The Break, The Comeback, The Gap.

It goes back to 2022–2023. Hackathons. Building React, Django, and Firebase projects. It was the pure love of late-night coding, winning competitions, and figuring things out without shortcuts. There was something undeniably real about that time.

Then came the JEE decision. A massive sacrifice. Physics, Chemistry, Math. Repeat. No GitHub commits, no hackathons, just pure grind for 18 months.

Got into IIIT Pune. I opened VS Code for the first time in 18 months. The world had completely changed, Claude, v0, Cursor were everywhere.

The immediate panic hit: did I just waste two years? My identity was caught between who I was before and who I needed to be now. But instead of freezing, I leaned in.

I started right, reading documentation, directing AI with real understanding (around 80%). Then the complexity grew.

The docs stopped. Tutorials stopped. The AI filled every single gap. My actual technical understanding dropped from 80% to maybe 30-40%. But the commit graph stayed solid green. Everything looked perfect from the outside.

I was learning to prompt. Not to create.

The ultimate question surfaced: if AI was removed right now, how far would I get rebuilding the systems I claimed to have built?

Not far.

I had been getting the reward without doing the reps.

The gap between what I could ship with an LLM and what I actually understood beneath the abstraction was massive. I made the decision to close it. In public.

This is what the channel documents. I follow the 70/30 rule. I'm using Stanford's CS229 as the math and theory backbone.

The deliverables are concrete: raw ML models, autonomous AI agents, real deployments, and rigorous evaluations. This is not a highlight reel. It's a transparent build log of doing the work.

"If you can't explain it, you don't own it."