LLMs for Engineers - Part 1
This is Part 1 of a series where I break down how AI actually works in simple terms,
You open ChatGPT , type something and hit enter.
And within seconds…you get a clean, confident answer.
It feels like you’re talking to something that understands you.
Almost like there’s a brain behind the screen.
But here’s the reality:
There is no thinking happening inside ChatGPT.
No understanding.
No awareness.
No intelligence in the way we humans experience it.
So what exactly is it?
The illusion we all fall for
The first time I used ChatGPT, I thought:
“This thing knows stuff.”
But that’s not true.
What it’s doing is much simpler and honestly, much more intersting way !!!
The simplest way to understand ChatGPT
Let’s break it down in one line:
ChatGPT is a very advanced next-word prediction system.
That’s it.
Let’s make it real
Imagine you type:
“Network automation is…”
Your brain instantly starts predicting what comes next:
important
growing
the future
You don’t consciously think about it — it just happens.
OK let’s take another example :
“the students opened their…”
Your brain instantly starts predicting what comes next:
books
laptops
exams
minds
Yes, you got the right answer :) i.e books . You are too Like LLM but with lot intellegence and experience.
👉 ChatGPT does the same thing.
But instead of your experience…
It uses patterns learned from a massive amount of internet text .
So what’s actually happening?
Every time you type something, ChatGPT is:
Reading your input
Breaking it into small chunks (we’ll cover this later)
Predicting the next most likely word
Then repeating that process… again and again
Until it forms a full response.
So instead of “thinking”:
👉 It is continuously guessing what comes next
Why does it feel so intelligent then?
Because of the scale.
It has seen:
blogs
documentation
coding examples
Q&A discussions
random internet conversations
So when you ask a question…
It doesn’t know the answer.
👉 It generates something that looks like the correct answer
Based on patterns it has seen before.
Engineer’s way to think about this ⚙️
If you’re from a tech background, this will click instantly:
CLI autocomplete → suggests commands
Google → suggests queries
Excel → predicts patterns
Now imagine:
👉 All of that
👉 Trained on internet-scale data
👉 Able to generate full sentences
That’s ChatGPT.
So what are you really talking to?
Not a human.
Not a brain.
👉 You’re talking to a probability engine trained on text
A system that has learned:
“Given this input… what usually comes next?”
Why this understanding matters
This one insight changes everything:
You stop expecting perfect truth
You understand why it sometimes gives wrong answers
You start writing better prompts
You use it like a tool — not magic
The one takeaway
If you remember just one thing, let it be this:
ChatGPT is not thinking. It is predicting.
What’s coming next
Now that you understand what it is…
The next obvious question is:
👉 Where did it learn all this from?
Because that’s where things get really interesting.
If this helped you see ChatGPT differently, you’re already ahead of most people using AI today.
Let’s go deeper in the next post
Smiles :)
Anurudh