To date, a handful of large companies have captured the value created by advances in AI. a16z general partner Martin Casado explains why with generative AI, that’s changing. 

[0:00] AI economics before
[2:30] The AI mediocrity spiral
[6:23] The economics of Generative AI
[11:53] The third epoch of compute
[13:35] Jevon's Paradox 

This conversation is part of our AI Revolution series, which features some of the most impactful builders in the field of AI discussing and debating where we are, where we’re going, and the big open questions in AI. Find more content from our AI Revolution series on www.a16z.com/AIRevolution.

We should all get ready for a new wave of iconic companies.

I don't think we know what they're going to look like,

but the economics are just too compelling.

Okay, so what has the narrative been for AI over the last 50 years?

The narrative is this episodic thing with summers and winters and all of these false promises.

I remember when I joined PhD in 2003, for my cohort that joined, I would say 50% of the people were doing AI.

This was when Bayesian stuff was super popular.

And then within 3 years, everybody was like, AI is dead.

And so it's kind of been, We've had this love-hate relationship with it for a very long time.

But if you look at all of the graphs, we've made

a tremendous amount of progress in the last 70 years.

Like way back in the 60s, we were doing expert systems, which are still used for diagnosis, right?

Like We're very good at beating Russians at chess.

We're doing self-driving cars, we're doing vision.

There's just a lot of stuff that we've solved.

And so much so it's become a cliche that every time we solve a problem, we're like, oh, well, that wasn't really AI, right?

So we've had steady progress, we've solved real problems.

And not only that, it's been a while now that we've been better at humans for some very, many very important things, for example, like, you know, perception or handwriting detection.

It's been about 10 years since we've been better than humans at entity identification.

And Not only that, we've actually gotten very good at monetizing this, particularly for large companies.

And so, as we all know, there's been a ton of market cap that's been added to companies like Meta and Google and Netflix by using AI.

So I think the question we should all ask ourselves is why hasn't this resulted in an actual platform shift?

And by platform shift, I mean, why has the value accrued to the incumbents?

a new set of kind of AI native companies that have come up and displaced them, which we've seen in many other areas, right?

We saw that in mobile, as Sarah mentioned, we saw that with the internet, obviously we saw it with the microchip, et cetera.

And for the first part of this talk, what I'm gonna argue is that the capabilities have all been there, but the economics just haven't for startups.

So if you step back and you look at the standalone case

what a big company can extract from it, but like a startup, it's actually not been great.

I mean, to begin with a lot of the sexier use cases are pretty niche markets, right?

Maybe it's a useful tool that you can apply to solving bigger problems, but that in

I actually think the second point is the most important point and it's pretty subtle.

Many of the traditional use cases of AI require correctness in the tail of the solution space.

1 of the reasons is if you have to be correct, and you've got a very long and fat tail, like, either you do all of the work technically, or you hire people.

For start-ups to start hiring people to provide solutions is a variable cost.

And the second 1 is because the tails of these solutions tend to be so long, think something like self-driving where there's so many exceptions that could possibly happen, the amount of investment to stay ahead, like, increases and

So we've actually done studies on this, and it turns out many companies that try to do this as startups end

They're lower margins, and they're just much harder to scale.

Of course, with robotics comes the curse of hardware, classically a very difficult thing for startups to do.

And if you really think like, what is the competition of most use cases of AI, not like traditional machine learning, but AI, it tends to be the human.

And traditionally, it's stuff like the human brain is really good at, like perception, right?

Like the brains that we have have evolved over a hundred million years to do things like whatever, pick berries and evade lions or whatever it is.

And it's incredibly efficient at doing that.

So this leads to something that most investors know, which we call the dreaded AI mediocrity spiral.

It's very simple, which is, Let's say a founder comes in and they want to do an AI company and they're going to use AI to automate a bunch of stuff.

Of course, correctness is really important and they want it to look good at first, so they hire people to do it instead of the AI.

Then they come to us, we invest in them, and I join the board.

Then I say, listen, this is great, you need to grow.

And they're like, oh man, we need to grow, this AI

Now you're on this treadmill of continuing to invest, hiring people, like the automation doesn't often happen, and if it does, it's only part of the solution.

This is 1 of the reasons why so many startups that

have tried to do this just haven't had kind of this breakaway economics and

that can actually sink these perverse economies of scale.

I think 1 of the great examples, just because it's illustrative of

But we've been working at this for decades.

We've invested $75 billion, and the unit economics are still not on par with a human Uber driver.

