What the heck is AI?

A beginner’s guide to what it is, how it works, where it came from, and why it matters.

May 01, 2025

By 2023, AI had already exploded into the mainstream.

Suddenly, every company was claiming they used AI, every startup pitch had “AI-powered” somewhere in the deck, and investor money seemed to have one clear destination: anything Artificial Intelligence.

At the time, I was already using AI too. A lot.

I was constantly amazed by what tools like ChatGPT could already do.

I even showed the account managers team I used to coordinate at the time how to use it to hit sales goals faster, brainstorm ideas, and handle tough objections more smoothly.

It wasn’t perfect, but it did feel like having an extra brain.

And in a high-pressure environment, that felt huge.

But even though I was already using GenAI everyday, if you had asked me back then to explain what it actually was and how it worked, I wouldn’t have been able to give you an answer.

That gap really caught up with me the weekend before I was supposed to speak at a student event about my startup, at the beginning of 2024.

I had already said yes to speaking at the event.

When I agreed, I thought I’d talk about my founder journey.

Later, I found out the whole theme of the event was around… you guessed it, AI.

And there I was, someone who used AI every day, but couldn’t explain how it worked, where it came from, or what made it different from regular software.

We weren’t even using AI in the startup I was building.

So the thought of suddenly having to integrate it into my workshop for the students made me panic.

I almost backed out.

But after talking with the event organizers and learning that I didn’t actually have to cover the AI part, because other speakers would handle it, I was relieved.

Still, I couldn’t walk into that room without at least understanding what the heck AI really is.

Luckily for me, I was spending that weekend working on our app with my technical co-founder.

I turned to him with a million questions, and he pointed me toward the resources that helped me finally start deep-diving for real.

And for the first time, I leaned into it.

That weekend flipped a switch in me.

I realized that if I wanted to stay relevant as a founder, a creator, and someone building for the future, I needed to truly understand the world of AI.

Not just use it.

Not just admire it.

Understand it.

Since then, I’ve been on a journey to make sense of it.

One that started with a lot of confusion and slowly turned into clarity and a lot of curiosity.

And if you’ve ever felt overwhelmed by AI, or just nodded along when people talked about it because it felt too complicated, stick around.

Today, I’m going to break it down, retracing the same rabbit holes and questions I had when I first started digging in.

We’ll talk about:

What the heck is AI
How does AI actually work
Where AI came from and how it evolved
Is AI close to replacing people
Why does understanding AI matter now, not later
Should we be excited or worried about AI

Subscribe to AI blew my mind

So… What actually is AI and how does it work?

Artificial Intelligence means machines that can mimic some kind of human thinking.

Simply put, AI is software that’s trained to notice patterns, make predictions, and get better over time by learning from examples, instead of following step-by-step instructions like traditional algorithms do.

Classic software runs on rules: if this happens, do that.

But AI doesn’t need to be told exactly what to do.

You give it enough examples, and it starts to figure things out on its own.

Here’s a simple example:

Let’s say you show a computer thousands of pictures labeled “cat” or “dog”.

You don’t explain what makes a cat a cat, you just give it a lot of labeled images. Over time, the system starts to learn the differences by picking up on tiny patterns.

It doesn’t actually “know” what a cat is, it’s just gotten incredibly good at guessing.

This kind of pattern-spotting is what powers things like Netflix recommending your next binge, Google Maps predicting traffic, Gmail catching spam emails, or your phone unlocking with your face.

All of that is AI.

More specifically, narrow AI, trained to do one job really well.

But most mainstream talk around AI today, especially in the last year or two, isn’t about spam filters or route optimization.

It’s about generative AI, the kind that can write, draw, plan, design, ideate, and even carry on a conversation.

So then, what’s generative AI?

Generative AI is a type of AI that doesn’t just analyze or predict, it creates.

It can write emails, draft essays, generate code, design visuals, even mimic voices and videos.

And it does all this using a kind of model called a Large Language Model, or LLM.

LLMs like ChatGPT are trained on massive amounts of text — books, websites, codebases, Reddit threads, Wikipedia articles, you name it.

They don’t “think” like humans.

What they’re doing is predicting, word by word, pixel by pixel, what’s most likely to come next in response to your prompt.

It’s not based on real understanding, just extremely advanced guessing, built on seeing billions of examples.

A quick technical peek (you don’t need to memorize this, but it’s helpful to know):

Training = feeding the model tons of examples and correcting it when it gets things wrong
Inference = using that trained model to make predictions in real time, like answering your prompt or finishing a sentence

I’ll be honest, I still didn’t fully get how LLMs work until I went down a rabbit hole of explainer videos.

If you’re starting from scratch, this short one (under 8 minutes) by 3Blue1Brown is a great place to begin:
Click here to watch it.

