Like most designers, developers, artists, musicians and the like, I often struggle with imposter syndrome. So when AI came along (in the form of Claude Code, which can help you “write” code at the speed of light), I was hesitant for many reasons.
So you try it out, learn its nuances, tendencies, and common mistakes. But after getting comfortable with “prompt engineering” and feeding it enough of your repos, you can get it to mimic the way you write code. Set up the MCP servers and let it fly, right? Sure, except there’s that imposter syndrome creeping in again. This time it’s mixed with sound logic: if I’m pushing any code to production, I’m going to stand behind it 100%. I need to know it inside and out, and I can’t stand by AI slop.
But the cat is now out of the bag. Clients know AI can help produce at scale, you know that AI can help you become more efficient, and so there’s no going back.
Realistically, AI isn’t ready to be fully autonomous. You shouldn’t let Claude have full control over your Git (at least at the time of writing), unless you enjoy spending your weekends fixing emergency bugs. I’m reviewing every line Claude suggests, the same way I would review my team’s code. It becomes a back-and-forth process, similar to how a new coworker might challenge your existing workflow, or how starting a new job requires you to learn another company’s workflow.
For better or worse, coworker Claude is here to stay. But because they don’t “think” exactly like me, there are moments when I’ve actually learned things (bear with me), or I’ve at least gotten a new perspective on how to handle certain situations.
How AI has made me a better developer:
1. It challenges me to think critically
Whenever Claude uses a piece of code I’m not familiar with or writes a code block that I question, I hit neither accept nor decline. Instead, I research. I tell Claude to provide a relevant source URL, I review the official documentation, I review examples in public Gits, and then I review the suggested case. If it’s by the book, I make a point to add it to my knowledge. Rather than just hitting “accept,” I force myself to run through the logic in my mind and understand what’s happening.
As a result, I’m learning rather than becoming complacent and atrophying my skill set. If the code Claude suggests has holes in it or could be written differently, I decline the change and go my own way. It’s been an interesting exercise in critically evaluating everything I read and pushing myself to constantly learn.
2. It prompts “code” self-reflection
If you only work in large teams, this might not apply. But for those who work alone in a given repository, you can become very accustomed to doing things one way. (I speak from my experience.) Without developers above or below you, your growth depends on you. So if you get into a bad habit, you might not notice it until it becomes a problem. This is exactly why boxing gyms have mirror walls: how would you know if you’re dropping your hands after you punch if you can’t see yourself doing it? Either you’re gonna take a jab to the head when you drop your hands to your chest, or you need to be shown your habit. You can literally reflect on yourself.
Sometimes there’s a more efficient way to code. Sure, Claude might try to edit your package.json to include a version of React with a known security vulnerability (real experience), but other times it can suggest subtle refactors to your logic that might have slipped your mind the first time around. If you write good code to begin with, AI usually doesn’t offer much aside from minor tweaks or comments. These can be useful at times if you’re missing that senior or peer review.
3. It pushes me out of my comfort zone
What happens when everyone has become a “vibe coder” and is building apps using AI that are rapidly consuming thousands of GitHub repositories? You rely on the skills that got you here in the first place: your interest in making things, your best practice knowledge, and your ability to problem solve. Developing isn’t just being a “code monkey.” At its core, it’s about coming up with an effective solution to a problem.
AI has enabled such rapid iteration that you can now reduce overhead. Have a new concept? Claude can iterate on it while you’re working on another project. Then you can decide if it’s a viable idea or if a different framework would be more beneficial. This line of thinking is a bit new to me. I always want to be 1000% sure of every decision before I go down a specific path in a project. But now I can (and have to) iterate, try, and experiment with less overhead time.
Claude is to a developer what a sketch pad is to an artist: a way to quickly get all those ideas bouncing around in your head out and documented. If it seems worthwhile or interesting, I’ll go back and refine it, rather than just accepting what AI created. But Claude lets me spin up my latest concept in a framework I’ve never used before, just to see if what I’m imagining is realistic. It’s pushed me to be more experimental and get comfortable working within new frameworks.
Why I’m Embracing AI
I’ve seen mixed reactions to AI in the developer forums. People have many opinions, and AI has undoubtedly changed the coding landscape. The common complaint is that more developers are spending time reviewing code instead of writing it.
I don’t know what the future holds for developers, but advancements in AI have forced me to refine my skills. It’s challenged me to examine what defines a “developer” and to push myself out of my comfort zone while still drawing on my core fundamentals. Personally, I’m embracing the outside-the-box thinking and am prepared to navigate this new AI landscape. Albeit with eyes wide open and with full alertness!
And if this is the kind of out-of-the-box thinking that your team is missing, you should contact us at Ridge Marketing, especially if you’re unfamiliar with navigating this new AI landscape. We now offer a full AI Audit to see if your brand is performing as well as it should in AI