I feel like an old man whenever I buy a new computer.

I mean, why does my laptop need sixty four gigabytes of memory?

When I learned to program on a PC that had twenty kilobytes?

No joke.

Kids these days don't understand how hard we had it back in the day.

Right, But it's also a nice feeling.

It tells me that we're making progress and that's good.

It's creating new worlds and new ways of life.

It's literally saving lives by accelerating science.

That's all great stuff, right, But how long can it go on?

What is the engine of this exponential growth in computing power?

And can we count on it to take us to the stars, to cure cancer and to develop self driving toothbrushes.

Today we'll dive into the physics underlying this trend and ask whether there are fundamental limits that could block us from achieving our dreams.

And we'll talk about whether there's danger and assuming technology will solve all of our problems.

Welcome to Daniel and Kelly's Extraordinary Universe.

Hello.

I am Kelly Wiener Smith.

I study parasites and space, and I realized when we were starting to do this episode that I wasn't one hundred percent clear on what mores Law meant exactly.

Hi, I'm Daniel.

I'm a particle physicist and I've been programming computers for more than forty years.

They get faster and I get slower.

Oh you're not slowing down yet, Daniel, you stab.

So my question for you today, Kelly, is what was your first computer?

Let's age Kelly.

Okay, So later when we talked to Adam Becker in our interview, he mentions that there was a while there where folks wouldn't get a computer because you'd wait as long as you could, because the computers kept getting so much better so quickly that if you could wait, your computer would be much better.

Yeah, and so my family waited way too long.

We didn't get one until I was in like high school, and I know, and I don't even remember what it was.

But in the meantime I had to write my essays on like it was like a brother typewriter, but it also had a little electronic screen, and so I could very slowly and laboriously click through my essays and then I would print it and something would be wrong.

It would take me forever to find where the error was.

It was very annoying.

But what about you, did you have like the first Apple computer?

Ever?

Oh?

Apple was way too advanced.

I go way before that.

What My first computer was a Commodore VIC twenty, which I think had twenty kilobytes of RAM, and we stored stuff on an audio tape, you know, like you write a little program and then you'd stored on these cassette tapes that you could later listen to and like, ooh what does that sound?

So yeah, we were very very early.

In fact, I remember hanging out with my dad in grad school while he was doing his research and he was literally feeding punch cards into those punch cards machines.

So I feel like I've personally experienced a huge fraction of the transformation of computers into the basically supercomputers we have today.

I mean, my smartphone is so much more powerful than anything my dad ever used in his research.

That is absolutely amazing.

I didn't even know that we were storing data on like cassette tapes.

Oh yeah, that's amazing to me.

Yeah, before magnetic floppies for sure.

So when you like are using your are you a MacBook guy?

I am?

Yeah.

It's like when you're using your Mac Do you every day think I am so lucky I'm not doing this on punch cards or you just do you take it for granted?

Now?

I think it's awesome.

It's incredible.

I mean every time I get a new MacBook, I'm like, while, this drive is ten times bigger than anything I've ever seen, and the memory is just shocking.

And it's also then incredible to me how rapidly our computational ambitions grow.

You know, my group does a lot of computation, and we're basically limited by computation, and so every time we get more powerful computers, we scale up our ambitions and solve bigger, harder problems, and so we're always at the edge of what the computers can do, right, Like, we have an infinite number of questions we could ask with harder computers.

So yeah, I'm in awe of the MacBook, not just because it's so much more powerful than anything I've used, but it's so reliable.

I mean, I spend hours and hours a day in front of this thing.

It almost never gives me problems.

So yeah, it's incredible.

What engineers have provided it is incredible.

And today we're going to talk about one way in which that incredible ability has been expanded, which has to do with Moore's Law, which I thought was about how much data you can store on your computer, and I think it's much more than that.

You're right, it's much more than that.

It's also about how things are growing over time and how long that will continue.

And there's this lore about Moore's Law which is permeated Silicon Valley and broader culture.

So in a minute, we're gonna also talk to Adam Becker about how this has impacted philosophy and politics and policy and how it might affect our future.

But first we wanted to know how long people thought Moore's Law might continue to make all of our computers faster.

So I went out there and I talked to our group of volunteers.

Here's what they had to say about the future of Moore's law.

So I'll give it six years and then I'm gonna sell my nvidious stock.

But quand computing will change the game.

We are restricted by things like housmow.

We can make stuff, so I think it's not true anymore.

My first thought was to sign no, but I know very little bit quantuine computing.

I would say until the computer speed, which just be the light maybe ten years, another good decade or so.

My understanding is that it's already done.

I don't think we are doubling in raw processing speed.

In my understanding, Moore's law kind of slowed down for laptop desktop chips a number of years back, but has continued with mobile just because they were a little behind.

But then you also have graphics and neural processing units that power our AI platforms of today.

So can we go into the future.

I think we can go for a number more years, given the innovations in transistor stacking.

I thought Moore's law had to do with cost decreasing as speed increased.

We do seem to be close to a tipping point with electrons being too large for the tiny circuitry.

However, it seems that optical circuitry might be a good replacement for that.

They're currently reaching the lower limits of workability before they start reaching quantum effects.

With silicon.

We are getting closer to the particle level and that stops us.

Honestly, I thought it had already stopped.

So do you think these are optimistic or pessimistic?

I mean I think they're realistic, which seems to be the option that always gets left out.

So I think there were a lot of people who said, you know, I thought we've already reached the limits.

So we're getting close to reaching the limits.

And I'll admit that I did not realize we were getting close to reaching the limits.

But it seems like a lot of our listeners are on top of this trend.

Yeah, and so we're not giving financial advice.

So I won't tell you whether or not to buy or sell in Vidia stock.

But you know, sometimes I wonder about these tech companies because their stocks also seem to follow Moore's law, Like, how can Google just keep getting more valuable?

I keep missing out on buying Google.

I can tell you that if you can make a time machine, one of the first things you should do is go buy in VideA and Google stuff.

All right, So let's dig into it.

