According to Business Insider, in 2025, four tech professionals with aspirational AI jobs all shared the exact same, remarkably simple piece of career advice for breaking into the field. Their unanimous tip was to stop trying to learn AI from a textbook or lecture hall and instead “get your hands dirty” by gaining real-world experience with the technology. The advice came from interviews covering topics from whether a Ph.D. is necessary to how to earn big as an AI contractor. The core idea is to treat AI like a new toy and tinker with it directly, an approach several described as essential for getting your foot in the door during the ongoing AI revolution disrupting white-collar jobs.
Why Theory Isn’t Enough
Here’s the thing: this advice is so common it sounds cliché. But there’s a reason for that. AI, especially the practical application of it, is moving way too fast for any formal curriculum to keep up. By the time a university course is approved, printed, and taught, the tools and best practices have already evolved. So what these pros are really saying is that your ability to experiment, fail, and figure things out on the fly is now the most valuable credential you can have. It’s not about what you know; it’s about proving you can learn and adapt.
What “Getting Dirty” Actually Means
But “get your hands dirty” is pretty vague, right? In practice, it means building something—anything. It means using an API from OpenAI or Anthropic to make a dumb chatbot. It means trying to fine-tune an open-source model on a dataset you care about, even if it’s just your own notes. It means automating some tedious part of your current job with a Python script, even if it’s imperfect. The goal isn’t to build the next ChatGPT. It’s to encounter real problems—data formatting, cost management, prompt weirdness—that you simply cannot read about in a theoretical way. That’s the dirt. And that’s what hiring managers want to hear about.
The New Path to an AI Job
This fundamentally changes the traditional career ladder. The implication is that you don’t necessarily need that fancy advanced degree as a prerequisite. What you need is a portfolio of experiments and a compelling story about what you learned from them. This is creating a new, more meritocratic (and chaotic) path into the industry. I think we’ll see more people “breaking in” from adjacent roles—the marketer who built a lead-classification model, the financial analyst who automated her reports. Their hands-on projects become their resume. So the real question is: what are you going to build this weekend?