And so it's remained in the domain of large companies.

And if you actually, the slide that I had previously, like other than the niche markets, all of these problems apply, right?

I mean, there's hardware you've got to deal with, it's incredibly long tail with RoboTaxi.

Even if you strip away everything and you just go to like, here's the processing unit that does self-driving today on the high-end systems, the amount of power that they consume, you know, the amount of hardware that's required, even then it's only like maybe a factor of 10 better than a human being, and the power is much, much higher, right?

The economics aren't that compelling, and certainly not the type of economics you see in the case of market transformations.

Like market transformations aren't created when

They get created when they're 10,000 times better.

So what is the learning from the last, say, 70 years?

It's not that the technology doesn't work.

It's not that we can't solve the problems.

Big companies are great at monetizing it.

It's that it's very, very hard for startups to break away.

And if startups can't break away, you don't get a transformation.

And I'm going to try to be pretty precise.

That model will take a natural language like English and will output something.

It would be a conversation, an image, a 3D model, music, whatever.

And as we know, this has been pushed into many different areas where we're already seeing productive viable businesses.

There's creativity, like any component of a video game, you can automatically generate.

There's companionship, which kind of more of an emotional connection, so that you have like characters to talk to, so it fills a social role.

And then there's a class that we call co-pilot, which is effectively it'll help you with tasks, whether there's a worst test or personal tasks.

So these are already emerging as independent classes.

So remember the slide that I showed you before of like the properties of AI previously that made it difficult to build a startup company.

So none of these really apply to this current model.

So the first 1, obviously, these are large markets that this is being applying to.

It's like arguably all of white collar work, even just like video games and movies

The second 1, again, I think is, you know, the most important point and maybe the most subtle.

In this domain, correctness isn't as much of an issue for 2 reasons.

1 of them is when you're talking about creativity, the first C, there is no formal notion of correctness, really.

I mean, what does it mean to be incorrect for like a fiction story

I mean, for sure, like you want to make sure like they have all their fingers, but even then, do you really?

And so we have absolutely adapted to use cases where,

you know, correctness is not a huge issue.

The second 1 is a little more subtle, but I just think it's so important, which is the behaviour that's developed around these things is iterative.

And so that human in the loop that used to be in a central company is now the user.

So it's not a variable cost to the business anymore.

And as a result, you can do things where correctness is important.

Like for example, developing code because it's iterative.

So, like, the amount of error that our crews gets less because it's a smaller subset of it.

But also, like, you're constantly getting feedback and correction from the user.

that we talked about, you know, clearly mostly software based, at least for now.

This does apply to robotics, but that's not what's really taking off.

And I want to talk to you about this brain portion, because I think it's so interesting.

I'm not a neuroscientist, but I think this is very interesting, which is for these types of tasks, the silicon stack is way better than the carbon stack.

So if you think about it, like Traditionally, a lot of it is doing stuff like the 100 million year old brain is doing, right?

The 1 that's been fleeing predators or picking strawberries or whatever it is.

And that's very, very hard to compete with.

Remember, if you have the CPU, GPU set up a self-driving car, Some of these kits are like, you know, 1.3 kilowatts, where the human brain is 15 watts.

So economically, that's very tough to compete with.

The new gen iWave, it's kind of competing with like the creative language center of the brain.

And it turns out it's incredibly competitive, so much so that you actually have the economic inflection we look for for a market transformation.

So let's just go ahead and break down the numbers very quickly.

So let's say that I, Martine, wanted to create an image of

So if I'm using 1 of these image models, the inference cost, let's call it a tenth of a penny, it's probably less than that, actually.

little bit more than that, but let's say that's how much

If you compare that to hiring a graphic artist, let's say that was 100 bucks in an hour.

a lot more money than that, but let's conservatively say that.

You've got 4 to 5 orders of magnitude difference in cost and time.

So these are the type of inflexes you look for, certainly as an economist, when

it's like there's gonna be actually a massive market dislocation.

I'll give you another example from Instabase.

So let's assume that you have a legal brief.

You throw it into this kind of unstructured document, LLM, and

then you ask questions for that legal brief.

Maybe it's a little more, maybe it's a little less.

Time to complete, maybe 1 second, maybe a little more, maybe a little less.

So, like, as someone who has actually spent a lot of money on lawyers hours, I

The first 1 is it takes more than 1 hour to iterate on this for sure.

And the second 1 is they're not always correct.

In fact, built in for any interaction I have with

a lawyer is cross-checking their work and double-checking their work.

