Can Designers Code? My Real Experience Using AI to Code for a Year

1. Why Should Designers Learn to Code?

I've been doing AI coding (Vibe Coding) for about a year now. Throughout this year, I barely took any breaks. Sometimes I'd spend several hours a day, for days in a row, just to build something.

Many people wonder: What can a non-engineer designer actually do with AI coding?

I want to share from four perspectives:

1. How I got started
2. What I can do after learning
3. What pitfalls I encountered
4. How you can start if you're interested

2. How I Started with AI Coding

I got into AI coding because I attended an AI meetup event. I saw many non-engineer cases there, with people using Google Apps Script to create simple but interesting tools.

Later, I discovered tools like Lovable and Cursor, which made the learning environment much friendlier. The AI meetup even spawned a coding study group, and I attended the first few sessions to learn how engineers approach problem-solving. It was incredibly valuable.

💡 Key mindset: Stay open-minded and observe what others are doing.

This era is great - many things can be approached in simple, low-cost, low-risk ways.

3. What Can You Do? From Imitation to Solving Your Own Problems

Many people think: "AI coding seems powerful, but I have no idea what to use it for."

My approach is: First, see what others have done and follow along.

In the process, you'll discover what makes it powerful and get inspired - realizing that this problem can also be solved with code!

Things I've actually done:

1. Figma Plugin Development 🎨

Figma has its own plugin system. I thought, since AI can write code, it should be able to write Figma plugins too. So I built features I couldn't find or wasn't satisfied with. I now have 5 plugins published on my account.

2. Personal Website 🌐

I've used various website builders, but due to too many customization needs, I never found a satisfactory solution. Hiring someone was too expensive. After learning AI coding, I let it help me realize ideas I couldn't achieve before.

3. Solving Small Work Problems 🛠️

For example, when I wanted to split a large PDF file, all the free online tools had file size limits. I asked AI to quickly write one for me, and it solved the problem immediately.

After this happened repeatedly, I gradually built confidence. Now when I encounter problems, besides my original professional skills, I can also think: Can this problem be solved with code?

4. What Pitfalls Did I Encounter? AI Coding Isn't Omnipotent

There are too many problems you'll encounter with AI coding. Because we're essentially "skipping grades" - jumping past many foundational concepts to start coding - there are definitely tons of pitfalls.

⚠️ Common pitfalls:

• Security issues: API key leaks, security risks
• Cost explosions: Accidentally running up huge API or server bills
• Deployment problems: Not knowing how to go live, sharing localhost links
• Can't see the interface: After coding in Cursor, having no idea what it looks like
• Only the shell: Looks good but only has UI, actual functionality isn't implemented

Many people give up early because of these issues, thinking "AI coding is just hype" or "influencers are just selling courses."

My perspective:

AI coding is truly powerful, but the pitfalls are real too.

This is normal. Because it's inherently challenging and valuable. If it were super easy from the start and worked perfectly, you should question whether what you're trying to solve is too simple, or if you're just assuming there are no problems.

It's like a child getting a gun, not knowing how to use it, randomly shooting and breaking furniture or hurting people, then saying "guns are useless garbage." That logic doesn't work. The real issue is not properly utilizing its true power.

AI coding is the same. And the best part is, after this year, there are more and more AI coding tutorials online. I personally love watching various free and paid content, learning from different people. Since I know little about this field, I can learn new knowledge from many sources and cross-validate them.

5. How to Start? Advice for People at Different Stages

📍 Complete Beginners

If you've never touched programming, start with free resources to understand "what programming can actually help you do" and see if you have such needs.

Try free tools first:

• Google AI Studio: Many features are free to experience
• Lovable: Daily free credits to try building small things
• ChatGPT / Claude / Gemini: If you have a subscription, you can write Apps Script code through Q&A

Read official documentation: Tools like Cursor and Lovable have official user guides. If you don't know where to start, check these first. With AI translation now, language isn't a big barrier anymore.

📍 Already Started, Want to Solve More Complex Problems

When you start solving your own or others' problems and writing more complex code, you'll encounter more issues:

• Why can't I deploy?
• Why does fixing A break B?
• What do I do when AI goes in circles?

At this point, you'll have a clearer understanding of your abilities and what you need to strengthen.

Some people have strong learning abilities and can solve problems through AI and Google on their own. But some might not have strong concepts or even know where their problems lie.

This is when courses can help you fill knowledge gaps in a more structured, logical way.

📍 Like Me, Coded for a Year, Want to Fill in the Basics

After a year of coding, I've indeed built many things - Figma plugins, personal websites, various small tools to solve work problems.

But precisely because I've done more, I clearly know I lack a lot of foundational knowledge.

Many times I can make the program "run," but I don't know why it runs. When I encounter problems, I might work around them differently, but I don't know the root cause. This state of "knowing how but not why" makes me stuck when facing more complex projects.

Things I've found I'm lacking:

• Underlying logic and processes of software development
• Systematic debugging thinking (not just random trying)
• Correct concepts about security and deployment
• How to evaluate whether a requirement is worth doing and what technical solution to use

These can't necessarily be learned naturally by "doing a few more projects." You need someone to help you build the framework and fill in the concepts.

This is why I need course assistance now - not starting from zero, but reorganizing the scattered experience accumulated over the past year with the correct framework.