# How I Started My AI Journey as a Developer (Without Knowing What GPT Even Means)

As a full-stack developer, I spend hours every day writing code, fixing bugs, testing features, and managing tasks. Somewhere along the way, I realized I was repeating myself—writing similar functions, Googling the same snippets, and setting reminders for routine tasks.

### But

At the same time, I kept hearing things like:

> *“AI will take developer jobs.”*  
> *“If you don’t learn AI now, you’ll be irrelevant tomorrow.”*  
> *“The future belongs to AI—developers who don’t adapt will be left behind.”*

---

![](https://cdn.hashnode.com/res/hashnode/image/upload/v1745147201783/80cd6d7c-bc9b-48f9-b1d9-53a4e870b153.gif align="center")

Honestly, it was overwhelming. I wasn’t a data scientist. I am just a simple developer.

I don’t know what is ***models, LLMs, GenAI***. But I knew how to **build things.** So I decided to stop overthinking and just start exploring.

### Where I Actually Began – With Curiosity, Not Theory

I didn’t know where to start. With a lot of enthusiasm and curiosity, I just started. As any developer, I go to Google, YouTube, and ChatGPT to help me learn AI.

### But

It is not easy… there was so much information, tons of articles, but most of it was packed with heavy terms like transformer, tokenization, Ollama, Hugging Face, fine-tuning RAG, and so on.

My goal wasn’t to build the next ChatGPT; I just wanted to understand the basics. What exactly is GPT? What is OpenAI? Everyone kept saying "AI, AI, AI," but I needed to know the fundamentals first.

> That’s where my journey truly began.

---

## Understanding the Basics

Once I started, I realized that to build anything useful, I first needed to understand some core concepts. It was shocking, but true—I finally got what **tokenization**, **transformers**, and **self-attention** actually mean.

> I used to thinkg ChatGPT and the OpenAI API would give the same answer to simple questions, like “What’s the weather in Delhi?”

## But I was wrong

> **ChatGPT gave me the answer. But the OpenAI API said... nothing useful. Why?**

## Here is the simple reason

🧠 **ChatGPT** is like a smart robot **with internet and extra tools**.  
📦 **OpenAI API** is like the same smart robot, but **without internet or tools**.

In simple terms, **ChatGPT is like a brain with a full working body** — it can think **and** take action. It can search the web, do calculations, read files — just like a body helping the brain get things done.

But the **OpenAI API is just the brain**.  
It’s super smart and can think really well, but it **can’t do anything by itself**.  
If you want it to act — like search the web or get the weather — you have to **build the body around it**..

## Wrapping Up

So that’s how my AI journey started with small steps, a curious mind, and a strong developer instinct.

From feeling overwhelmed by AI buzzwords to understanding how tools like ChatGPT actually work

> And trust me, if I can do it, **you definitely can too.**

In the next post, I’ll break down one of the most basic but powerful concepts in AI:

> **What is tokenization?**  
> (Hint: It’s like teaching a computer how to read!)