What in the end is Moore's law?

So Moore's law was something postulated by Gordon Moore.

He has a first name.

And he was one of the found of Intel, so a big dude in like, you know, semiconductors and electronics.

And he suggested initially that the number of transistors you could squeeze onto a chip would double every year.

And it's a little bit more complicated than that.

He also was talking about the power usage and the cost, but roughly speaking, he was talking about the density of transistors on chips getting higher every single year, which means the speed of these computers is growing very quickly.

So transistors are about speed and not storage, or are they about both?

They're about both.

The transistor fundamentally is a tiny programmable switch.

And the reason that computers got small and got fast is because we were able to make transistors small and make them fast, which allows us to have lots and lots and lots of switches in a small area, which is what allows the computer to be complex and to be fast.

And so essentially it's saying we can make computers denser every year, and that makes computers faster and more powerful.

Okay, so this has to do with why we went from computers that took up entire rooms to something you can now stick in your bag and take with you.

Yeah, exactly, And we'll dig into that in a minute.

But the history here is that in nineteen sixty five More predicted this and then ten years later he revised it.

He was like, well every year, maybe that's too optimistic.

Let's go for every two years.

And so that was the prediction in seventy five, and you'll see that it mostly held up until fairly recently.

It's sort of an extraordinary prediction in that sense, though.

You know, anytime there's a prediction that holds up, you got to wonder, like, well, what were the other predictions this person made, Like you just spew predictions constantly, Eventually you're goind of get one Ris.

Yeah, yeah, well, especially he gave himself another decade to like fit the trend line.

That was pretty generous to himself.

But okay, exactly, So let's dick into what a transistor is and why it allows computers to be faster, because that's crucial to understand why More's law has worked, how we've made it work, and whether it's going to work in the future.

Basically, a transistor is a programmable switch like computers operate on digital logic.

I have a number in the computer, the number four.

They store it in binary.

But to store things in binary you need a physical system.

But that can store a zero or a one right the way you can like write a digit on a piece of paper.

That's like I'm representing the number four by scratching this graphite onto this sheet of paper.

I want to store things on my computer and zeros and ones because binary is the code for computers, and physically that means a switch, you know, as you can imagine either just like literally like a light switch, but here we're doing an electronic switch.

Okay, And so just to if we switched from using transistors to things like DNA to store data or quantum computing, could you still apply More's law, like if we switched to some other method or is More's law specifically about the transistors that we're talking about now, Yeah.

Great question.

You're talking about fundamental changes in how we do computing.

So currently computing operates on bits, zeros and ones, and we're saying those are represented by transistors, which is like a physical implementation of that bit.

You switch to quantum computing, the fundamental element of that is a cubit, which isn't necessarily a zero one.

It has the probability to be in several different states, and so it requires a different physical system to model that.

We don't use transistors or not even like quantum transistors.

In fact, transistors are already relying deeply on quantum mechanics, so quantum transistor is redundant.

But yeah, cubit, there's no guarantee that you can like build cubits and then build them more densely and more rapidly.

There's certainly no More's law for quantum computing that's a guarantee.

And biological computing like DNA is super awesome as an idea, but there you have like four possibilities, right, DNA is basically base four, and so it's a question of like how do you encode numbers into DNA?

Do you use all four bases?

Do you group them into two to make binary?

The technology is fundamentally different, So again you wouldn't expect necessarily further to be more solved, but you might get some other law which could be better.

So but yeah, Mores law reflects the details of the technology we're using to represent the fundamental element of computing, which is a zero or one, and then crucially the logic that operates on those zeros and ones.

Let's get into that logic.

Yeah, because what you want to do is represent like numbers in your computer.

I want to put the number four in but also I want to calculate stuff.

I don't just want to write four into my computer.

I want to be able to add four to two.

I want to be able to compare four and seven.

Right, That's what allows you to program a computer for it to do useful computation.

And if you know something about computing, you know like the basics of computation is a Turing machine which can like read in numbers and write numbers onto this infinite tape.

And so in order to do logic, you need to be able to have things that respond to different inputs.

So in logic you have things like gates.

Like a not gate is something which if you give it to zero, it responds a one.

If you give it a one, it responds to zero.

It's like a logical map from inputs to outputs, or an a gate.

Right, an and gate gives you a one if both inputs are one and a zero otherwise.

Or the converse of that is a nand gate NA n D, which is the combination of an N gate and a knot gate.

And the really cool thing is that if you can build a nand gate, you can build any logical map nandgates are like the basis function of logic.

So if you have nands, people have shown that you can build any map from inputs to outputs and essentially any sort of computer logic.

So you can build knot gates and and gates out of transistors.

Transistors are like this digital switch, and we'll go into the detail of the physics of how they work, but essentially they're a programmable switch.

You can turn them on or off in response to other stuff.

So from that you can build logic, and from that you can build nandgates, and from that you can build literally anything like adders and comparitors and anything you need in computers.

So this is like the basically the smallest little lego brick of computing is a switch, a programmable switch that goes from zero to one.

And that's what a transistor is an implementation of.

And it didn't have to be a transistor.

It could have been something else.

It could have been DNA, it could have been whatever.

But this is like the best, fastest, smallest thing that we have invented, and this is what revolutionized our society.

What does the transistor look like?

Yeah, what does the transistor look like?

It looks like nothing because it's super duper tiny, right, Like the ones that we're building these days are order nanometers right, so like you put one on your finger, you can't see it.

The number of transistors on a typical chip is billions, so you can't see an individual one.

There used to be able to, like when they were first building them in the fifties, you would like make one, you know, on a bench.

You could think of it as sort of like three wires coming together.

You have a source, a drain, and then a gait and the gate basically decides do I connect the source and the drain Do I open or close this switch?

And so it's sort of like a wire with a lever in it that you know you can open or close, and then another wire that determines what whether or not that's open or closed.

So that's not a physical description of what they look like.

We can get into like the semiconductors in a minute, but that's sort of the logical construction.

