The hottest take in tech right now is that AI is coming for developers' jobs. Every week brings a new breathless article about how coding assistants will make software engineers obsolete, how anyone can build an app with a prompt, how the entire industry is about to be disrupted beyond recognition. I've read most of them. I've nodded along with some of them. And I'm not worried.
I say this as someone who has gone deeper into AI-assisted development than most. Over the past six months, I've built WitFoo's next-generation analytics platform almost entirely with Claude Code. I've written two detailed blog posts documenting the process (one on building from scratch, one on migrating legacy code). I've trained my team on these workflows. I use Claude for research, troubleshooting, strategy, and storytelling. I am, by any reasonable measure, a power user who has fully embraced the technology.
And I'm telling you: Claude is not a threat to WitFoo, and it's not a threat to my career. Here's why.
The Wizard Still Wears the Hat
I've written before about the three personas I use to navigate my professional life: The Wizard, The Warrior, and The Poet. The Wizard devises strategy, envisions the grand scheme, and charts the course. The Warrior executes with discipline and precision. The Poet provides meaning, reflection, and inspiration.
Claude is an extraordinary Warrior. Give it a well-defined plan, clear standards, and the right guardrails, and it will execute with speed and consistency that no human can match. My two previous posts document this in detail. The structured prompt workflows, the phased approach to PR execution, the systematic testing and documentation (all of that is Warrior work amplified by AI).
But the Wizard's work? That still comes from me.
When I sit down to figure out how WitFoo's Empathetic Processing engine should handle a new category of cybersecurity data, Claude isn't generating that insight. When I recognize that our Conductor product can disrupt the data pipeline market by reducing compute costs by 98%, that recognition comes from a decade of watching customers struggle with overpriced, underperforming tools. When I see the pattern that connects cybersecurity operations to the Five Eyes intelligence framework and envision a national CyberGrid strategy for New Zealand, that's not something I can prompt my way to.
The Wizard's power comes from years of lived experience, cross-domain pattern recognition, and the kind of strategic intuition that only develops through thousands of experiments, hundreds of customer conversations, and more than a few spectacular failures. Claude can research any topic I point it at. It can synthesize information beautifully. But it can't look at a broken industry and see the thing nobody else has seen. That's the Wizard's job, and the Wizard is still very much human.
The Poet Cannot Be Automated
Here's where the "AI replaces everything" narrative falls apart most completely.
The Poet in my framework is the persona that provides meaning, context, and inspiration. The Poet weaves present endeavors into the fabric of history and time. The Poet challenges the expedient paths chosen by the Wizard and the Warrior, seeking to ensure that actions align with a righteous and meaningful trajectory.
Claude helps me research and structure my writing. It can find data points, suggest frameworks, even draft passages. But the Poet's work is irreducibly human. When I write about sitting in the wonder at Hagley Oval watching cricket as a New Zealand sunset paints the sky, that's not a prompt output. That's a lived moment filtered through a specific consciousness. When I tell the story of WitFoo's founding (the frustration, the conviction, the years of proving it could work), the emotional truth of that narrative comes from having actually lived it.
The Poet's role is to ask the profound, often ineffable questions that guide the Wizard and the Warrior. Why are we doing this? What does it mean? Is this the right path? These aren't questions Claude can answer for me, because they require something Claude doesn't have: a stake in the outcome. Claude doesn't care if WitFoo succeeds. I do. My team does. Our partners and customers do. That caring is the source of the Poet's power, and no amount of language modeling can replicate it.
Claude assists beautifully in the mechanics of storytelling. The soul of the story still has to come from somewhere real.
The WitFoo Way Is Not a Prompt
One of the arguments for AI disruption goes something like this: "Now anyone can code, so the barrier to entry for software companies has collapsed." There's a kernel of truth here. Claude does make coding more broadly available. Someone who couldn't write a line of Go last year can now produce functional code with the right prompts.
But here's what that argument misses entirely: code is not a product. A product is the code plus the understanding of the problem it solves, the years of research that informed its architecture, the battle-tested operational knowledge, the go-to-market strategy, the customer relationships, and the ability to sustain and evolve all of it over time.
At WitFoo, we call our approach the WitFoo Way. It's a living document that captures our philosophies, architecture, and practices. It specifies our frameworks, our testing strategies, our security posture, our performance requirements. It represents the distilled knowledge of nine years of research, 4,000+ experiments with Fortune 500 companies, universities, and government agencies, and a US Navy Cyber ANTX exercise win that validated our approach against funded competitors.
