The Pragmatic Approach
to AI in Frontend
Realistic workflows, common traps,
and how to actually gain from AI in Frontend.
Juan Andrés Núñez
Senior FE
-
Understand
-
Don't do
-
Do
-
How I use AI
4 key parts
Understand
Challenge your attitude
Understand
Stop fearing the machine. AI won’t replace you if you adapt—resisting is the fastest way to make yourself obsolete.
Massive competitive opportunity
Understand
The reset button has been pressed: everyone is starting fresh. This almost never happens. It’s groundbreaking.
Passive acceptance
is not enough
Understand
Either you go all-in… or you’re already out. We are lucky to live this paradigm shift — it’s our Internet-going-mainstream moment.
Don't do
Be aware of FOMO
Don't do
-
The tool isn’t the magic — it’s just a tool.
-
You don’t need to be everywhere (don even try 🙏).
-
Depth beats breadth.
Ignore gurus
Don't do
-
Nobody knows where AI is going — anyone who says otherwise is lying.
-
Most of it is just fucking content marketing, fighting for your attention.
-
If someone sells you a “method” or binary rules → 🏃♂️➡️
Avoid "vibe coding"
Don't do
-
Professional programming is still hard and not everyone can do it. AI doesn’t change that.
-
False sense of achievement — it feels like progress, but it’s cosplay.
-
Like playing doctor: you wear the coat, you sound smart, but all you’re really doing is embarrassing yourself.
Trust, but verify
Don't do
-
Rubber-stamping the model is the easiest way to self sabotage.
-
Blind delegation = perfect candidate for replacement.
-
If you never verify, you’re not a professional — you are stupid.
Do
You plan. AI execute. Always.
Do
- Start with intent. Everything begins with your thinking, not the model’s.
- AI writes the first draft. From composables to unit tests to UI scaffolding.
- No plan = no ownership. If you let the model blindly decide, you deserve the mess that follows.
Augment your workflow. Don't replace yourself.
Do
-
Create artificial resistance. If AI is too embedded, you stop thinking.
-
Keep programming. You may work differently, but you still write code.
-
Cognitive debt is real: if you outsource everything, your skills atrophy. Use it or lose it.
Scratch your own itch.
Do
-
Start from a real need. If it’s not solving your problem, you won't stick long enough.
-
Learn by doing. Build something, break it, debug it, and repeat until it actually works.
-
Share what breaks. The edge cases and the friction are where the real insight lives.
How I use AI
How I use AI
From skepticism
To career shift
-
Fair chance.
-
Write a lot.
-
Share a lot.
-
"Fixing" FE learning.
My journey
How I use AI
My toolset
Auto-completion
Ubiquitous
Augmentation
How I use AI
My workflow
Explore
Plan
Execute
Preserve
Clean
Gather
Contextualize
Document
Operate
Reset
How I use AI
- Plan deeply, execute in small, focused batches.
-
Treat the model as your teammate (a person).
-
Think of AI as a junior dev on steroids.
-
Use AI to get the gist of any codebase in minutes.
-
Leverage images, links, and structured docs.
-
Delegate the boring, double down on what makes you thrive.
-
Invest in what can’t be replaced: your creativity.