Here's something extremely random.

The fax machine was invented more than 30 years before the telephone, identical twins don't have the same fingerprints, humans share more than 60% of our DNA with bananas, except those facts aren't actually random.

I mean, they're surprising, but what's actually random is radioactive decay, the exact speed at which the bus will pass a given point on any given day, or white noises, and these guys.

It turns out there are very few things in the universe that are truly random because the laws of nature are actually pretty predictable.

And that's important because some really crucial parts of your life and mine, from sending an email to medical trials to the fate of the stock market, all rely on randomness.

Hey, smart people, Joe here, and today, we are diving into what randomness really is and what it isn't, and whether anything in the universe is really truly random because that's up for debate.

And how every day, thanks to our scientific quest to find true randomness, your very identity is protected by things like a lava lamp.

Far out, man.

(bright music) Random, you know, the very sort of nature of the word makes it kind of hard to define, but random does have a scientific definition.

At its core, randomness is about unpredictability.

You won't be able to predict what happens next based on anything that's come before.

Take a coin toss, for example, the outcome's totally random, right?

I mean, the result of one coin toss won't tell us anything about whether the next toss will land heads or tails.

Each result is independent from the last time or the hundred last times.

Not so fast.

What if we could account for all of the infinitesimal variables involved before we guess the coin toss?

Maybe the minuscule differences of mass on either side of the coin, the friction of the air, the pull of gravity, the force and angle that we flip and hold the coin.

Theoretically, if you could measure all of these variables, you could use that information to accurately predict whether the coin toss would land heads or tails.

In fact, in 2024, researchers showed that a coin toss isn't exactly even 50/50 odds.

In their experiment, they calculated a 50.8% chance that the coin will land the same way up as it was in the coin tosser's hand.

And that's due to slight biases in the way that humans induce rotation as we flip it up in the air.

So even though for you and I, a coin toss is as good as random, it isn't truly random in the scientific sense.

That's because it's technically what we call a deterministic system.

If we give it a certain input, with enough information, we are able to determine the output.

A really simple way to imagine a deterministic system is like a vending machine.

You press A1, you're gonna get a Snickers every single time.

Snickers, get a Snickers every single time, nailed it.

Okay, now, instead of our deterministic vending machine, imagine a gumball machine.

You turn the crank and you might get red, blue, or green.

If we know how many gumballs of each color are in there, then we know the odds of getting a certain one.

This is a probabilistic system.

We can predict the probability of a certain outcome, but not the exact outcome of a single turn.

But again, I mean, if we knew the exact mechanics of our machine, if we knew the starting state of the gears and which gumballs were where, I mean, heck, let's say we had access to any physical measurement that we wanted, would we be able to confidently predict the color of the gumball?

In other words, if we did know everything about a given system, does that mean that everything is deterministic?

French mathematician Pierre-Simon Laplace gave us a really vivid way to picture this back in the early 1800s.

He proposed that if there was some hypothetical super intelligent, omnipotent being, what we now call Laplace's Demon, and they knew absolutely every characteristic of every atom in the universe, then theoretically, it could predict everything that would ever happen.

Now, this kind of thing is like candy for philosophers.

Many people think that Laplace's little thought experiment is a pretty strong argument for the idea that everything in the universe is deterministic or predictable, and that the things that we think are random only look that way because we're just limited by our measurements and our humanness.

But if we did have the tools and the ability to measure everything, then the whole universe would be predictable.

But there's a pretty compelling argument in the other direction too.

So it's been more than two centuries since Laplace came up with his demon, and since then, physicists have learned that when we get down to the scale of the smallest particles and how they behave, things are fundamentally unpredictable.

Basically, at the quantum level, particles do some really weird stuff, like quantum tunneling, where a particle can cross a barrier that it really shouldn't be able to cross, at least according to the laws of classical physics, or there's the Heisenberg Uncertainty Principle, which, true to its name, means that you can never know the precise position and the speed of a particle.

As soon as you start measuring one of those variables, you lose your ability to measure the other one.

Imagine you're trying to measure the speed and position of a ripple on a pond.

To see its position, you'd have to freeze the water like a snapshot.

But of course, then the ripple stops moving and you lose all the information about its speed.

But to measure how fast it's traveling, you need to watch it move over time, but then it's no longer at any single position.

Of course, at the large scale of, I don't know, a car, this kind of uncertainty is so small that it doesn't meaningfully affect our ability to predict the car's position or speed.

But at the tiniest scales, like subatomic, these minuscule uncertainties do make a big difference.

This is exactly what happens with quantum particles.

The very act of pinning down their location destroys the information about their speed, and measuring their speed means we can't know their position because they're moving.

It's the fundamental nature of reality.

This sort of quantum weirdness means that the behavior of the universe at the absolute smallest scales truly might be unpredictable.

So, if the smallest particles behave like fuzzy little probabilities, maybe there is something random in the universe after all.

