What is AI actually good at?
Strategic Thinking? Problem Solving? Reasoning?…. Nope!
Let me explain. This will be interesting. So stay with me till the end.
As a developer, you spend around 8 to 12 hours in the office. And if you are a nerd with an inability to say “NO“, you will probably be spending 16 hours in the office.
But what tires you the most? Writing Code? Meetings? Coffee Breaks?…. Not at all.
You spend 80% of the time coding, and the remaining 20% is actually the thing that makes you tired at the end. It’s your effort in coming up with various approaches to solve a specific problem at hand, assuming you actually put in the effort.
In other words, STRATEGIC THINKING.
So, 80% Coding & 20% Strategic Thinking.
Now, in this AI era, what do companies actually expect from you?
Modern LLMs like ChatGPT and Claude are trained on massive amounts of content — Stack Overflow, blogs, research papers, thousands of GitHub repos, and more. Because of this, they’re extremely good at recognizing patterns. That’s why their answers usually make sense.
So we can safely say: AI is really good at implementation.
It can understand your codebase reasonably well and generate acceptable solutions.
In many ways, it automates a big chunk of the work — writing code.
But what it still struggles with — and likely will for a few more years — is thinking in terms of:
- Scalability
- Reliability
- Reducing redundancy
- Choosing the right design patterns
- Managing trade-offs
That’s where you come in.
The strategic part: analyzing impact, exploring different approaches, designing systems and architectures, researching, and making decisions.
AI can help with the 80% - The Implementation
You focus on the thinking - the 20%
That’s the real value now.
And sometimes, I even turn off agentic mode and code by myself.
Honestly… it feels good.
Now, here’s where you might think you’re screwed.
Companies have realized this.
They strongly believe that AI writes better code and that it writes faster.
Result? Tight deadlines, more features shipped in a month, fewer bugs & of course, fewer employees.
And we can’t blame them.
Their competitors are employing AI to speed up the work and in their products. So they have no choice but to do the same — More & More AI.
They just want you to write code faster and ship it, whether you use AI or not. If AI takes care of writing code — The implementation, then what are you paid to do?
Simple, it’s STRATEGIC THINKING
Your ideas, your research, your learnings — That’s what matters.
If you are being hired, that’s what you are expected to do.
Makes sense?
But you know what? You don’t have to worry too much.
If you think about it, it has always been this way.
You were never hired just to write code. You were hired to solve problems — coding is simply the tool.
That’s exactly why interviews focus on DSA and System Design. They don’t really care which language you use. What they’re evaluating is your thinking — how you approach a problem, how you reason about it, and how strong your fundamentals are.
If that part is solid, you’re good.
So the point is simple: in this AI era, companies are looking for problem solvers — people with deep knowledge in a domain who also know how to leverage AI to get things done.
In my opinion, that’s what you should aim for.
Expertise.
Build that, and you’ll be just fine.