And when I think about more's low, I think, well, what is exactly is the connection between more transistors and speed?

Like it's cool to have things small because then you can put a computer in your watch or whatever.

But why do smaller computers operate faster?

Because that's really the crucial key.

When you sit down at your laptop, you're not like, wow, the transistors are super dense.

You're like, wow, you know word opened in a bill a second instead of you know, spinning my beach ball forever.

Yeah, So it's the speed that's really crucial, and that's really transformed society.

Right, it's computational power and miniaturization means faster operation for a few reasons.

Number one, things just don't have to go as far.

Right, Electronics is limited by the speed of light.

It's not instantaneous.

You close a switch, the electrons don't move instantly.

Right, the current doesn't change instantly, and so we are still limited by the speed of light.

And so if the distances between the transistors are smaller and the transistors themselves are smaller, things just happen faster because there's a speed limit to information in the universe.

That's awesome.

I guess I hadn't imagined that as a limiting factor.

Okay, super cool, what's next.

Yeah, that's one.

The other is you can have wider data paths, like instead of just using thirty two bits to store your numbers.

You can use sixty four bits, right.

Remember bits are this essential element of binary numbers, and so if you have like a two bit number, you can only store between zero and four.

If you have an eight bit number, you can store many more numbers.

You have thirty two these days computing a sixty four one hundred and twenty eight bit If you hear about these numbers as the sort of the core the computing of your CPU or your operating system, that's what it describes, like what size numbers are we operating on?

And this is important because basically it's how much your computer can do in parallel.

Like if you can add two hundred and twenty eight bit numbers, it's really one hundred and twenty eight bit wise operations done in parallel instead of if you're doing sixty four bit numbers, then you're only doing sixty four operations in parallel, and so you can do more operations in parallel, You can pass more data at the same time, and so data flows more quickly.

Another thing that really limits the speed of computers is how long does it take to get the data into the actual CPU.

Right, Like you have these numbers in memory, you want to do some calculation on you got to slurp them from memory and put them into the registers in your CPU.

They're actually doing the comparisons or the adding or the subtracting or whatever.

And so the wider the data path, the faster the data gets loaded, and the faster the computation happens.

And CPU probably means senorebditis pyro wetting underwater.

What does CPU mean?

CPU means central processing unit.

It's a thing on your computer that does the actual crunching, does the adding or subtracting or comparing, or loading or unloading or writing to memorates the closest thing we have to a digital brain.

Okay, but there's another sort of mechanical element to like why speed means faster computers.

And you know, back in the nineteen fifties, people were doing electronics, and they're doing it sort of the way you might do it in your garage.

You got resistors, you got capacitors, you solder them together, you make these big sort of physical circuits.

But in the late nineteen fifties people invented what's called the integrated circuit.

Integrated circuit is just like you know, it's a big green board and it's got the whole circuit printed onto it.

You don't like solder the components together, and this really allows for like the embedding of these transistors and other components inside these protective layers, which enhance their reliability.

And so that means you can make them smaller, you can make more complex circuits that you didn't have to like wire together yourself with dripping hot bits of solder, and so this makes them more reliable, so you don't need as much error correction, et cetera.

And that allows things to be smaller and to be faster.

So you've got integrated circuits, you got wider data paths, you got shorter distances to travel, and you have faster switching.

All these things are why more transistors means faster computing.

Okay, And so when did we get our first transistor?

Yeah, so the transistor was invented in Bell Labs in nineteen forty seven, I think it was.

And there was a lot of research in the forties different kinds of technologies for transistors.

Try this, try that, try the other thing.

But the basic concept was invented in the late forties in Bell Labs and you know, Bell Labs is one of these like elements of another era, an institution that I really miss.

You know, it's a privately funded research lab that did basic research.

You know, this is an arm of the telephone company.

But they just like gave nerds money and said, hey, play around, figure stuff out, and they came up with things like the transistor, which is I think a trillion dollar idea would be underestimating it, right, Like, it's literally the foundation of our entire economy.

Its transformed the way we live.

Wow.

And so I think even if every other piece of science was a waste of money, this one brings the average up like this one idea like means all of science has been worthwhile just from a purely economical, cynical point of view.

And that's the way science works, right, like a lot of fuzzes out and occasionally a huge, huge payoff.

Anyway, it was the late nineteen forties people figured this out.

And you know, we've only had quantum mechanics for a couple of decades.

Then.

People had ideas for making transistors before then, but weren't able to make it work.

But at Bell Labs, smart guys figured this out.

One Nobel Prizes.

It was really pretty awesome.

Awesome, And when we get back, let's talk about how we went about shrinking these transistors.

All right, So in nineteen forty seven, Bell Labs creates the transistor just in time for us to use it to get to space, which is the most important topic that we have to keep getting to every episode.

All right, So now we've got the transistor, how do we go about shrinking it?

Yes, the transistors are built out of semiconductors.

You know, you hear the semiconductor industry everywhere, And what does that really mean?

Well, we understand what conductor is, right, It's something where electricity can flow.

An insulator is something where electricity cannot flow.

And to understand that, you have to take your vision of the atom where you have like electrons orbiting around the nucleus or being in fuzzy quantum mechanical states, and think about what happens when you put a lot of atoms together, Like, what is the energy level of an electron around an iron atom?

Well, it's a bunch of levels.

What happens when you have a billion iron atoms and a lattice, what happens to those electrons?

Well, they don't really belong to any individual nucleus anymore.

They sort of like move around the iron super highway.

They can flow around from here to there.

And what distinguishes a conductor from an insulator is whether or not there's a big gap between energy levels, like can the electrons get up to those energy levels where they can flow around between all the atoms or not.

If they can get up there, then it's a conductor.

If there's a really big gap so they can't get up there, then it's an insulator.

Semiconductors are things that are sort of halfway in between.

They have a medium sized gap between the energy levels where the electrons are stuck around individual atoms and the ones where they're just flowing across the super highway.

And so that's something you can control.