You can explore more of their videos after that, and if they start to feel too technical, don’t worry. There’s a ton of other beginner-friendly content out there.

Already enjoying this post? It’s public, so feel free to share it with a friend who’s curious about AI too.

Who invented AI and how did we get here?

Let's rewind.

1950: Alan Turing asks “Can machines think?”

British mathematician and computer scientist Alan Turing poses a bold question and lays the theoretical foundation for AI. He develops the idea of a “universal machine” (now known as the Turing Machine) and proposes the famous Turing Test, a way to assess a machine’s ability to mimic human conversation.

He died in 1954, long before his ideas became real, but his legacy shapes the entire field.

1956: The Dartmouth Conference

John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon organize the Dartmouth Conference, where the term “Artificial Intelligence” is officially coined. Their proposal? Every aspect of learning and intelligence could, in principle, be simulated by a machine.

1957: The Perceptron

Frank Rosenblatt builds the Perceptron, an early neural network capable of learning simple patterns, a foundational step toward modern machine learning.

1966: ELIZA, the first Chatbot

Joseph Weizenbaum creates ELIZA, a natural language processing program that simulates conversation by mimicking a therapist. It amazes and unsettles the public, planting early questions around machines and emotion.

1980: Expert systems go commercial

Expert systems like XCON are deployed in industry, automating complex decision-making tasks. This marks AI’s first major wave of real-world business adoption.

1997: Deep Blue defeats Kasparov at chess

IBM’s Deep Blue beats world chess champion Garry Kasparov, proving machines can out-strategize top human players in structured games.

2005: DARPA Grand Challenge

Autonomous vehicles complete a 100-kilometer off-road course, signaling a breakthrough in self-driving technology.

2006: Deep Belief Networks

Geoffrey Hinton and his team introduce deep belief networks, reigniting global interest in neural networks and setting the stage for deep learning’s rise.

2011: Watson Wins Jeopardy!

IBM’s Watson defeats the top human players on Jeopardy!, showcasing AI’s rapid progress in language comprehension and data processing.

2012: AlexNet and the ImageNet Revolution

AlexNet, a deep convolutional neural network, dominates the ImageNet competition. Its success sparks a global boom in deep learning and computer vision.

2014: GANs — AI Starts Creating

Ian Goodfellow invents Generative Adversarial Networks (GANs), giving AI the ability to create realistic images, audio, and more from scratch.

2015: AI Beats Humans at Image Recognition

For the first time, an AI system outperforms humans in the ImageNet Large Scale Visual Recognition Challenge.

2016: AlphaGo Conquers Go

Google DeepMind’s AlphaGo beats the world champion of Go — a game with near-infinite possibilities. This is considered a landmark moment in AI history.

2016: WaveNet Brings Natural AI Voices

Google DeepMind develops WaveNet, producing more human-like speech than previous text-to-speech systems.

2017: The Transformer Model

Researchers publish “Attention Is All You Need” introducing the transformer architecture, the foundation of today’s most powerful language models like BERT and GPT.

2018: BERT and Self-Driving Taxis

Google releases BERT (Bidirectional Encoder Representations from Transformers), pushing the boundaries of natural language understanding. Meanwhile, Waymo launches the first commercial self-driving taxi service in Phoenix, Arizona.

2018–2023: Generative AI and Large Language Models

Models like GPT-2 (2019), GPT-3 (2020), and ChatGPT (2023) bring generative AI to the mainstream, enabling everything from content creation to coding to real-time conversation.

2023: ChatGPT Goes Mainstream

OpenAI releases ChatGPT based on GPT-4. With millions of users, it becomes the most widely adopted generative AI tool to date, demonstrating new levels of reasoning, creativity, and collaboration.

And while we’re still far from Turing’s original dream of machines that truly “think,” the pace of progress is accelerating faster than ever.

Is AI close to replacing humans?

If you’ve seen the headlines lately, it’s easy to think the answer is yes.

Duolingo CEO Luis von Ahn recently shared that the company is going AI-first, using AI not just in the product, but to assist with performance reviews and hiring decisions.
At Shopify, CEO Tobi Lütke announced that teams now need to prove a human is truly required before hiring. Otherwise, AI will be the default.
Salesforce CEO Marc Benioff has said that in 2025, they might pause hiring software engineers entirely, because AI copilots have boosted productivity so much.

So yeah, it’s already here.

But maybe not in the way you expect.

AI isn’t “replacing humans” the way robots replaced assembly lines.

It’s not one machine swapping out one worker. It’s more subtle than that.

What’s actually happening is this:

Tasks are being reshaped. Workflows are being rewired. And many companies are asking:

“Do we need a full-time person for this? Or can AI handle 80%, and leave the human to focus on the harder, more nuanced parts?”