So again, we have 4 to 5 orders of magnitude difference in cost and time.

And if you wanna have like an example of like how the extremely nutty that this can get.

Like I see no reason why you can't generate an entire game.

The 3D models, the characters, the voices, the music, the stories, et cetera.

Like there are companies that are doing all of

And if you compare the costs of like hundreds of millions of dollars and years versus, you know, a few dollars of inferences, now we have current internet and microchip level asymmetries in economics.

Now listen, I'm not saying this is gonna happen, I'm not saying this is gonna happen soon, we're not there yet.

What I'm saying is this is the path that we're on.

And these types of paths are what you look for with big transformations.

So it's little wonder why we're seeing so much takeoff the way that we have.

These are the fastest growing products and some of the fastest growing companies we've seen in

And it's because, again, it's less a capability, which we always focus on the capabilities, and it's much more that the economics work.

So whenever the marginal cost of something drops this much, the industry changes.

And by marginal cost, I mean, like for whatever good you're producing, to produce more of that good, that price converges on 0.

So listen, this may sound hyperbolic, but I really think that we could be entering a third epoch of compute.

I think that the first epoch, of course, is the microchip.

And so, and just because the numbers are interesting, at least to me.

And so, before the advent of the computer, you actually had people calculating logarithm tables by hand.

And that kind of ushered in the compute revolution, right?

number of companies that were either totally transformed like IBM or totally net new.

So the microchip brought the marginal cost of compute to 0.

The internet brought the marginal cost of distribution to 0.

So listen, in the 90s, when I wanted to get a new video game,

And if you look at the full cycle from that box, leaving whoever saw it, This is weeks.

And again, I don't have the math up here, but if you actually calculate the price per bit relative to DSL in the late

90s, it's about 4 or 5 orders of magnitude again, relative to actually shipping it.

So I think it's a pretty good analog where you say, these large models actually bring the marginal cost of creation

Some very fuzzy, vague notion of what creation means.

But for sure, we could talk about it of, like, content, conversation, whatever it is.

And like the previous epochs, When those epochs happened, you had no idea what new companies were going to be created, right?

You just knew something was going to happen, right?

So I think, listen, I think we should all get ready for a new wave of iconic companies.

we know what they're going to look like, but the economics are just too compelling.

There's always a question when you have market dislocations, like they're staring you in the face.

And Because this is an economic talk, I'm going to give an economic answer, which is there's something called Jevon's paradox.

It says if the demand is elastic, It turns out there's unlimited demand for compute.

Even if you drop the price, the demand will more than make up for it.

This is absolutely the case of the internet, right?

So you get kind of more value, more productivity, et cetera.

And I personally believe when it comes to creating any creative asset or any sort of kind of work automation Clearly the demand is elastic, right?


Speaker 0: We should all get ready for a new wave of iconic companies. I don't think we know what they're going to look like,

Speaker 1: but the economics are just too compelling.

Speaker 0: Okay, so what has the narrative been for AI over the last 50 years? The narrative is this episodic thing with summers and winters and all of these false promises. I remember when I joined PhD in 2003, for my cohort that joined, I would say 50% of the people were doing AI. This was when Bayesian stuff was super popular. And then within 3 years, everybody was like, AI is dead.

And so it's kind of been, We've had this love-hate relationship with it for a very long time. But if you look at all of the graphs, we've made

Speaker 1: a tremendous amount of progress in the last 70 years.

Speaker 0: And along the way, we've solved a

Speaker 1: lot of very real problems, right? Like way back in the 60s, we were doing expert systems, which are still used for diagnosis, right?

Speaker 0: Like We're very good at beating Russians at chess. We're doing self-driving cars, we're doing vision. There's just a lot of stuff that we've solved. And so much so it's become a cliche that every time we solve a problem, we're like, oh, well, that wasn't really AI, right?

Speaker 1: So we just keep moving the goalposts.

Speaker 0: So we've had steady progress, we've solved real problems. And not only that, it's been a while now that we've been better at humans for some very, many very important things, for example, like, you know, perception or handwriting detection. It's been about 10 years since we've been better than humans at entity identification. And Not only that, we've actually gotten very good at monetizing this, particularly for large companies. And so, as we all know, there's been a ton of market cap that's been added to companies like Meta and Google and Netflix by using AI.

