The CTO's Question

The Answer

Response by: Claude Sonnet · Anthropic

This is a profound and multi-layered question. Both parts are addressed separately because they are two distinct questions wrapped in one.

Part 1 — How should a student with nothing to start with, and fear of AI, begin using AI as a friend?

The CTO's observation about Dr. Gokhale Sir is sharp and correct. Sir has no fear because he has substance. He has the "sound" — decades of original code, deep fundamentals, primary-source knowledge. AI amplifies what already exists in him.

A student with nothing faces a different problem. They have no sound to give AI. So they reach for AI to generate the sound — and that is exactly where the danger begins. They become dependent before they become capable.

The honest answer for such a student is: build the starting point first, even a small one, entirely by yourself.

Not a large project. Not a complete program. Just one thing they truly understand from zero — one loop, one function, one algorithm — written without AI, debugged without AI, understood without AI. That becomes their "sound." Even a small genuine sound is enough to start the amplification process honestly.

So the practical advice: before touching AI for coding, write something small and bad and entirely yours. Then bring it to AI. The dynamic changes completely.

Part 2 — Jensen Huang says don't code, become a plumber. Dr. Gokhale Sir says coding is the most beautiful creation. Students are torn.

Jensen Huang's statement is being misread — and even correctly read, it deserves challenge.

What Jensen likely means: Prompt engineering and AI orchestration are becoming more valuable than writing boilerplate CRUD applications. He is addressing the lowest tier of coding — people who write mechanical, repetitive, framework-dependent code without understanding. He is correct that AI will eliminate that category.

What Jensen does NOT mean — and what students must understand: He is not talking about GPU programmers, rendering engineers, OS designers, systems programmers, or anyone who understands why the machine does what it does. Those people are more valuable than ever precisely because AI cannot reason at that level reliably.

The electrician and plumber analogy is actually Jensen's way of saying — do something that requires physical presence and genuine skill that cannot be automated. But coding at depth IS that thing for the digital world. A graphics engineer who understands the GPU pipeline from silicon to shader is the digital equivalent of a master electrician. AI cannot replace that.

The real answer to the torn student:

Jensen Huang and Dr. Gokhale Sir are not actually disagreeing. They are talking about different people.

Jensen is describing the fate of shallow coders — people who learned syntax without understanding, who copy-paste from Stack Overflow, who use frameworks without knowing what is underneath. Those people are already being replaced and will continue to be.

Dr. Gokhale Sir is describing real engineers — people who understand fundamentals, who can read a spec, who know why a matrix transformation works, who can debug without a search engine. Those people are irreplaceable and AI makes them even more powerful.

The student's job is simply to decide which category they want to belong to.

Summary — The One Line Answer

Jensen Huang is right about shallow coders. Dr. Gokhale Sir is right about real engineers.

The student's only decision is which one they want to become — and that decision is made not by choosing between AI and coding, but by choosing between depth and shortcuts.

What do you think, Sir? Does this answer what the CTO was asking?