In some roles, that’s exciting. In others, unsettling.

So to make sense of it, it helps to zoom out and look at the three levels of AI, and where we are right now.

The 3 levels of AI (and why they matter)

Most of what we call “AI” today falls into the first level. But the levels are worth knowing, especially when people throw around terms like “AGI” or “superintelligence”.

1. Narrow AI (this is where we are now)

Also called “weak AI”, these are systems trained to do one thing really well.

Think:

ChatGPT generating text
Midjourney creating images
Google Maps finding the fastest route
Duolingo helping you practice Spanish

Narrow AI is everywhere. It’s impressive, helpful, and evolving fast, but it doesn’t understand in a human sense. It can’t generalize, reflect, or transfer knowledge across tasks. It’s pattern-matching at scale.

2. General AI (we’re not there yet)

Also called AGI (Artificial General Intelligence), this would be a system that can do anything a human can do: learn new things, adapt to brand-new tasks, reason, maybe even have preferences or emotions.

We’re not close to this yet. Some researchers believe it’s possible within decades, others say it may never happen.

3. Superintelligence (sci-fi territory)

This is what people fear (or fantasize about), a future AI that’s vastly smarter than all of us, across every field. Science, leadership, art, emotions, everything. Think Her or Ex Machina or The Matrix.

Right now? Pure speculation.

So no, AI isn’t about to replace humans entirely.

But is it already changing jobs? Yes.

Is it reshaping what certain roles look like? Absolutely.

Is it forcing companies (and individuals) to rethink what’s uniquely human? 100%.

If you’re feeling unsure about where you fit in all this, that’s normal.

The good news is that we’re still early. And understanding what AI can do (and what it can’t) is the first step toward staying relevant and resilient in the shift ahead.

Why AI matters for you and your future

AI isn’t just a Silicon Valley obsession. It’s reshaping how we live, work, and build across industries, roles, and everyday routines.

Jobs are evolving fast.

AI is taking over many of the repetitive tasks that used to soak up hours of human time (scheduling meetings, updating spreadsheets, reviewing legal documents), which frees people to focus on other higher-value work.

Entirely new roles are emerging.

Job titles like AI ethicist, prompt engineer, data translator, and model trainer didn’t exist just a few years ago. Now they’re becoming essential in modern teams, guiding how AI gets trained, used, and kept in check.

Businesses are transforming too.

What once required entire departments can now be done by a solo founder with the right AI tools. Startups are using AI to write marketing copy, build product prototypes, automate customer support, and speed up decision-making, staying lean and moving fast.

And in everyday life, AI is everywhere.

In healthcare, algorithms help detect disease earlier and more accurately. In education, adaptive platforms adjust to each student’s pace and learning style. Even how we shop, commute, and communicate is being shaped by invisible AI assistants running in the background.

And the pace isn’t slowing down. According to the World Economic Forum, while 85 million jobs may be displaced by AI by 2025, an even greater 97 million new roles are expected to emerge.

The bottom line?

Understanding AI isn't optional anymore. It's becoming a fundamental skill, a kind of superpower, that can help you navigate, adapt, and thrive in the future economy.

What about you? Is AI already changing how work happens around you? Share your take in the comments.

Final thoughts: should we be excited or worried?

Here’s the honest answer: a bit of both.

AI is powerful. That much is clear. It’s already reshaping how we work, learn, create, and communicate. But it’s not magic. And it’s definitely not a monster.

What we’re seeing now isn’t some sudden takeover, it’s the slow, messy, very human process of integrating a new kind of tool into our everyday lives.

Arvind Narayanan

puts it, AI isn’t some superintelligence waiting to escape. It’s a “normal technology”, like electricity or the internet.

That means it’s human-built, human-directed, and deeply influenced by how we choose to use it.

The real question isn’t what AI will do to us, but what we choose to do with it.

Learning about AI, and learning to work with it, is no longer optional. It’s quickly becoming the default across many industries. This isn’t some futuristic skill. It’s a present-tense capability. The companies, teams, and individuals who learn how to use it thoughtfully will have a huge advantage. Not because they believe the “hype”, but because they understand the reality.

So, should you be excited? Yes.

There’s real opportunity here to save time, unlock creativity, and build things that once seemed out of reach.

Should you be cautious? Also yes.

We need to stay curious, ask questions, and think critically about how AI systems are built and where they’re being used, especially when it comes to ethics, bias, privacy, and power.

The bottom line: this isn’t about waiting for the future to arrive. It’s already here.

If you understand what AI is, how it works, and why it matters, even at a basic level, you’re already ahead of the curve.

And if you’re still figuring it out? Start small, but start now. The gap only grows if you don’t.

AI blew my mind