If you tweak it a little bit by like adding them a little bit of germanium or this other kind of thing, you can control that gap.

And so what you want when you're building circuits is you want places where things conduct really well and then places where things conduct really really terribly.

So rather than having to have different kinds of material, like if I build a circuit in my garage, I use copper for the wires and then to use rubber for the insulators.

It's better if you can have a single kind of material and sort of tweak it and like, okay, I'm going to make this part of it conductor and that part of it an insulator because it allows you to print circuits onto your material.

Okay, And so what is the material you use?

So we use silicon.

Silicon is the semiconductor of choice, and then you dope it with various things to change its behavior to make it a conductor.

And the way that we have shrunk transistors from pretty big stuff you could see on your garage bench to tiny stuff almost the size of atoms is through a technique called photolithography, which essentially prints a circuit onto a piece of silicon.

We grow these huge silicon wafers that are like ten inches and then you want to print a circuit onto it, and you want to print like billions and billions of transistors, and you want them to be as small as possible.

For the reason we just point it out, So like, how do you print this stuff onto a piece of sar silicon?

So this is what photolithography is.

Essentially, you design your circuit on the computer, and then you print on the surface of the silicon this thing called a photo mask, and the photomask like protects the silicon from the next thing you're gonna do to it, which is blast it with really high energy light.

So you shoot like super high energy light at the silicon which is partially covered by this mask, and the parts that are exposed get a little bit chemically changed.

Then you dip the whole thing in like acid, and the parts that we're exposed get like eaten away, for example, and so what you're left with is just the pattern that you wanted.

That's like a very hand wavy explanation of how photolithography works.

But the things to understand is that it's limited by those photons.

Like if you use photons that we really wide wavelengths, then you're going to get a fuzzy picture.

If you use photons with really narrow wavelengths, which means high energy photons, right, then you're gonna get a much crisper picture.

And so over the decades, we've been trying to shrink these transistors to get more and more transistors on these chips and have faster computers.

And one way to do that is to crank up the energy of those photons, and so now we're in the like extreme ultraviolet limit where the photons have a wavelength of like thirteen or fourteen nanometers.

Wow, And that's hard because it requires like special optics.

You can't just use normal lenses to bend this kind of light.

It's why it's very hard to do, like X ray optics.

Also, the higher the energy light, the harder it is to bend it.

Have we maxed this out?

We probably have maxed this out, because anything beyond this requires insane optics.

Like already the optics are insane.

You know, making a single mask for these things costs like hundreds of thousands of dollars, and there's like a few places in the world you can do this kind of stuff.

The equipment is extremely expensive, the operating conditions are very very particular.

You have to have specialized clean rooms.

Like this is really the pinnacle of technology.

It's incredible, and that's why you know a few of these players in this field, like the Taiwannee semiconductor industry is so important for the worldwide computing industry.

Like a single company goes down and like we can't make computers anymore.

Wow, Oh my gosh.

Right, all of the geopolitical tensions just came into focus exactly.

That's one reason why Taiwan is so important because a lot of this stuff is done by Taiwan, these firms.

All Right, so we figured out photo lithography and we've kind of reached the limit.

Yeah, is that the end of the story.

It's not quite the end of the story.

And you know, I'm more to say about like how impressive it is.

Like in the mid nineties, we were doing things at like three hundred and fifteen nanimeters scale, which sounds pretty awesome, Like that sounds pretty tiny.

And then late nineties it was like one hundred and eighty animeters, and the two thousands it was sub one hundred nanimeters.

These days we're getting down to like ten nanimeters single nanometers.

It's amazing, but it's getting harder and harder because we're already beyond the wave length of the light that we're using, right, and we're approaching the size of the atom.

Right, silk and atoms are like zero point two nanometers across, so like you're going to be bild a transistor out of something.

It's like you know, you can't make things out of legos if you only have a few of the bricks, right, And so it's challenging to make transistors smaller than about a nanometer because you're really reaching that fundamental limit of the size of the silicon atom.

And every year it gets harder.

Like it's true that we've increased the transistor density every two years, we've doubled it, but the amount of money spent in this research has increased by a factor of ten or twenty, so it's not like a constant effort every year to achieve this.

We have to ramp up the energy and the creativity.

And that's great, you know, it's like inspired all sorts of cool things and spinoffs and whatever.

But it gets really really complicated.

And the sort of cutting edge of this is to now start stacking these transistors, so well, don't just think of it as a plane.

Let's go up in the third dimension.

Let's make the transistors more powerful by shrinking them further and then allowing them to grow in sort of the third dimension above this sort of plane.

And the leading edge of technology right now are these transistors called fin fet so FET, which stands for a field effet transistor and then a finn meaning like they literally have this like fin over the gate that controls it like a physical thing.

It looks like a shark fin.

That makes it possible to be efficient while shrinking even further, so you can make the sort of footprint of it smaller while keeping its effectiveness because you have this third dimension.

And so that's what stacking is.

And really people think that we've reached the limit of what we can do technologically, and as some of the listeners have said, we're going in other directions, like instead of making your CPU more dense, you just have multiple cores, or you start building other dedicated stuff like graphics processing units that are really good at linear algebra, which is needed for graphics and also for machine learning.

And so we're sort of like simultaneously trying to go in many directions at once to improve the power of computing.

But it's not clear that we can keep doing this, and a lot of people think that we really are at the edge of what we can do to improve computing speed.

Wow, and so stacking isn't going to be the magic solution because there's like limits on stacking.

Yeah, exactly, like stack can get you a little further.

But if we're going to keep doubling, then it's hard.

And you know, I think there's something to be said about the sociological impact of this doubling.

You know, Moore's law is not something that comes out of like the fundamental laws of physics.

It's something that was predicted and that we maintained really over decades, which is really incredible.

Like one of Intel's earliest processors, the four thousand and four, had twenty three hundred transistors in it, right, whereas like the eighty three eighty six, which I've spent a lot of time programming on as a teenager, had like hundreds of thousands of transistors in this MacBook I'm sitting in front of has billions.

