What is Machine Learning, actually?

When the average person hears “AI”, their minds may jump to sci-fi ideas…think SkyNet from Terminator, or maybe Ultron from The Avengers. Maybe even Alicia Vikander in Ex Machina.

Super-intelligent robots bent on global domination.

Thankfully, and anti-climactically, the truth is far less dramatic. (At least thus far.)

AI, and more specifically, machine learning is all about finding patterns in data.

These patterns are what empower the digital world around us today. Ever notice how Netflix or YouTube seems to know exactly which series or video to recommend to you? That’s machine learning at work.

Self-driving cars? Blame machine learning. Siri or Alexa having a conversation with you? You get the point.

Machine learning didn’t just go global; it moved in. In fact, it’s hard to find an industry it hasn’t touched. ML is now in finance, farming, fashion, and probably your grandma’s thermostat.

It’s kind of like an intern who showed up to Earth, got bored, learned everyone’s job, and started automating things — except it never leaves the office.

Here’s the formal definition of Machine Learning, according to Wikipedia.

Let me translate that to non-jargonese.

Machine Learning is when computers figure things out for themselves, so that we don’t have to spell it out like we’re talking to a toddler.

Machine Learning might feel like a modern marvel, but its real history goes pretty far back. Like, I’m talking the 1950s.

Ever heard of this guy?

That’s Alan Turing. He was a British mathematician and all-around brainiac who’s often considered the godfather of computer science — and by extension machine learning.

He’s the one who posed the question: Can machines think?

While he didn’t build the actual machine learning algorithms themselves, he laid the intellectual groundwork that researchers would build upon decades later. Things like:

It was the Turing Test that was his real mic-drop moment. A machine that can learn behaviorally? This sparked the entire field of AI, and paved the path for what machine learning would eventually become.

Then in 1959, came this dude named Arthur Samuel. This guy decided to expand on Turing’s theoretical framework, and thought to himself: “I know! I’ll teach the machines to play checkers”. Literally.

Samuel developed a self-learning AI that learned to get better one game at a time. Eventually it got so good…that it beat Arthur himself.

Turns out Arthur had developed an early answer to Turing’s question if machines can think: Yes.

From there, things got nerdy in the best way. The '80s and '90s brought neural networks and algorithms that could start learning from data—though they still needed babysitting.

But the real kicker in the right direction? Computing power.

Thanks to GPUs (yeah, the gaming kind), we got the power to train massive models fast—and with open-source tools like TensorFlow and PyTorch, machine learning blew up into a global playground for coders, startups, and even the occasional curious artist.

And that brings us to now: a world where Machine Learning is everywhere and constantly evolving. From its humble beginnings in checkers strategy and academic theory, ML is now powering the digital backbone of modern life.

Think implicit vs explicit.

Traditional programming is explicit. And no, not in a Joe Pesci in Goodfellas way, but in a “you tell the computer exactly what you want it to do” type of way. It’s like you’re mapping out step-by-step, every logical decision you want your computer to make. (Imagine a bunch of if / else statements).

Machine learning is implicit. It flips the script; instead of defining out the rules for the machine, you “feed” it a bunch of examples (data), and it figures out the rules on its own. Remember: machine learning is about finding patterns in data.

Here’s an example:

Imagine you want a computer to tell the difference between a cat and a dog. Traditional programming would be like writing a 10,000-page guide: "If it has pointy ears, maybe it's a cat... unless it's a Doberman... unless it meows… wait, what?"

Machine Learning? You just show it a thousand pictures of cats and dogs, and it magically learns: “Oh, the cat is the one that looks pissed off 90% of the time. Got it.”

The word “mathematics” might trigger some less-than-pleasant thoughts. It might bring up memories of confusing homework assignments, alien-looking equations, and make you feel like you got hit full-on by the Confusion from one of those psychic-type Pokemon.

Or maybe you like math. In which case, kudos to you.

However, there is no doubt about it: Math is essential to machine learning.

It’s the why behind the wow. Machine learning simply wouldn’t be able to do what it does without it; algorithms rely on math to be able to spot patterns, make decisions, and improve over time.

Luckily, you don’t need to be Einstein or Archimedes to learn math for machine learning.

Really the focus falls on these 3 disciplines:

Once you get to know the math, it’s really quite beautiful.

Of course, I’ll also be covering all these in the future, in digestible bites.

Machine learning is the quiet superhero of the digital age. Except instead of wearing tights and a cape, it appears as code and math.

Notice how AI seems to be taking over everywhere? It wouldn’t be able to do so without machine learning. It’s what powers voice assistants like Siri and Alexa, helps your emails filter spam, allows cars to self-drive, and so much more.

For companies and businesses, it helps figure out what customers want, where they want it, how they want it, even why they want it.

It takes the oceans of data that are out there, and uses it as a knowledge base to help make smart decisions on autopilot.

Machine learning is shaping the future of nearly every industry we can think of. So from solo developers to giant companies, everyone’s racing to get smarter with ML - so you reading this means you’re at the right place.

Alright, I’ve used up enough RAM in your brain for now.

Until next time!

Want a free shoutout?

Tell me who you are, your background, any personal or career achievement, and maybe an interesting quirk about you.

I’ll hype you up, just like this dude below.

Just reply to this email (or send me an email).