I mean, if everything in our macro world is built on quantum stuff, that minuscule randomness could ripple upwards into our world.

Unpredictable quantum behavior gets amplified as it cascades through larger and larger systems, ultimately influencing the visible world.

It's like a quantum butterfly effect, and that means that many everyday events, like the weather in your backyard, are fundamentally random, not just practically unpredictable.

Okay, philosophers, I hear you down there in the comments.

What if we could theoretically know every single thing about every single particle, even on the quantum level, including aspects of their behavior that we don't even know exist yet?

What then?

Would we be able to predict that too?

Maybe all of these things that we think are random aren't truly random.

They're just random to us.

Well, there's actually a whole branch of science that studies where these predictable and unpredictable parts of our universe meet chaos theory.

- I'm still not clear on chaos.

- Oh, oh, it simply deals with predictability and complex systems.

- Yeah, what he said.

It's the study of how seemingly chaotic or random systems can still give rise to patterns or organization.

Now, it can sometimes feel like these questions are just to show off how much you like to think, but these are not just thought experiments.

We actually need real randomness for a lot of important things in our lives these days.

So it's important that we understand how randomness works, and the place where randomness is most important to you and me is.

- [Announcer] The internet, the information highway, the new electronic frontier, the worldwide web.

- Think of the internet as a network of communication networks.

- Here's the thing about the internet, okay?

The earliest versions of the public worldwide web, they were not created to guard information.

The web was invented to make knowledge sharing as easy and seamless as possible, but nowadays, we all use it for super sensitive stuff, right?

I mean, banking, medical data, passwords, private messages, and so much more that honestly, I don't wanna know about.

Every time that you enter your password into your banking app or send a text to your mom, that data travels through many different networks to get from you to its destination.

For example, an email from me here in the US to a colleague in London might travel from my computer, through my wifi router, to my email provider's servers across undersea cables, and then down a similar chain on the other end to my colleagues inbox.

That all happens in less time it takes me to have a sip of coffee, but once I begin composing that email, it's vulnerable to being intercepted at any point on that journey.

Imagine sending a postcard through the mail.

Anyone along the way can pick it up and read your message.

That is why we need encryption.

You know how pretty much every address on the internet starts HTTPS?

Well, that S stands for secure, and it's telling you that the data you're feeding into that site is protected from being spied on or stolen, and that protection comes from encryption.

There are lots of different kinds of encryption, but the whole point of all of them is to scramble that message, that information so that anyone who intercepts it along the way just sees a bunch of nonsense, but that you and the computer you're talking to, can decipher.

Now, you could protect your message by say, locking it in a box that only can be unlocked with a special password, and then you send a copy of that password key to your friend so that they can unlock the box.

I mean, that would work, sort of.

Problem is anyone who steals the key can unlock everything.

That's too risky to run the internet on.

So what actually happens is that your computer and the computer you're trying to talk to, they lock up your message with several different keys, one that they share out in public and one that they each keep secret to themselves.

Now, those keys aren't physical keys, they're not even passwords really.

They're made with mathematical calculations, and that math depends on random numbers.

- Whoa.

- Here's basically how it works.

Your computer says hi to the other computer it's trying to talk to.

"Hello, this is what kind of computer I am, and what kind of secret codes I know."

Then the other computer says, "Hi there.

Here's the kind of secret code method we're gonna use to talk, and also here's a certificate that says I am who I say I am and not some hacker pretending to be the website that you want to talk to."

Now, all we've done up to this point is verify that each of the computers is who they say they are and agreed on what flavor of secret code they're gonna use.

What happens next?

It depends on exactly which flavor of code they happen to agree on, but here's basically what it boils down to.

You have a message, I give you a public key, and you use that key to do some math to your message and lock it, making it unreadable.

Now, you send me your scrambled message, and I have the ability to reverse your math with my private key that only I have and unscramble your message.

Now, when I send you a message, we can reverse this whole thing.

You gave me the public key, I use that to scramble my message with math, and then lock it.

I send it to you and only you can unscramble my message with your private key that only you have.

That is a very simplified version of what happens.

If this kind of doesn't feel like it should even be mathematically possible, that you can encrypt something using a totally different key than you use to decrypt it, I feel you, because it is pretty weird.

The actual math involved, it's not the kind of math that normal people do.

It involves things like information theory and modular arithmetic.

And if you really wanna dig into those things, I don't know, go to college and become a mathematician or something.

But that's what makes this math so effective, it's hard to do, but more importantly, it's even harder to undo.

Working the math backwards to break this kind of encryption would take even the most powerful supercomputers that we have, like millions of years.

So it's not literally impossible, but it is practically impossible.

The important thing is, those keys, that all this mathematical encryption and decryption depend on, they rely on random numbers.