It's incredible, but it's sort of guided the field.

I think people because they thought this was possible and maybe even inevitable, they worked for it.

It's a target, you know, And so if you think something as possible, then like you stay late and you push hard and you come up with new ideas, and so in some sense it's a self fulfilling prophecy.

Okay, so first of all, have we hit the limit to More's law?

Already, or you just think we're going to hit it soon, Like when is the first year you think where we'll be like Moore's law gone.

I think we're right at that inflection point.

You know, we're still seeing improvements in speed, we're still seeing big boosts and productivity, but we're sort of running out of avenues and so I still see that, Like my MacBook is faster than the one I gave my son, which is my two year old MacBook, But I don't know that the one I'm going to get in two years is going to be as much faster.

So I think we're right at that inflection point.

So that feels a little scary to me.

So, like, you know, over time, we've gotten computers that are better and so at least, you know, in my field, almost every five years you expect, you know, the statistical models of the systems that we study to get more complicated so that we can get a better understanding out of each one of our data sets.

Are we not going to be able to do that anymore?

Or do you think in twenty years our computers are just going to start getting bigger again until they fill up a room, Because we're going to want to keep getting more complicated in our analyses.

Yeah, well, I think we're already seeing our computing getting bigger.

I mean, think about like the data centers that are being built by Meta and Microsoft is like trying to turn back on nuclear reactors because they need the power for their AI data centers.

These things are vast and they're consuming huge amounts of our resources.

So I think, yeah, our appetite for computing is just growing, and even if our computers don't get faster, we're just going to keep building them bigger and bigger.

But I also think that for those of us who do things that are not directly just computing, that there are other ways to increase speed.

I was talking to Katrina about this, and she was saying that Moore's law also kind of applies to genomics.

You know, like the first study of a human genome costs like how many millions for one genome, And then the NIA had a target like it should cost less than one thousand dollars to sequence a human genome, and they hit that target and now it's cheaper than one thousand dollars.

And where does this come from?

Part of it come from computing, but also part of it comes from like the miniaturization of biology.

And I've seen this just like observing her field.

Something that used to be like a PhD level of work then in a few years becomes a little box on the lab bench.

You press a button and it's done while you're at lunch, right, Yeah, and that allows you to now do things that were impossible ten years earlier.

And that kind of transformation of the scope of the capacity of the field enables broader, deeper thinking.

And that's not just computing, right, that's the menigtization of like the actual biology, like micro little bits.

It's essentially like what therapnose was tapping into this feeling like, oh, eventually we should be able to diagnose diseases with tiny life jobs of blood in this kind of sense.

So I think that there's lots of dimensions that we can follow for improving our scientific and technological industrial capacity.

Is not just is my computer faster.

So when someone says, like you just did such and such follows Moore's law, do they essentially mean we do it better with smaller stuff, and like we do it exponentially better in particular.

Yeah, I think it's about exponential growth that's a crucial thing because you know, exponential growth builds on itself.

You know, it's like putting a dollar in the bank.

Every year you have more dollars, and those dollars earn more dollars, and eventually you have all the dollars.

Whereas like if you're just selling lemonade and you're making a dollar every day, you're making the same amount of dollars every day.

It's not increasing.

So it's all about that exponential growth.

And I think that that's what people mean when they refer to Moore's law sort of more colloquially than just like the density of transistors on a chip.

That's kind of interesting because like Moore's law isn't really a law, like it's an observation.

And so it seems like now anytime we see exponential growth, we say the words Moore's law instead of just saying exponential growth or am I being negative?

No?

I think you're right, And I think it says something about our aspirations.

You know, we live in a time when we expect our children's lives to be very different from our lives and our grandparent lives, and that's really unusual, Like most of human history.

You could tell your kids what their life was going to be like, because it's going to be basically the same as yours and your grandparents for like the last ten thousand years, right, because like change was inconceivable because nobody had ever experienced it.

But now we live in a time when, like we know that's not true, and so I think it leaves us with this like gap in our wisdom.

And then we project forward, and some of us are optimistic and we're like, Yay, this is going to change our lives in a way that solves all of our problems.

And as you'll hear from Adam, some of us are less optimistic, you know, about what this means and whether it's the right way to place our bets.

I do feel like that was a slightly simplified view of history, But this isn't Daniel and Kelly's historical universe.

So we're moving on.

Hey, I have to fit it into about one minute, so I'm not going to do a deep dive.

But yeah, I mean, do you disagree with me about the broader assessment of the way that human experience has changed.

I do think human experience was similar for a really long time.

You know, like when our ancestors moved out of Africa, there was probably a lot that changed in a couple generations, and the Industrial Revolution and climate.

Yeah, yeah, I think there's probably been a lot of moments where things were like, oh crud, but usually probably they were getting worse, whereas now we're hoping that it's getting better.

But anyway, so when I was reading Adam Becker's new book More Everything Forever, there was a discussion on Moore's Law where I realized, like, oh my gosh, I fundamentally didn't understand Moore's Law very well or what like underpinned Moore's Law, and I didn't realize that we were, you know, perhaps reaching the end of Moore's Law.

And so we reached out to Adam Becker and asked if he would talk to us about sort of the implication of, you know, the death of Moore's Law.

I'll be super dramatic about it, but how this expectation of exponential growth impacts our view of the future in ways that are not always necessarily realistic.

Let's say, all right, so then we're very happy to welcome to the podcast.

Adam Becker, who is an astrophysicist turned author.

He wrote the widely acclaimed book What Is Real, one of my favorite books about quantum chinnings.

If you write me to ask for a book about quantum mechanics that explain stuff in an accessible way, often recommended.

And he has a new book out called More Everything Forever about the rise of techno utopiasts and how we can project our future and the future of technology.

Adam, Welcome back to the podcast.

Thanks, it's great to be here.

So let's start just by talking about Moore's law.

It's the foundation of so much of the techno utopian movement.

Why do you think that it has inspired sort of this broader fanaticism, especially when it's just like an empirical observation, not like a deep law of the universe.