So I think the question we should all ask ourselves is why hasn't this resulted in an actual platform shift? And by platform shift, I mean, why has the value accrued to the incumbents? And why hasn't there been

Speaker 1: a new set of kind of AI native companies that have come up and displaced them, which we've seen in many other areas, right? We saw that in mobile, as Sarah mentioned, we saw that with the internet, obviously we saw it with the microchip, et cetera.

Speaker 0: And for the first part of this talk, what I'm gonna argue is that the capabilities have all been there, but the economics just haven't for startups. So if you step back and you look at the standalone case

Speaker 1: for AI economics, not like

Speaker 0: what a big company can extract from it, but like a startup, it's actually not been great.

Speaker 1: I mean, to begin with a lot of the sexier use cases are pretty niche markets, right? Like, you know, it's great

Speaker 0: to beat Russians at chess. Like that's not a market. Maybe it's a useful tool that you can apply to solving bigger problems, but that in

Speaker 1: and of itself is a market.

Speaker 0: I actually think the second point is the most important point and it's pretty subtle. Many of the traditional use cases of AI require correctness in the tail of the solution space. And That's a very hard thing for

Speaker 1: a start-up to do for

Speaker 0: a couple of reasons. 1 of the reasons is if you have to be correct, and you've got a very long and fat tail, like, either you do all of the work technically, or you hire people. So often, we hire people, right? For start-ups to start hiring people to provide solutions is a variable cost. And the second 1 is because the tails of these solutions tend to be so long, think something like self-driving where there's so many exceptions that could possibly happen, the amount of investment to stay ahead, like, increases and

Speaker 1: the value decreases, right? You have this perverse economy of scale.

Speaker 0: So we've actually done studies on this, and it turns out many companies that try to do this as startups end

Speaker 1: up with non-software-like margins. They're lower margins, and they're just much harder to scale.

Speaker 0: Of course, with robotics comes the curse of hardware, classically a very difficult thing for startups to do. And if you really think like, what is the competition of most use cases of AI, not like traditional machine learning, but AI, it tends to be the human. And traditionally, it's stuff like the human brain is really good at, like perception, right? Like the brains that we have have evolved over a hundred million years to do things like whatever, pick berries and evade lions or whatever it is. And it's incredibly efficient at doing that.

So this leads to something that most investors know, which we call the dreaded AI mediocrity spiral. And what is it? It's very simple, which is, Let's say a founder comes in and they want to do an AI company and they're going to use AI to automate a bunch of stuff. Of course, correctness is really important and they want it to look good at first, so they hire people to do it instead of the AI. Then they come to us, we invest in them, and I join the board.

Then I say, listen, this is great, you need to grow. And they're like, oh man, we need to grow, this AI

Speaker 1: is hard, the tail is long,

Speaker 0: I'm going to hire more people. Now you're on this treadmill of continuing to invest, hiring people, like the automation doesn't often happen, and if it does, it's only part of the solution. This is 1 of the reasons why so many startups that

Speaker 1: have tried to do this just haven't had kind of this breakaway economics and

Speaker 0: the value accrues to large companies

Speaker 1: that can actually sink these perverse economies of scale.

Speaker 0: I think 1 of the great examples, just because it's illustrative of

Speaker 1: so many of these things is RoboTaxi.

Speaker 0: RoboTaxi is fantastic. It's great value. They're driving outside of here. But we've been working at this for decades. We've invested $75 billion, and the unit economics are still not on par with a human Uber driver.

And so it's remained in the domain of large companies. And if you actually, the slide that I had previously, like other than the niche markets, all of these problems apply, right? I mean, there's hardware you've got to deal with, it's incredibly long tail with RoboTaxi. Human brains are very good at this. Even if you strip away everything and you just go to like, here's the processing unit that does self-driving today on the high-end systems, the amount of power that they consume, you know, the amount of hardware that's required, even then it's only like maybe a factor of 10 better than a human being, and the power is much, much higher, right?

The economics aren't that compelling, and certainly not the type of economics you see in the case of market transformations. Like market transformations aren't created when

Speaker 1: the economics get 10 times better. They get created when they're 10,000 times better.

Speaker 0: So what is the learning from the last, say, 70 years? It's not that the technology doesn't work. It's not that we can't solve the problems. All of that has always been the case. It's not even that we can't monetize it.

Speaker 1: Big companies are great at monetizing it.

Speaker 0: It's that it's very, very hard for startups to break away. And if startups can't break away, you don't get a transformation. Okay. So, why are we all here? So, this wave is very, very different.

And I'm going to try to be pretty precise. So, you have a model. That model will take a natural language like English and will output something.

