AI transcript
– Hi everyone, welcome to the A6NZ podcast.
I’m Sonal, happy new year.
Today’s episode is on why we should be optimistic
about the future, because it features two
of the most optimistic people together in conversation.
A6NZ co-founder, Mark Andreessen,
is interviewed by Kevin Kelly,
founding executive editor of Wired Magazine and more.
The conversation originally took place
at our most recent annual innovation conference,
the A6NZ Summit, and it was also previously released
on YouTube if you’d like to check it out there
as well.
– Good afternoon.
Thank you, Mark, for answering some questions.
I have a bunch of questions,
which I hope that we can talk about.
These all have to do about the future, where we’re going.
I want to start with a question about the past.
You know, a generation ago,
a lot of smart people didn’t think
the internet was gonna work,
and therefore they were unprepared for its benefits.
What are we smart people today not prepared for?
– Yeah, so you may remember actually generating,
it wasn’t even just that a lot of people thought
that the internet was gonna work,
a lot of smart people didn’t think that.
In fact, the inventor.
(laughing)
I can’t resist.
I can’t resist on the story.
They actually, the inventor of Ethernet,
which is a foundational technology for the internet,
spent the ’90s actually predicting the internet
would crash, would collapse,
and what we call it would be the gigalapse,
would take down the internet by like 1996, 1997.
He wrote a column at the time for a magazine
called Info World, and he said that if he was wrong,
by, I think it was like,
if the internet hadn’t collapsed by 1997,
he would eat his column.
And to his enormous credit in 1998,
he actually went on stage at a conference,
he actually ripped his column out of the magazine,
he put it in a blender with water,
blended it up, and he drank it on stage.
So it’s one of the more shining examples
of intellectual honesty I’ve ever seen.
As it turns out, he was wrong.
It turns out the internet did work.
So I think the big thing,
I’ve been thinking about this a lot,
you know, it feels to a lot of people
like things are getting strange.
And maybe I’m the only one who feels that way,
but if you read the news,
or just track things happening in the world,
just things feel kind of weird and different
over the last few years.
I actually think there’s like,
there’s that actual generational thing that’s happening,
and you alluded to the generational component.
Like it did take 25 years to get everybody online.
And like we’re not quite there yet,
but we’re getting very close.
Like I think the most exciting thing happened
in the world right now is Mukesh Ambani,
who’s the richest man in India,
has this program called Geo,
where he is literally providing internet access
to the 500 million lowest income Indians,
like literally it’s like free for six months.
And then it’s like a dollar a month.
It’s like the most amazing thing.
And it’s like, it’s working incredibly well.
And so we are very, very close to every,
at least every adult on the planet being internet connected.
But it took 25 years to get there.
And so for me, it’s like, okay, so then what?
One interpretation of that is, okay, we’re done.
We did it.
The other interpretation of that is actually,
okay, that’s just the beginning point.
– Right.
– That’s like the beginning point of what?
– Right.
– And I think it’s the beginning point of like, okay,
like what if you actually interconnect
with everybody on the planet?
Like what, you know, there’s like the metaphor
of the global mind of the global brain.
Like what if you actually connected everybody together
and let everybody find out what everybody else was thinking?
It’s one of those things that people think
sounds good.
And then they encounter it face to face
and they’re like, I don’t know.
– Right, right.
That was like, during my time,
that wired people were kind of concerned
about the digital divide.
And I said, the digital divide is going to cure itself.
The thing you should be worried about
is what happens when everybody is online?
So you think we’re not prepared
for what will happen when everybody is online?
– No, and I think we’re not prepared.
And then I think it’s going to be very exciting.
I mean, I think we’re already seeing that in many ways.
I think the, and then I think we’ve kind of figured out
collectively that it’s going to be different.
And so the initial impulse to say things
are going to get much worse.
And I don’t think that’s right.
I think things are going to get very different.
I think things will be much more positive.
And we’ll talk a lot about that today, hopefully.
But things are definitely going to be different.
– I think one lens that I’ve been trying to put on lately
is kind of think about it through a cultural lens.
Right, sort of what happens to culture
because culture, you know, Ben just wrote this book
about culture being kind of the foundation of the behavior.
And I think that’s really true certainly in companies,
but I think it’s also true in countries and globally.
And it feels like the internet’s impact on culture
is just beginning in the sense of like a world
in which culture is based on the internet,
which is what I think is happening.
It’s just at the very start, right?
‘Cause it had to get universal
before it could set the culture,
but that’s actually happening now.
– Okay.
And at the same time, a generation ago,
well, there was a few people
who actually did think the internet was going to work,
but they were also, like myself, expecting VR
and conversational AI to happen tomorrow.
So what are we expecting to happen now
that it’s not gonna happen?
– Yep, so I object to the question.
(laughing)
Your Honor.
So this is one of those things in our business
that you deal with a lot,
which is ’cause you find yourself,
these entrepreneurs come in and they pitch an idea
and you kind of feel like you should draw judgment
on whether the idea is gonna work or not.
