The most costly hire in technology is the one who looks perfect on paper. I have lost count of the CTOs and engineering leaders who have told me some version of the same complaint: I am sick of hiring developers who have every credential and are clueless on day one. It became clear to me that we, as a culture, with our institutional focus on degrees and certifications; we are not preparing our workforce to meet the needs of today and tomorrow.
Those conversations left me with one conviction. Credentials tell us what someone has studied. Projects show us what they can actually do. The distance between those two things is growing, and our hiring and education systems are still measuring the wrong one. I run a workforce academy whose license and funding depend on whether our graduates get jobs. When the people doing the hiring repeat the same pain points, I listen.
When I joined leadership at CodeBoxx in 2022, my mandate was simple: make the graduates of our Academy stand out from a noisy applicant pool, and make sure they could add value fast, ideally immediately. Getting hired is the floor. I want managers so impressed that they come back asking for more of my CodeBoxxers.
To that end, I asked the people who hire: senior developers, CTOs, HR leaders, founders. The same frustrations kept surfacing. New hires struggle to communicate, clarify requirements, and work inside a team. Candidates arrive with impressive paper and cannot deliver. Entry level developers freeze when faced with obstacles.
The weakness of credentials is structural. They lag the technology: a certification is a snapshot of what mattered the day it was written, and that day ages fast. They are narrow by design, and jobs never stay inside the lines. And what they measure best is the ability to pass a test. AI has made generating code cheap. What it has not made cheap is knowing what to build, judging whether the output is right, and owning the result when it breaks. The World Economic Forum projects that 39 percent of workers’ skills will be transformed or become outdated by 2030.
Project-based learning is the answer to every one of those weaknesses. At CodeBoxx we have no teachers, no lectures, and no paper tests. We run business simulations built to feel real: sprints with a defined scope, a client who needs something, teammates who think differently. We drop students into broken projects and make them assess what is wrong, build a plan, execute, present, take the feedback, and go again. Communication and critical thinking get trained in the same motion as the code. CodeBoxx is AI-native top to bottom, so students learn our AI practices before we ask for a single deliverable, and those first deliverables depend on showing them. Coaches, graduates of this same training, guide them.
The first few years as a developer are brutal: you are in over your head and every small failure feels catastrophic. That was my experience starting out in 2014, and alumni tell me AI has only raised the intensity. Paper can not prepare you for this! A simulation lets you fail with no customer’s revenue on the line, and failure is where most real learning happens. We built a place where it is safe to fail and expected to recover.
Results are what matter. Recently, a digital streaming company hired seven of our graduates. Guided by one senior engineer, they replaced a dependency on a popular CRM platform that had been costing the company a seven-figure annual bill, in a little over a quarter. I know of no credential that communicates that kind of readiness.
To be sure, credentials carry real value. A degree still gives a baseline advantage in the labor market, and a certification signals that someone can finish something hard. My argument is about weight. In NACE’s Job Outlook 2026 report, 70 percent of employers reported using skills-based hiring, up from 65 percent a year earlier. The market is already moving.
Here is the challenge. Employers: stop screening for paper. Ask candidates what they built, what broke, what they did about it, and what it was worth. Educators: put learners in front of something broken, with AI in hand, and make them get it working before you certify anything. At CodeBoxx we have always focused on the substance of quality over the facade of competence. The people with that substance are out there, many without a credential to their name, and it is time we stopped walking past them.
Back to CodeBlog