Speaker 1: It would be a conversation, an image, a 3D model, music, whatever.

Speaker 0: And as we know, this has been pushed into many different areas where we're already seeing productive viable businesses. I like to call them the 3 Cs. There's creativity, like any component of a video game, you can automatically generate. There's companionship, which kind of more of an emotional connection, so that you have like characters to talk to, so it fills a social role. And then there's a class that we call co-pilot, which is effectively it'll help you with tasks, whether there's a worst test or personal tasks.

So these are already emerging as independent classes. So remember the slide that I showed you before of like the properties of AI previously that made it difficult to build a startup company. So none of these really apply to this current model. So let's go through them.

Speaker 1: So the first 1, obviously, these are large markets that this is being applying to.

Speaker 0: It's like arguably all of white collar work, even just like video games and movies

Speaker 1: is like $300 billion market. These are massive, massive markets.

Speaker 0: The second 1, again, I think is, you know, the most important point and maybe the most subtle. In this domain, correctness isn't as much of an issue for 2 reasons. 1 of them is when you're talking about creativity, the first C, there is no formal notion of correctness, really. I mean, what does it mean to be incorrect for like a fiction story

Speaker 1: or a video game?

Speaker 0: I mean, for sure, like you want to make sure like they have all their fingers, but even then, do you really? In sci-fi? And so we have absolutely adapted to use cases where,

Speaker 1: you know, correctness is not a huge issue.

Speaker 0: The second 1 is a little more subtle, but I just think it's so important, which is the behaviour that's developed around these things is iterative. And so that human in the loop that used to be in a central company is now the user. So it's not a variable cost to the business anymore. The human in the loop has moved out. And as a result, you can do things where correctness is important.

Like for example, developing code because it's iterative. So, like, the amount of error that our crews gets less because it's a smaller subset of it. But also, like, you're constantly getting feedback and correction from the user. The primary use cases

Speaker 1: that we talked about, you know, clearly mostly software based, at least for now. This does apply to robotics, but that's not what's really taking off.

Speaker 0: And I want to talk to you about this brain portion, because I think it's so interesting. I'm not a neuroscientist, but I think this is very interesting, which is for these types of tasks, the silicon stack is way better than the carbon stack. So if you think about it, like Traditionally, a lot of it is doing stuff like the 100 million year old brain is doing, right? The 1 that's been fleeing predators or picking strawberries or whatever it is. And that's very, very hard to compete with.

Remember, if you have the CPU, GPU set up a self-driving car, Some of these kits are like, you know, 1.3 kilowatts, where the human brain is 15 watts. So economically, that's very tough to compete with. The new gen iWave, it's kind of competing with like the creative language center of the brain. It's like 50K years old. Like it's much less evolved.

And it turns out it's incredibly competitive, so much so that you actually have the economic inflection we look for for a market transformation. So let's just go ahead and break down the numbers very quickly. So let's say that I, Martine, wanted to create an image of

Speaker 1: myself as a Pixar character, right?

Speaker 0: So if I'm using 1 of these image models, the inference cost, let's call it a tenth of a penny, it's probably less than that, actually. Let's say it takes 1 second. Maybe it's a

Speaker 1: little bit more than that, but let's say that's how much

Speaker 0: it costs. If you compare that to hiring a graphic artist, let's say that was 100 bucks in an hour. I've actually hired graphic artists to

Speaker 1: do things like this. It tends to be

Speaker 0: a lot more money than that, but let's conservatively say that. You've got 4 to 5 orders of magnitude difference in cost and time. So these are the type of inflexes you look for, certainly as an economist, when

Speaker 1: it's like there's gonna be actually a massive market dislocation.

Speaker 0: I'll give you another example from Instabase. So let's assume that you have a legal brief. It's in a PDF. You throw it into this kind of unstructured document, LLM, and

Speaker 1: then you ask questions for that legal brief.

Speaker 0: Again, the inference cost, say, a tenth

Speaker 1: of a penny. Maybe it's a little more, maybe it's a little less. Time to complete, maybe 1 second, maybe a little more, maybe a little less. Time to complete, maybe 1 second, maybe a little more, maybe a little less. But again, I

Speaker 0: mean, it's relatively cheap. Relative to the cost of a lawyer. So, like, as someone who has actually spent a lot of money on lawyers hours, I

Speaker 1: want to point out a couple of things.

Speaker 0: The first 1 is it takes more than 1 hour to iterate on this for sure. And the second 1 is they're not always correct. In fact, built in for any interaction I have with

Speaker 1: a lawyer is cross-checking their work and double-checking their work.