And it’s something I’m really leery of doing anymore.
And the reason for that,
and I think you know this from all of your reading,
every successful technology that I’m aware of,
the things that are like all of a sudden,
like the next big thing, like the iPhone in 2007,
or just as an example,
they all have this like incredible 25 or 40
or 50 year backstory to them.
And you sometimes have to go back and excavate, right?
Because you haven’t heard a lot of the backstory
as the previous efforts failed, right?
But if you go back and look,
like there’s often a multi-generational run-up,
and so I’ll just give you a few of my favorite examples.
So iPhone hit big in 2007.
I for years went around saying,
well, IBM is, there was a 20 year project,
IBM shipped the first smartphone in 1997 called the Simon.
I thought that was true.
It actually turns out it’s not true.
I found the other day, RadioShack had a smartphone in 1982
with their, they literally had a phone version
of their TRS-80 mini computer.
They sold about four of them.
But it was a thing, right?
So that had a 25 year fuse on it.
Video conferencing, video conferencing goes back
at least to the mid ’60s, to the World’s Fair.
The telegraph was invented in the 1870s,
and then sat on a shelf for 100 years
before the Japanese turned it into an industry.
And then another favorite is fiber optics.
Nominally, or you can kind of stretch,
you could say fiber optics were invented in the 1840s.
Paris had a optical telegraph network under the city.
You could actually do, you could actually do under,
you could do telegraphy in the 1840s in Paris.
And it was literally, they were shining flashes of light
through glass tubes.
So there’s this like this incredibly rich back story
to all these things.
And so as a consequence, it’s actually less a question
of like, what’s the new idea?
It turns out the idea is probably already out there
somewhere. – Right, okay.
– And then it’s less the question of like,
is it going to work?
It’s more of the question of like,
when is it gonna work? – Right.
– And I pushed it so far, and people in our office
have heard this.
I pushed all the way to the point where I just think
we should assume that whatever
we’re being pitched is going to work.
(laughing)
It’s just a question of timing.
Then of course, timing turns out to be the hard part,
but it at least focuses the conversation.
– Right, right, right.
So it was the same idea of kind of looking
at the history of things.
One wonders, who really made all the money
when electricity came along?
It probably wasn’t the people
necessarily generating electricity.
Who do you think is gonna make the money
when AI comes along?
Is it the AI providers?
Is it the AI service?
Is it the algorithmic writers?
Who’s gonna be making money in AI?
– Yeah, so we think that there’s two
obvious business models, and probably others,
but the two obvious.
One is to be sort of a horizontal platform provider,
infrastructure provider, you know,
for AI kind of analogous to the operating system
or the database or the cloud.
You know, the other opportunity is kind of in,
would say in the verticals,
and so the applications of AI.
And there’s certainly a lot of those.
So that’s the general answer.
I think that the deeper answer is
there’s an underlying question
that I think is an even bigger question about AI
that reflects directly on this,
which is, is AI a feature or an architecture?
Is AI a feature?
We see this with pitches we get now,
which is just like we get the pitch,
and it’s like, here are the five things my product does,
right, and bullet points one, two, three, four, five,
and then oh yeah, number six is AI.
Right, and so you go, it’s always number six, right,
’cause it’s the bullet that was added
after they created the rest of the deck.
(laughing)
And so it’s like, okay, if AI is a feature,
then that’s actually correct,
which is like every, basically everything
is just gonna kind of have AI sprinkled on it.
There’ll be AI features kind of in every product.
That’s possible.
We are more believers in the other scenario,
that AI is a platform and is an architecture.
If, in the same sense that like the mainframe
was architecture, or the mini computer was an architecture,
the PC, the internet, the cloud have been architectures,
we think there’s very good odds that AI
is the next one of those.
And if that’s the case, then it means that basically,
when there’s an architecture shift in our business,
it means basically everything above the architecture
gets rebuilt from scratch.
Because the fundamental assumptions
about what you’re building change, right?
And so you’re no longer building a website,
you’re no longer building a mobile app,
you’re no longer building any of those things
you’re building instead an AI engine
that is just like, in the ideal case,
is just giving you the answer
to whatever the question is.
And if that’s the case,
then basically all applications will change,
along with that all infrastructure will change.
Basically the entire industry will turn over again,
the same way that it did with the internet,
and the same way it did with mobile and cloud.
And so if that’s the case, then it’s just,
it’s going to be like an absolutely explosive period
of growth for this entire industry.
– ‘Cause it means then that all the incumbents,
suppose the incumbents really aren’t incumbent at all.
– Yeah, the products just won’t be relevant anymore.
I mean, I’ll just give you an example.
There are lots and lots of sort of business applications,
just business apps as an example.
There’s lots of business apps,
where you basically, you type data into a form
and then it stores the data
and then later on you run reports
against the data and get charts.
And that’s been the model of business software
for 50 years in different versions.
What if that’s just not needed anymore?