Now, if there were any predictable pattern to choosing those numerical keys and people wanting to listen in on that conversation, could guess or reproduce it, which would allow them to unlock the box that they're not supposed to be able to unlock.

And here lies our really big problem because computers can't generate random numbers, at least not on their own.

- Classical computers are fundamentally deterministic, right?

Like computers, they're calculators, that's what they do.

Like they do two plus two is four, and they do it every single time.

This is the thing that makes computers so great, but it also means if you say, "Hey, pick a random number," like, they're not good at that.

- Computers are deterministic systems because we've designed them to be.

When you press a key or run a program, the computer follows an exact sequence of logical operations processed through predetermined pathways etched in silicon, and then it produces an output according to the rules that we programmed into it.

So computers are incapable by design of generating truly random numbers, unless they have help.

See, computers can incorporate outside data into their algorithms, and if that data's unpredictable, say, I dunno, the precise measurement of the temperature at the top of Mount Everest at this very second, or the difference in electrical current across a neon gas tube, or any of a bunch of other natural phenomena, these are unpredictable, at least to us, because again, we're not Laplace's Demon.

Cloudflare here is a company that powers the infrastructure that the internet runs on, and they came up with some pretty interesting ideas for how to insert random data from the natural world into all of their secret code math.

They use physical installations in their offices, things like double pendulums, wave machines, and lava lamps.

In one of their offices, Cloudflare has cameras trained on a wall full of lava lamps, converting their random undulations into data they can use to power encryption algorithms.

- A thing we do at Cloudflare with these random walls is thinking about how to take that randomness from the natural world and bring them into a computer.

- So why lava lamps?

Well, because they rely on some pretty simple physics to create something infinitely complex.

Now, the two substances here in the glass vial are two different densities, and they're also what we call immiscible liquids, meaning they won't mix together.

Although the exact live lamp recipe is a closely guarded proprietary secret, this is some kind of dense wax suspended in a lighter density liquid.

When you add energy to this system, like from the heat source down here in the lamp's base, that denser liquid becomes lighter and lighter, rising to the top of the vessel.

There, it's further away from that energy source, right?

So it cools down, it becomes denser, and sinks back to the bottom again, just looping over and over in these really satisfying globs.

This is actually a form of fluid dynamics that physicists called Rayleigh-Taylor Instability.

It's part of a field of physics called Bubble Dynamics, which is honestly just so fun and cute, isn't it?

The Rayleigh-Taylor process is affected by any tiny change in the materials that are used in the recipe and by everything around the lava lamp, like the temperature of the air, if the chemicals in the wax are degraded, how much heat is being emitted by that light bulb down here?

And so much more.

Predicting or controlling for all of these variables would be so computationally difficult that it is virtually impossible.

Basically, no two lava lamps have ever or will ever behave in exactly the same way, which is why these and these and these and these all have a role in protecting about 20% of the internet.

- When you think about how to take an image of these lava lamps and turn them into a number, an easy way to imagine that is to say, "Okay, I'm gonna take a picture, and it's a digital picture, so you know, evaluate each pixel, decide what color it is."

You're basically giving each pixel a number, and then you bring all of that together.

When you take a digital picture of it, it gets converted into numbers.

All we have is a camera that looks at all these lava lamps or double pendulums in our London office or wave machines in Portugal, and because they are unpredictable, the picture of them, when you convert it into a big number, is unpredictable.

- And Cloudflare isn't the only company doing this.

Others use different methods to generate randomness, like atmospheric noise, others use radioactive decay.

Some even use the timing of users' mouse movements and keyboard strokes, all of which are random enough, at least with our current limitations of measurement and computation, even if some would argue that they aren't truly random.

This scientific drama about whether randomness truly exists, like on a philosophical level, it has a real world purpose other than just making your brain feel like a melting lava lamp.

If we're thinking about quantum stuff and whether it's truly random, then we have to talk about quantum computers.

Because quantum computers have the potential to be exponentially more powerful than any classical computer that we've ever built.

They could actually solve the incredibly complex and difficult math that is currently what makes modern encryption methods work.

What would take a classical computer, say millions of years to solve, a quantum computer may be able to do in a few minutes or even seconds.

Quantum computers exhibit the kind of true randomness that we've been talking about.

They could be used to create quantum random number generators.

It sounds pretty cool, devices that directly harness quantum uncertainty to produce truly random numbers, paving the way for an entirely new era of quantum encryption.

It's Schrodinger's future, one where encryption is both broken and saved.

- We can have lots of debates about when quantum computers will be practical and when they'll be able to break encryption.

Fundamentally, encryption are these math problems.

There's still math problems that are hard for quantum computers to go one way, there are math problems we know are hard for classical computers, so we're gonna take the randomness that we have and use that to create great encryption that, you know, works that is hard for both classical and quantum computers to solve.

- So next time you log into your email or make an online purchase or watch a video like this one, remember that was pretty random.

Stay curious.