\

Code Like a Driver, Prompt Like a Cowboy

  • Jordi Torras
  • Blog

The Dual Lives of Modern Engineers

After living in Dallas TX, for more than a year, I’ve had the chance to embrace some of the local culture—most notably, horse riding. There’s nothing quite like saddling up on a fine Texas horse and hitting the trails. However, if there’s one thing I’ve learned, it’s that horses, no matter how well-trained, have a mind of their own. You might pull the reins and expect them to follow your lead, but sometimes they decide to go their own way. It’s an experience that’s both exhilarating and humbling, reminding you that you’re not always in full control.

One day, while I was driving back from an equestrian center to my favorite BBQ restaurant, I realized something. When I turned the steering wheel, the car turned. When I pressed the brake, it slowed down. The car did exactly what I wanted—no surprises, no attitude. It’s a straightforward experience: what you do is what you get.

This contrast is a lot like the difference between software engineering and prompt engineering. Writing code is like driving a car: it’s structured, predictable, and things go as planned. But prompt engineering? That’s more like riding that stubborn horse. You give the AI a prompt, and who knows where it’s going to take you. You’ve got to adapt, nudge it back on track, and sometimes just hang on for the ride.

Control vs. Collaboration: Who’s Really in Charge?

When you’re coding, you’re the one in charge. Every piece of code does exactly what you tell it to do. It’s like driving a car: you steer, you brake, you accelerate—all according to your plan.

But prompt engineering? That’s a whole different ballgame. It’s more like a collaboration with the Language Model. You give it direction, but the AI might decide to interpret your input in its own way. It’s less about control and more about working together, like a cowboy trying to understand and respect the horse to get the best ride.

AI as a Co-Driver: When the Horse Can Drive the Car

Interestingly, AI is not just a horse; it’s a horse that can actually drive the car for you! AI has become incredibly proficient at writing code. It can generate code snippets, help you navigate libraries and parameters, and even translate between programming languages. But just like you wouldn’t trust a car to drive itself without checking the GPS, you shouldn’t assume that the AI-generated code is perfect. Always review what the AI has produced, ensure you understand the code, and verify that it works as intended. AI might be a great coder, but it’s your job to make sure that code is bug-free and fit for purpose.

Using AI to write code is also a fantastic way to learn. It’s like having a tutor that guides you through complex programming tasks, helping you understand new concepts in real-time. AI may be stubborn sometimes, but it’s an amazing horse that can take you on a learning journey like no other.

Wrangling the Wild Horse, Mastering the Open Road

In the end, the dual lives of modern engineers—driving the car of software development and wrangling the horse of AI prompt engineering—require a blend of precision and adaptability. Mastery in both arenas doesn't just mean controlling the outcome; it means knowing when to guide with a firm hand and when to collaborate with a mind of its own. As AI continues to evolve, engineers will need to sharpen their skills in both disciplines, harnessing the power of code while learning to ride alongside AI's unpredictability.

So, whether you’re behind the wheel or in the saddle, remember: your success depends not just on where you want to go, but on how well you understand and engage with the tools at your disposal. Embrace the ride, enjoy the journey, and never forget to keep learning along the way.

Make AI work for you

Empower your vision with our expertise. Me and my team specialize in turning concepts into reality, delivering tailored solutions that redefine what's possible. Let's unlock the full potential of AI. Effectively.

Contact us