Yeah, that's a good question.

I mean Moore's law.

Yeah, it is an empirical observation.

But it's so regular, it's so comforting, and it has you know, the fact that Moore's law held more or less accurately for what about fifty years.

It did change a lot of things about the world, and it took computers from being these large, slow, you know, refrigerator sized things that live in mainframe rooms at corporations to you know, tiny little things that live in our pockets, are on our wrists and have much more power than all of the main frames that existed, you know in the nineteen seventies combined, right, caused all sorts of changes in our society, some for the better, so much for the worse.

But you know, living through that, it seemed like clockwork, right, you know.

I mean I only lived through like the last part of it, But I remember when I was a kid, it seemed like, you know, computers were just always getting smaller and faster and better every single year, and you could just get you know, the advice was wait as long as you can to get a new computer, because the longer you wait, the better it'll be.

Right, And it was this amazing thing, and it made a lot of people a lot of money, and a few people truly enormous amounts of money.

And so you put all of that together and it's it kind of makes some sense that some people would take it extremely seriously as this general thing, because it seemed to be, you know, if you lived a comfortable middle or upper class life, it seemed like the most important thing in the world.

In the late twentieth century, right, and it wasn't really like anything that you'd seen before.

It was easy to think, oh, this is just going to continue.

So Ray Kurzweil is this inventor and futurist who you know.

He made like real serious contributions to text to speech technology and like assistive devices for the visually impaired, and I think hearing impaired as well.

You know, he made serious contributions to the field of electronic instruments, like you know, musical instruments.

But he is best known as a futurist.

He is best known as somebody who you know, makes these forecasts about what the future is going to be like.

So he's a retired electrical engineer, you're saying, essentially, yeah, I have a lot of those in my inbox.

Yeah, me too, man.

I'm pretty sure that if you put anywhere on the Internet that you have a PhD in physics, you get a lot of retired electrical engineers in your inbox.

Guys, my inbox has pictures of feces from people who want to note the parasite infections.

I'm feeling pretty low on sympathy right now.

Oh, but I once got so sorry for you guys.

Yeah, no, we should.

We should have a separate episode just talking about what's in our inboxes, because I have some crazy stuff in any event.

All right, So you're telling us how Ray Kurzweild was thinking about how Moore's law is transforming technology and that's the engine of transformation of society and predicting the future of society more broadly exactly.

Yeah, And like Kurzweil extends Moore's Law in his Forecasts of the Future and says, oh, this is part of a more general trend in the history of technology and the history of you know, even life in the universe.

And he calls it the law of accelerating returns, where he says, you know, once you have better technology, it's going to allow you to make the technology that you've already got even better, and then that'll just be a self reinforcing cycle that leads to this exponential trend.

And More's law is just one manifestation of that trend, and it's going to you know, he says, it's something that you can see if you look back through the entire history not just of human technology, but evolution of life on Earth, because you see the same thing with biological quote unquote technology, and he says, you know, this is going to continue, and in short order we are going to reach this point that he calls the singularity, which is where you've got, you know, technology that has developed to such an advanced degree that it gives us, you know, godlike powers of creation and destruction and transform and just changes the fundamental nature of life on Earth and in the universe.

Well, a lot of what you said sounds reasonable.

R There is evolution, and there is transformation, and things are changing more rapidly.

But from reading your book and from your tone, I'm guessing you don't agree with Cursewil about singularity and how we're all going to be techno gods in the future.

Why not?

Why will Daniel not be a techno god?

Yeah?

I mean, look Daniel in particular, Yes.

Yeah, I have a personal stake in this question.

Now, yes, Daniel in particular, Yeah, you are not going to be a techno god, Daniel, because you know, by having me on this podcast, Ray Kurzwild is going to put you on his list, and then you know he won't allow you to ascend to God.

I knew this was a mistake.

Yeah, exactly, worse than trying to fight a land born Asia.

Huh, yes, exactly.

Yeah, No, that's number two.

Now number one is inviting Adam Becker onto your podcast.

But I mean, look, Curzwild is taking this exponential trend and just extending it out into the future and saying it's going to last forever.

And the one thing that's always true about exponential trends is that they end.

Right.

If you see any sort of exponential trend in nature or in you know, technology or whatever, your first thought should be, oh, that can't last, because it just doesn't.

There are not enough resources, there's not enough space, there's not enough anything to allow exponential trends in general to continue forever.

One of the examples that Churzwild gives in his book The Singularity Is Near which is probably his most famous book from about two thousand and five.

Doesn't he have a few books like The Singularity is Near, The Singularity is near Er, the Singularity is near Ish?

Yeah, yeah, yeah.

The Singularity is Near Er came out last year, and when I tell people that that's the title, they usually don't believe me.

But that is actually the title.

He wrote a book called the singularity is nearer.

Next year, the singularity is near your ear.

Yeah, neariest.

But the classic example in biology of exponential growth is something like bacterial growth in a petri dish and yeah, if you chart the number of bacteria in this you know, nutrient rich medium over time, Yeah, it grows exponentially until they fill the dish and eat all of the agar and then they die.

To try to play Devil's advocates, so when I was talking to space settlement folks, they would say something like, you know, the reason we need to go into space is because exponential growth does end at some point.

But our species is so amazing that we can see when we're getting close to the like asymptope and the exponential curve, and so we can go out to space and get resources and we can be more proactive about it.

What is wrong about that argument?

Yeah?

I mean where to start?

Ye?

What you got to pick somewhere?

Okay, I'm going to pick on you know, I'm going to do what we should all strive to do, or what I strive to do, and punch up right, I'm going to pick on somebody bigger than me.

Jeff Bezos makes the same argument, right, Yeah.

Jeff Bezos says that we need to go out into space because of exactly this.

He says, you know, we are using exponentially more energy as time goes on, and if that trend continues as it has for decades, if not centuries, then in about two three hundred years, we're going to be using all of the energy on Earth that we get from the Sun and will have used up all of the non renewable resources.

