will create technical debt. He's experienced enough system failures to recognize patterns that might cause problems in production. He's built enough integrations to know which approaches will be maintainable six months from now.
This is what separates someone who uses AI tools from someone who has mastered the craft of working with AI.
Your Competitive Advantage
The future doesn't belong to people who can prompt AI well (though that's useful). It belongs to people who combine deep domain expertise with the ability to leverage AI as a powerful tool.
If you're a leader, your value isn't in having all the answers instantly available. It's in knowing which answers matter, understanding the second and third-order effects of decisions, and bringing the wisdom that comes from having navigated similar challenges before.
If you're a developer, your value isn't in writing every line of code from scratch. It's in architectural thinking, in understanding system design tradeoffs, in knowing how to build maintainable solutions that scale.
If you're in any knowledge work, your competitive advantage is the depth of understanding that lets you evaluate, refine, and improve what AI produces.
The Challenge
Here's my challenge: invest in learning the craft of working with AI, not just using AI tools.
Experiment with different AI tools across multiple projects. Learn where each excels and where each falls short. Develop your ability to prompt effectively, evaluate output critically, and integrate AI-generated work with your own expertise.
But most importantly, never stop developing your core competencies. The deeper your domain knowledge, the more valuable you become in an AI-augmented world.
I've been intentionally working with different AI tools across multiple projects - not because I want to be an AI expert, but because I want to be an expert at my craft who happens to leverage AI effectively.
When someone needs real expertise three years from now, they won't be looking for someone who can use ChatGPT. They'll be looking for someone who has built real solutions, learned from real failures, and developed the judgment that only comes from experience.
AI is fantastic at processing information and generating starting points. Humans are special at bringing wisdom, context, and judgment to complex problems.
Use AI to highlight and support your specialness, not replace it.
Josh Anderson
Editor-In-Chief
The Leadership Lighthouse
From the Podcast Studio
This newsletter was inspired by a conversation Bob Galen and I had on our latest episode about the parallels between the current AI boom and what happened during the Agile transformation years. We dove deep into the "AI versus HI" debate, shared real stories from the trenches, and discussed why human intelligence remains irreplaceable even as AI tools become more powerful.
If you enjoyed this perspective, you'll love hearing the full conversation where we get into the weeds about AI slop, the dangers of zero human intelligence workflows, and why we think this moment feels eerily similar to the Agile cash grab of the 2000s.
Listen wherever you get your podcasts or watch on YouTube.
Got thoughts on AI and human expertise? We'd love to hear from you - drop us a line or leave a comment on the episode.