The Driver, Not The Car

There's a moment in every Formula 1 race where the car hits a speed that no human should be comfortable with. The engine is screaming, the tyres are at their limit, and the margin between a podium finish and a catastrophic crash is measured in milliseconds. The car doesn't make that call. The driver does.

I think about AI the same way.

The machine is extraordinary. That's not the hard part.

We're living through an inflection point in software engineering. The tools we have access to right now — large language models, code generation, image synthesis, autonomous agents — are genuinely extraordinary. A year ago, half of what I build daily wasn't possible. The machine has gotten very, very fast.

But fast doesn't mean safe. And capable doesn't mean correct.

I've spent the last year deep in AI Engineering — not just using the tools, but building systems around them. Delivery pipelines where AI generates code, creates imagery, writes content, and deploys infrastructure. The kind of end-to-end automation that sounds like a pitch deck but actually works in production.

And here's what I've learned: the technology isn't the bottleneck. The judgement is.

Without a skilled driver, a fast car is just an expensive crash

AI can hallucinate. It can ship bad code with absolute confidence. It can automate the wrong thing at scale and do it faster than any human could clean up the mess. I've seen it happen — not in hypothetical scenarios, but in real production systems where someone let the machine run without anyone at the wheel.

The companies that are winning with AI right now aren't the ones with the most advanced models. They're the ones with engineers who understand when to push and when to pull back. Who know the difference between letting AI accelerate a workflow and letting it drive off a cliff.

That's the skill. That's what's actually scarce in the market right now.

I'm a driver for hire

My name is Chris Peters. I'm the founder of Menoko, a platform engineering company in Sydney, and I've spent the last two decades building software systems — from .NET enterprise platforms to modern cloud-native architectures on Azure.

Right now, my work sits at the intersection of platform engineering and AI. I build the systems that let AI do useful work safely: content pipelines, image generation services, automated deployment workflows, and the shared infrastructure that ties it all together. Everything I build, I dogfood — including this site. The images you see here were generated by Menoko's AI service. The content pipeline that produced this article is AI-assisted from journal to publication.

I don't think AI is the Terminator. I also don't think it's "just a tool" — that undersells what it can do. I think it's a Formula 1 car: extraordinary, powerful, and absolutely capable of causing damage if there isn't a skilled human making the decisions.

If your organisation is trying to figure out how to put AI to work — actually to work, in production, generating value — you need someone who knows the machine. Someone who's been in the seat, who's felt it try to go off track, and who knows how to bring it back.

That's what I do.