But as artificial intelligence rapidly reshapes the workplace, that promise is starting to look less certain.

A new survey of more than 200 engineering leaders, conducted by tech training nonprofit CodePath and shared exclusively with Fortune, shows entry-level tech hiring is slowing. More than one-third of respondents, 38%, said their company has reduced the number of entry-level hiring over the past year, and nearly 1 in 7 reported pausing Gen Z hiring altogether.

At the same time, 18% said hiring had stayed the same, and 8% reported an increase. Despite the overall slowdown, CodePath CEO Michael Ellison—a Y Combinator alum—argues telling people to avoid tech right now would be a mistake.

“That’s just kind of like taking crazy pills if you end up choosing not to invest in the tools that make you the most powerful—of telling computers what you want them to do in an age where computers are becoming exponentially more powerful,” Ellison told Fortune. “So to me, it’s like saying, ‘don’t learn how to use the internet.’”

Ellison’s argument reflects a broader shift in how computer science fits into the AI economy. As generative AI tools become more capable, understanding how software works—and how to direct, customize, and integrate AI systems—is increasingly seen as a foundational skill rather than a specialized one.

That demand is already showing up in the labor market. AI literacy topped LinkedIn’s list of the skills professionals are prioritizing and companies are hiring for right now. And a Lightcast analysis of more than 1.3 billion job postings in 2024 found roles advertising at least one AI or generative AI skill offered an average of $18,000 more in annual compensation that those that did not.

Notably, the majority of those roles were outside the tech sector. Some 51% of jobs requiring AI skills were in non-tech industries, up from 44% in 2022—a sign coding and AI fluency are becoming relevant far beyond Silicon Valley.

The new secret to landing a tech job

Still, slowing hiring doesn’t mean aspiring technologists should give up. Instead, the CodePath data suggests candidates may need to rethink what they emphasize—and what they leave off—when applying for tech roles.

When asked which signals matter most outside the interview process, engineering leaders indicated proof of real-world skills matter far more than formal credentials. Side projects or portfolios topped the list, cited by 38% of respondents, followed by internship experience (35%), and public code portfolios like GitHub (34%).

Traditional markers of achievement, by contrast, carried far less weight. Just 4% of leaders said credentialing programs were a top influence in hiring decisions, while only 23% cited a candidate degree or academic focus and 17% pointed to school prestige.

The shift away from pedigree suggests employers are seeking evidence candidates can actually do the work. Greater fluency with AI tools and frameworks was the most common skill expectation for early-career hires, followed by faster time to writing production-ready code and the ability to learn new tools or programming languages quickly.

And despite buzz about tech layoffs, job opportunities do still exist. The U.S. federal government, for example, recently announced it would be hiring about 1,000 new engineers, data scientists, and AI specialists. No degrees or work experience is required—and salaries will range from $150,000 to $200,000. Meta has also still been hiring young talent in recent weeks, with job postings for roles such as product software engineers.

Ellison’s advice for those seeking roles is simple: Opportunities are out there as long as you are willing to dig in deeper—and build a portfolio that hiring managers are looking for.

“People are rewarded for being aggressive and for going after what they want,” he said. It’s surprising the opportunities that are hidden in plain sight.”