🔮 Now that you’ve had a recent, and potentially even your first experience using AI, let’s break it down. How does this thing work? Is it a crystal ball that knows my future?
Ha! Thankfully not, let’s jump in!
But first, I promise the conclusion gives you the one thing that matters most for actually using AI in your business, it also has my call to action for you. Now, a question:
Do you remember playing Mad Libs as a kid? A sheet of paper with some text, you’d fill in the blanks: “The [adjective] cat [verb] over the [noun]”. Sound familiar?
If you answered yes to that question, then congratulations, you understand how ChatGPT works!
Quick note: I've planned these Foundation editions to build on each other. If you're jumping in here, you might want to catch up on the previous issues first. It'll make way more sense! here.
Also, if you find this newsletter helpful, share it with someone in your network who's curious about AI. The more people I can help, the better content I can create for all of us! 💪
Let me explain. ChatGPT is a Large Language Model, or LLM for short. This may or may not be a new term for you, don’t worry if it sounds technical, it’s simpler than you think! Note: pretty much everyone says LLM
Here's what LLMs do: they predict the next word. That's it. Over and over again.
“What? You lost me Cam.” I know, I know, it makes no sense. I still feel that way at times. Let’s break down how this is possible using a series of analogies.
1. The Language Learning:
First, imagine a baby learning to talk. They don’t start with grammar rules or long lists of vocabulary. They start by listening to conversations.
“More milk, please”
“Where are your shoes honey?”
“Good morning, my sweetheart”
After hearing enough examples, patterns emerge. The baby notices that “please” often comes after requests, “what” usually starts a question, and certain words flow naturally together. Eventually, the baby learns to speak.
That is exactly how AI learns. It’s pattern recognition at scale, significant scale.
2. The Training Process:
Now imagine the same baby, but instead of hearing a few thousand conversations from friends and family, they read every book ever written…. every website ever published… and every article ever printed.
We're talking billions of examples of how humans use language.
I still can't wrap my head around that scale 🙃. And yes, this costs companies like OpenAI a LOT of money and computing power to do.
The LLM studies these patterns like a student cramming for the biggest test:
How do people typically start emails?
Which words usually come after "The weather today is..."?
How do recipe instructions flow from step to step?
And on and on the training goes until the Large Language Model is extremely good at, you guessed it, Language.
3. The Practice Phase:
Just like a piano student practices thousands of times until their fingers know where to go automatically, the LLM practices predicting the next word in sentences millions of times.
It starts terrible, like a toddler saying "Me want cookie now!" But gradually, through endless practice, it learns to sound more natural and sophisticated.
Think of the practice phase as the feedback and corrections provided to the training of the LLM to improve it. It iterates and improves with time.
Most people call this whole process "training the model."
Let’s put it all together
After all this training, something amazing happens. The AI becomes so good at patterns that when you start a thought, it can finish not just the sentence, but entire paragraphs.
It's like finishing your best friend's sentences, but it happens every time, lightning fast, and sounds surprisingly… human.
Here's the simple version: You type a message, ChatGPT uses everything it learned, it predicts what you need to hear back.
To say this is oversimplifying it is an understatement 😂 but what’s important is understanding the concept of how the LLM functions. Through trial and error (training), it becomes the ultimate prediction machine.
Conclusion (The MOST Important)
Now that you know AI is just playing an elaborate prediction game, you get to help it play. The better the clues you give it, the better its guesses become.
Think of yourself as a car passenger providing instructions to the driver; the more specific your instruction, the less chance of confusion (my wife would roll her eyes if she read this).
In the coming weeks, you’ll read the word 'context' so much it'll make your eyes roll.
Here's what I want you to remember: The magic isn't in the machine. It's in what you tell it. The context it operates with.
This is a foundational principle you will need to remember throughout all types of AI tools. Always consider the context the AI is working with.
If you can master this concept and remember it consistently, you’ll be among the top users of AI today with that alone. That’s how you’ll find leverage for your business. You need to keep stepping forward and learning.
Take the next steps and start trying to integrate AI use into your everyday life. Set aside 10-20 minutes to experiment.
You got this,
Cam
Quote of the Week:
“The public praises people for what they practice in private.” - David Senra, host of my favorite podcast, Founders.
It’s quite similar to Tony Robbins’: “You’re rewarded in public for what you do in private”, but David’s resonates more with me for some reason. Which do you like more?
P.S. You’ll notice I said ChatGPT is an LLM, but it is not the only LLM. It is by far the most used and widespread. Take a moment and look up Google’s Gemini and Anthropic’s Claude LLMs.