# The Skill Level Of Young Developers Is Dropping. And Barely Anyone Wants To Admit It. > Source: > Published: 2026-07-09 20:39:32+00:00 Yesterday, during a class on advanced machine learning models, my professor said something that stuck with me. *"The skill level of young developers is declining year after year."* Not because they're less intelligent. Not because the curriculum is worse. But because they no longer take the time to think things through by themselves before handing their tasks off to AI. He added something even more concrete: **he sees this decline firsthand in the new interns he mentors.** It's not an abstract trend for him it's something he watches happen in real time, intern after intern. At first, it sounded like a slightly jaded professor's remark, the kind of comment every generation hears ("kids these days don't know how to do anything anymore"). But the more I think about it, the more I believe he's right, and that the problem runs deeper than we'd like to admit. There's something essential about learning to code that we're losing: **the struggle itself.** Spending two hours stuck on a dumb bug. Reading the same stack trace over and over. Searching, testing, failing, trying again. It's frustrating, sure. But that exact process is what builds real competence. It's by hitting the wall that you understand *why* a solution works, not just *that* it works. Today, a lot of young developers never hit that wall anymore. They copy an error, paste it into an AI chat, grab a solution, and move on. The code works. But the understanding never actually happened. I'm not saying AI is bad for learning to code that would be absurd, and a bit hypocritical since I use it myself every single day. The problem isn't the tool. The problem is using it to *avoid* effort instead of using it to *support* effort. There's a huge difference between: The first approach is a learning accelerator. The second is technical debt you take out against yourself and you'll pay it back later, with interest. Here's where I want to go with this, and it's the part that worries me the most. If this trend continues, the average skill level of developers will keep dropping. Not dramatically or overnight, but slowly, generation after generation. More and more people will know how to "make code work," but fewer and fewer will actually understand what they're writing, be able to debug a complex problem without assistance, or design a solid architecture from scratch. Meanwhile, AI itself will keep improving, taking up more space, automating more tasks. That creates a scissor effect: more people relying entirely on AI, fewer people capable of supervising it, correcting it, or solving the problems it can't solve on its own. The result: in 10 or 15 years, the people who kept real technical depth — who kept understanding the fundamentals instead of systematically outsourcing their thinking will be **extremely rare**. And a rare skill, in a market that will always need it, comes at a very high price. I think we're heading toward a split: That second group will end up grabbing pretty much every interesting opportunity and the salaries that come with them. If this reasoning holds up, the real question isn't "should we use AI or not." The question is: **are we using it to learn faster, or to stop learning altogether?** I think the best long-term strategy for any developer starting out today is to keep imposing difficulty on themselves on purpose. Solve problems without AI from time to time. Understand the "why" before asking for the "how." Treat AI like a demanding mentor, not a permanent shortcut. Because in a few years, "knowing how to use AI" won't be what sets anyone apart. Everyone will know how to do that. What will matter is: **who still has real technical skill behind it.** And those people will be few. Which is exactly what will make them valuable. **Do you notice this trend around you, in your classes, or with people you mentor? Curious to hear your take in the comments.** *This is part of a series where I share my honest thoughts on AI, learning, and building in tech.* *Follow along on Facebook*