And so at that point we need to go out into space, if not before, then otherwise we're going to have what he calls a civilization of stasis and rationing.

And you know, he's not wrong about the first part.

If somehow we continue that exponential trend in energy usage, then yeah, and I think it's in about three four hundred years, we'd be using all of the energy available to us on Earth.

And also we'd be using so much energy that like the waste heat from our energy usage would like boil off the oceans.

We can't we can't do that, right, it's not possible.

I mean, putting aside that, you know, it's it's implausible that that trend will continue.

The problem is with the second half, because yeah, Okay, we get like three to four hundred more years here on Earth if you continue that trend.

So Bezos says, we have to go out into space, and you know what he doesn't say is where where resources are unlimited, but you know he implies it.

The problem is that if you really want exponential growth to continue, going out into space, it doesn't actually help you that much if you're looking on a timescale of centuries, because if you do that, like about I think it's like one thousand years after we hit that point of using all of the sunlight that hits Earth, we get to a point where we're just using the entire energy output of the Sun.

And then if we spot Bezos and company a warp drive so they can go faster than the line to try to amass even more resources very very quickly outside of the Solar system, which we shouldn't spot them a warp drive.

There's no reason to think that you can build a warp drive, and a lot of reason to think that you can't.

But if we do spot them a warp drive, that only gets you like about another two thousand years before you're using all of the energy in the observable universe.

Wow, so you know there are limits, growth ends, and the fact is that you know, all of that is wildly implausible.

It's not like the lesson that I want people to take away from all of this is, oh, well, we better keep in mind that growth has to end at some point to the next like three thousand days years.

The answer is, oh, no, growth has to end a lot sooner than that, because you know, going out into space has lots of problems, even putting aside the lack of warp drive, just living in the Solar System is an extraordinarily difficult and dubious proposition.

To give Bezos a little bit of cris after ragging on him just now, one of the things I like that Jeff Bezos has said is he makes fun of Elon Musk for wanting to go to Mars because Mars sucks.

But Bezos's solution is.

You know, for going out into space, it's not considerably better, which is to build like hundreds of thousands or millions of enormous city size space stations and then have everybody live inside of them.

This is also not a great idea for many many reasons.

All right, So it's reasonable, I think to make these arguments against like the strongest version of those claims.

You know, exponential growth will last forever.

Sure, and you're right, that's obviously practical because the universe is finite, or the observable part of it is finite at least.

Yeah.

Yeah, But what if we just like water down those claims a little bit and we just say, you know, technology is transforming society very rapidly, and even the future you describe as refuting exponential growth.

That sounds pretty awesome.

Like if in two thousand years we're tapping into all the energy from the Sun and nearby stars and have an incredible, you know, star spanning civilization a lot of people out there, and be like, that sounds great.

What's wrong with that?

The prospect of large numbers of people living and working in space has an enormous number of technological and social and political questions tied to it that are very very difficult to solve and may not be solvable.

And some of those problems are sort of irreducibly time consuming.

You can't solve them without doing like lengthy experiments involving things like radiation exposure and low gravity exposure.

And things like that.

And I see Kelly nodding.

And you know, Kelly may know more about this than I do, because you know, this is one of the subjects in my book.

Kelly and Zach wrote an entire book about this, an excellent book that I really like.

I do always find a way to pull the conversation back to space settlement.

So sorry for derailing us, but you do a great chapter on it in your book.

Yeah, thank you, and you have nothing to apologize for it's in my book.

But let me maybe highlight a difference between the takes you guys have in your books.

Kelly and Zach say that you know, we're maybe not ready to settle space, that we haven't done the necessary legwork, and we shouldn't get over excited and jump too fast and send people to Mars now, because there's a lot of stuff we need to figure out, but that it's possible and if we do it right, maybe you could figure this out.

We just aren't there yet.

But I feel like your book goes a step further and suggests that you know it's dangerous to make these projections.

You know, somebody out there listening might say, all right, Adam, maybe we won't get there, you know, to as far as these guys project, but however far we get will be great.

What do you say to that person?

Is there a danger in this kind of thinking?

Yeah?

I mean this This gets back sort of to the last question that you asked me as well, because we don't know that it's possible to have large numbers of humans living off of Earth.

Because it's very possible that that's not, you know, something that we can do.

We need to find a way to live safely and healthily within the limits imposed by Earth.

We can't just assume that we're going to be able to leave.

The danger is that this rhetoric of oh, it's always going to be possible to expand out into space and grow forever can be used, and in fact it's not hypothetical.

It is being used to justify this sort of logic of rapacious consumption that is not sustainable here on Earth.

And because there's a very.

Good chance that we cannot in any meaningful way leave Earth, we need to stop doing that and find a way to live here.

That's not to say that we shouldn't explore space.

I think robots in space are amazing.

Like the voyager probes make me cry, you know, I'm a cosmologist by training.

I think getting data from space is really important and interesting.

I'm not even saying that we shouldn't send people into space to you know, the Apollo missions were amazing and really interesting.

They were, of course, you know, not primarily missions of scientific discovery.

It was about the Cold War.

But still like the fact that we did like a crude sample return mission to the Moon several times and nobody died is amazing.

But the visions that we have of the future are used to justify all sorts of things right here and now, and so we need to be careful about what we think the future is going to look like and whether that's remotely plausible.

And I really think that the things that Musk and Bezos and these other tech billionaires are talking about are sort of like saying, you know, yeah, well it's okay that we're doing what we're doing right now, because in the future we're all going to live in like Hogwarts and have broomsticks and magic wands and like, it's roughly the same level of plausibility.

And so to try to get us connecting More's law back with where we are in the conversation.

To me, I see the connection being that you've got this thinking that we're going to have exponential growth and our ability to do everything.

So like when I was talking to space settlement people, they'd be like, I'd talk about a problem and they'd say, well, AI is going to solve that everything is expanding.

Our ability to do anything related to technology keeps expanding exponentially, and so you know, we've talked about how we have limits and so you can't expect exponential trends to go on forever.