Speaker 0: So again, we have 4 to 5 orders of magnitude difference in cost and time. And if you wanna have like an example of like how the extremely nutty that this can get. Like I see no reason why you can't generate an entire game. There are companies today working on it.

Speaker 1: The 3D models, the characters, the voices, the music, the stories, et cetera. Like there are companies that are doing all of

Speaker 0: these things. And if you compare the costs of like hundreds of millions of dollars and years versus, you know, a few dollars of inferences, now we have current internet and microchip level asymmetries in economics. Now listen, I'm not saying this is gonna happen, I'm not saying this is gonna happen soon, we're not there yet. What I'm saying is this is the path that we're on. And these types of paths are what you look for with big transformations.

So it's little wonder why we're seeing so much takeoff the way that we have. And these are

Speaker 1: the fastest going open source projects.

Speaker 0: These are the fastest growing products and some of the fastest growing companies we've seen in

Speaker 1: the history of the industry.

Speaker 0: And it's because, again, it's less a capability, which we always focus on the capabilities, and it's much more that the economics work. So whenever the marginal cost of something drops this much, the industry changes. And by marginal cost, I mean, like for whatever good you're producing, to produce more of that good, that price converges on 0. So listen, this may sound hyperbolic, but I really think that we could be entering a third epoch of compute. I think that the first epoch, of course, is the microchip.

And so, and just because the numbers are interesting, at least to me. And so, before the advent of the computer, you actually had people calculating logarithm tables by hand. Like that's where the word comes from. They were computers, they would compute. Then we created ENIAC, along

Speaker 1: with other machines, but let's look

Speaker 0: at ENIAC. So ENIAC was 5,000 times faster than

Speaker 1: a human being doing it. Right? There's your 3 to 4 orders of magnitude.

Speaker 0: And that kind of ushered in the compute revolution, right?

Speaker 1: And this gave us a

Speaker 0: number of companies that were either totally transformed like IBM or totally net new. So the internet is something. So the microchip brought the marginal cost of compute to 0. The internet brought the marginal cost of distribution to 0. So listen, in the 90s, when I wanted to get a new video game,

Speaker 1: I would go to a store and buy a box.

Speaker 0: And if you look at the full cycle from that box, leaving whoever saw it, This is weeks. And again, I don't have the math up here, but if you actually calculate the price per bit relative to DSL in the late

Speaker 1: 90s, it's about 4 or 5 orders of magnitude again, relative to actually shipping it.

Speaker 0: So I think it's a pretty good analog where you say, these large models actually bring the marginal cost of creation

Speaker 1: to 0. Some very fuzzy, vague notion of what creation means. But for sure, we could talk about it of, like, content, conversation, whatever it is.

Speaker 0: And like the previous epochs, When those epochs happened, you had no idea what new companies were going to be created, right?

Speaker 1: You just knew something was going to happen, right? Like nobody predicted Amazon. Nobody predicted Yahoo. Like I remember when this happened.

Speaker 0: So I think, listen, I think we should all get ready for a new wave of iconic companies. I don't think

Speaker 1: we know what they're going to look like, but the economics are just too compelling. Forget the capability. Economics are just too compelling.

Speaker 0: So I have 1 final point. There's always a question when you have market dislocations, like they're staring you in the face. You know it's coming. What happens to the value? What happens to the jobs?

What happens to the people? Right? And Because this is an economic talk, I'm going to give an economic answer, which is there's something called Jevon's paradox. It's very simple. It says if the demand is elastic, It turns out there's unlimited demand for compute.

Even if you drop the price, the demand will more than make up for it. Normally far more than make up for it. This is absolutely the case of the internet, right? So you get kind of more value, more productivity, et cetera. And I personally believe when it comes to creating any creative asset or any sort of kind of work automation Clearly the demand is elastic, right?

Speaker 1: I think the more that we make that the more people consume and so

Speaker 0: I think that we're Very much looking forward to massive expanse of in productivity a lot of new jobs a lot of new things I

Speaker 1: think it's gonna follow just like

Speaker 0: the microchip and

Speaker 1: just like the internet.

Speaker 0: And so with that, I welcome you all here today. I think there's a lot of work for all of us

Speaker 1: to do, and thank you so much for taking the time. Thank you. Thank you. Thank you. Thank you.

See all a16z transcripts on Youtube

The Economic Case for Generative AI with a16z's Martin Casado