Can someone use Claude to build a cybersecurity analytics platform? Sure, in the same way someone can use a hammer to build a house. But the WitFoo Way isn't about hammering nails. It's about knowing which nails to hammer, where to place them, and why that particular configuration will hold up when the storm comes. Our ability to solve complex systems and human problems (the interplay between technology, process, and people in cybersecurity operations) didn't come from code. It came from a decade of sitting with customers in their worst moments and figuring out what actually helps.
Claude can't replicate that. Neither can anyone who prompts Claude to "build me a cybersecurity platform."
And sustainable go-to-market strategies? Our channel partner model, our Five Eyes expansion approach, our decision to position New Zealand as a test laboratory for national cyber defense? Those are strategic choices informed by relationships, geopolitical understanding, and a specific theory about how cybersecurity should work at national scale. That's Wizard territory, not Warrior territory, and no AI is generating those insights from a prompt.
Spinning Faster, Not Spinning Out
Here's what Claude actually does for me and for WitFoo: it makes us faster.
My 7 Rules of Efficiency include "Slay; Don't Wound" (finish what you start), "Think then Slay" (compartmentalize planning from execution), and "Part with My Toys" (delegate and automate when possible). Claude is the ultimate expression of Rule #4. It's the most powerful delegation and automation tool I've ever used.
When I'm in Wizard mode, Claude accelerates my research. I can explore a new technology landscape, synthesize competitive intelligence, and map strategic options in hours instead of days. When I'm in Warrior mode, Claude executes the plan with a speed and precision that would take a team of developers to match. When I'm in Poet mode, Claude helps me structure narratives and find the right data to support the story I need to tell.
Claude helps me personally spin faster in understanding problems and delivering solutions. It helps WitFoo organizationally do the same. Our small team punches far above its weight because Claude amplifies every member's capabilities. We're a rock band, not an orchestra (to borrow my own metaphor), and Claude is like giving every musician a better instrument.
But faster isn't the same as different. We're still playing the same music. We're still the ones who know the songs.
The Wit Is Still Human
Here's the thing about our company name. WitFoo isn't arbitrary. "Foo" is the hacker's placeholder, the variable, the building block. It's the craft, the technical execution, the code and the data and the systems. "Wit" is the intelligence, the insight, the human judgment that makes the Foo meaningful.
Claude provides a massive upgrade to the Foo side of the equation. It writes cleaner code faster. It documents more thoroughly. It tests more systematically. It refactors legacy code without complaining (unlike every human developer I've ever managed, myself included). The Foo has never been better.
But the Wit? The Wit is still in the human thought and experiences. The Wit is knowing that a 98% reduction in compute costs isn't just a technical achievement but a market-reshaping competitive advantage. The Wit is understanding that cybersecurity isn't really a technology problem but a human coordination problem. The Wit is recognizing that sustainable business practices matter more than hockey-stick growth curves, even when every VC in Silicon Valley is telling you otherwise.
The Wit comes from 10 years in the US Navy, from watching cybersecurity mature (and fail to mature) across an entire career, from building a company through a pandemic, from relocating to the other side of the world and seeing your industry through entirely new eyes. No language model, however sophisticated, is generating that kind of insight. It can accelerate the expression of that insight. It can help organize and communicate it. But the insight itself remains stubbornly, irreducibly human.
Wrap Up
I'm not worried about Claude replacing me or threatening WitFoo for a simple reason: Claude makes everything we already do better and faster, but it doesn't (and can't) do the things that make us us.
The Wizard's strategic vision. The Poet's search for meaning. The WitFoo Way's hard-won operational knowledge. The relationships, the intuition, the conviction that comes from nearly a decade of proving something works when the industry said it couldn't. Those are the moats, and they're not the kind that AI drains.
If anything, Claude has made me more valuable, not less. I can now execute at a pace that matches my vision, which has always been the bottleneck. The Wizard always had more plans than the Warrior could handle. Now the Warrior has a force multiplier, and the Wizard can finally think as big as the problems demand.
So to anyone asking whether AI coding tools are a threat to experienced practitioners and their companies: relax. Sharpen your Wizard. Feed your Poet. Let Claude handle the Warrior's workload. And remember that the most important things in any business (the strategy, the meaning, the relationships, the hard-won knowledge of what actually works) are exactly the things that can't be automated.
The Foo has been upgraded. The Wit remains ours.
Charles Herring is co-founder and CEO of WitFoo, a cybersecurity company building collective defense solutions. He is a US Navy veteran and speaks regularly at security conferences including DEFCON, GrrCON, and Secure360. You can find him on LinkedIn.