Do you connect then this kind of Moore's law thinking with techno optimism and this these sort of views of the future.

Or have we just gotten off on a different topic.

No, No, No, I think these things are connected, right, Like, there's a reason why all of these different things are.

In my book, one of the things that I like to remind people about when we're talking about Moore's law is that More's law.

It's not just that it's an empirical observation rather than a law of nature.

More's law was a decision.

More'slaw is a choice that the leaders of the semiconductor industry made, and then they continued making it for decades, you know, and there was a road map and lots and lots of different, you know, plans made in order to ensure the continuation of Moore's law for as long as possible.

There are massive, massive amounts of money and corporate resources poured into this, and in fact, Moore's law is not even an example of accelerating returns, as as Kurzweil would have, but in a sense it's an example of diminishing returns because they got, you know, the semiconductor industry got less bang for their buck over time.

They had to spend more and more money, even adjusting for inflation, just to get the same doubling of the number of processors crammed into the same space.

The techno utopian sort of ideas that Kurzweil pushes using Moore's law, as you know, sort of the justification and this you know eternal expansion into space stuff that we've just been talking about, they all sort of traffic in the idea that the future of technology is not just eternal exponential growth and expansion, but that it's inevitably that not that that's you know, something that we could do, but that it's it's what we have to do.

It's what is going to happen, and the only alternative, if there is one, is the extinction of the species.

And you know, again Musk is extremely clear about this.

Musk has said the only choice we have is eternal expansion out into cosmos or extinction.

And when he's pushed on this, he, you know, he brings up the fact that you know, in about half a billion or a billion years, it's going to get so hot on Earth because of you know, the sun getting hotter, that the oceans will boil off.

And yeah, that's not wrong, but you know, a lot's gonna happen between now and then.

Not only is it not a particularly pressing problem, but it may not even end up being our problem at all, because there are many other things that could cause humanity to go extinct between now and then, like say, civilizational collapse due to global warming, for example, a problem that tech oligarchs and other billionaires have done a lot of work to try to prevent humanity from solving.

But instead Musk says that the solution is to leave Earth.

And this is the sort of rhetoric that I was talking about, you know, in terms of like, this is what this eternal expansion idea gets you.

But it's also I think part of the connection with the logic of taking Moore's law as this law of nature that we can always count on these exponential trends, and we can always count on human ingenuity and technical knowledge and know how to get us out of any problem.

If you believe that.

Account for all of the problems in the world today, Like, there's so many problems that we have that are not amenable to technological solutions that people have tried to solve for a long time, that are fundamentally social in nature, or you know, had a technological component but also have a social component, like climate change.

Right, we have a lot, if not all, of the technology that we need to address climate change, but we haven't yet as a species, and that's primarily a social and political issue, not an issue of technology.

So, to paraphrase your argument, I think you're saying, it's not that computers won't get faster and the technology can't help us in the future.

It's just that we can't rely on it always doing so to magically solve all of our problems, and doing so distract ourselves from the real problems we face in the more immediate future.

Yeah, yeah, I mean, also, more's laws over.

I mean, come on, we have the transistors down about as small as we can get them.

You know, you can't make a silicon transistor smaller than an atom of silicon.

But you do see a role for technology in shaping our future.

I mean, it's not that you don't want chet gpt to cure cancer.

I definitely hear that there's a role for technology in shaping our future.

Technology is a big part of how we shape our future.

I'm going to just pretend that you didn't say the thing about chatch ept curing cancer.

Though.

God, there's this tweet, like one of my favorite tweet and responses, effor is where Sam Altman said something like be me, build chatch ept to cure cancer or whatever, and then people start criticizing you, and then he like goes on and on and like has a pity party for himself.

And then somebody just responded with did you cure cancer or whatever?

But you know, there has been actually great progress made and treating cancer just in the last few years, right, you know, like these I don't remember the names of the drugs because like I'm not a cancer guy, but like these approaches of like getting cancer patient's own immune systems to properly recognize and attack the cancers in their own bodies has been like incredibly successful and is really promising for further developments.

It's really amazing, and like there have been all sorts of really amazing biomedical advances that are currently being destroyed by RFK Junior and Trump.

Like mRNA vaccines are one of the great success stories of you know, biomedical science in the last twenty years.

And I think that's important, and I think in general, vaccines are great.

You know, there's all sorts of really wonderful technology that we've created that has made the world generally a better place, or has at least enabled.

People to make the world a better place.

Right.

In general, technology is a tool, and there are questions about how you use it, right.

You know, nuclear power can be used to build nuclear power plants, but it can also be used to make bombs.

YadA, YadA, YadA.

I think I just YadA YadA, nuclear apocalypse.

But yeah, you did, Yeah whatever.

I'm a physicist.

Of course that's what I'm gonna do.

But the point is, yeah, of course there's a role for technology to play in shaping our future.

It's just not two things.

Technology is not the only thing that shapes our future, and the development and future direction of technology is not inevitable.

Technology is something that humans make, and the future development of technology is filled with contingency and human choice.

It is not like we build every single technology that it is physically possible to build.

It's not on rails.

It's not like it's you know, the analogy I make in the book, It's not like a tech tree in civilization, right, where like the future of technology is just sort of revealed to us and we have we just make a choice about which branch we're going to pursue first.

That's not how anything works.

All right.

Well, thanks Adam for coming on, and let's hope that chat GBT desk your cancer before any of us get it.

Thanks for being on the show, Adam absolutely.

Daniel and Kelly's Extraordinary Universe is produced by iHeartRadio.

We would love to hear from you, We really would.

We want to know what questions you have about this Extraordinary Universe.

We want to know your thoughts on recent shows, suggestions for future shows.

If you contact us, we will get back to you.

We really mean it.

We answer every message.

Email us at questions at danieland Kelly.

Dot org, or you can find us on social media.

We have accounts on x, Instagram, Blue Sky and on all of those platforms.

You can find us at d and K Universe.

Don't be shy, write to us