Like what if in the future what you’ll do
is you’ll just give your AI and your business access
to all, email, all phone calls, all everything,
all business records, all financials in the company
and just let the AI give you the answer
to whatever the question was.
And you just don’t go through any of the other steps.
Google’s a good example of this.
Like they’re pushing hard on this.
Like the consumer version of this, right, is search, right?
So search has been, it’s been the 10 blue links
for 25 years now.
What Google’s, they talk about this publicly,
what they’re pushing towards this is just like,
no, it should be that answer.
Which is what they’re trying to do
with their voice UIs.
And so that concept might really generalize out, right?
And then everything gets rebuilt.
– Right.
So one of the new interfaces to AI
that people are talking about is voice
as the new interface.
What are we likely to get wrong about voice?
– Yeah, so I think the thing that,
if we’re gonna get something wrong about voice,
I think it’s gonna be that it would be a one-to-one
replacement for existing user interaction models
so that it would be like a replacement for keyboard
or that it’d be a replacement for the mouse or for touch.
Probably not, ’cause it’s a different modality, right?
It’s, you know, we know exactly what to keep.
After all this time, we know what the keyboard is for,
we know what touch is for and for voice
to displace those, seems like a stretch.
On the other hand, to the previous question,
there has been this turning point reached,
it feels like in AI applied to language
and from there to voice, right, to text and to speech.
Which is, it feels to us in the technology
like the natural language processing methods
that people have been working on for, again,
for 50 years, computer scientists have been working
on getting computers to understand basically speech.
And what we’re seeing now is in the technology
is that that now has started to work
in the same way that machine vision
started to work about seven years ago.
And so if that’s the case, then all of a sudden
the conversational UIs are about to get much better.
And again, and then you couple that with, okay,
what are you actually trying to achieve
when you talk to a computer?
Are you actually trying to like, you know,
are you trying to write a document?
Are you trying to read an email?
Are you trying to like do all these other things
that you do today?
Or are you fundamentally gonna be doing something different?
‘Cause the machine’s gonna be so much smarter.
And I think that’s a very interesting open question.
– When I think about the AR mirror world,
I find it very hard to imagine it without it
having a voice component where we can understand
what you’re saying besides what you’re looking at,
that is that an essential part of the AR world?
– Yeah, I think actually I’d go so far as to say
it may be the case that voice actually is the key
to the AR world, like voice may be the thing.
Voice may actually be the foundation of the whole thing.
You know, for, this is kind of a cliche at this point,
but like the Apple AirPods,
I think were a fundamental breakthrough.
Like it’s again one of these funny things where it’s like,
okay, wireless headphones, okay, cool.
Like wireless headphones where there’s, you know,
there’s not even a wire connecting the two things.
Cool, okay, it seems like more of the same,
but you know, if you want, the experience you can have now
is like you can wear one of these things basically all day
and you can talk to it all day.
And you know, they’re getting, you know,
the new versions are getting better.
You know, and Siri and Google Now and Croton
and all these things are getting really good really fast.
And so it may be that we have just this constant
ongoing running dialogue.
This is kind of, you know, basically the machine
talking to our ear.
And then, you know, the visual overlay of AR
will obviously be important and valuable,
but it might be, it might,
the visual overlay might be supportive
on top of the voice experience.
– And we could very quickly have universal language
translation speaking over the years.
And I think people underestimate the change
that that would bring about in the world.
You’d have millions of people who are highly skilled
in everything except the skill of English.
Now being able to participate in a global economy.
We were talking about unexpected and unexpected things.
Biology, which is a million times as complicated as digital.
We’re now talking about a biotech revolution.
Are we misunderstanding what biotechnology actually is?
– Yeah, so that’s the big bet that we’ve made
with our bio effort that we started a few years back.
We think biological science is at a turning point
at the scientific level and we think it’s at a turning point
from basically being a process of discovery
of how biology works to being able to engineer biology.
And up to including literally being able
to program biology, being able to actually basically
be able to use electrical engineering and computer science
and these mechanical engineering and these kind of fields
for engineering things and be able to apply
those kinds of skills to biology.
If we’re right about that, then the whole concept
of kind of how bio and biotech work might be
on the verge of really changing.
The most obvious application that would be in pharmaceuticals,
there’s this concept of drug discovery.
It’s always the word discovery.
It’s always like, discovery sounds great.
It’s like, it’s optimistic.
It’s like, ooh, discovering things is fantastic.
The problem is, discovering, they literally call it that
’cause they literally have to run all these experiments
and try to discover the drug that works.
Like try to kind of reverse engineer back from nature.
And the problem is sometimes they discover it
and sometimes they don’t.
So the example we always give is we talk about
with computers, we’ve been on this kind of 50 year track
of what’s called Moore’s Law.
We’re at chips beginning faster and cheaper
every year for a long time.
In biology, in drug discovery,
there’s what they call e-rooms law,
which is more spelled backwards, e-room.
And it’s the cost of discovering a new drug.
And it’s exactly the wrong direction.
It’s up and to the right, billions of dollars now.
And so if you could actually engineer biology,
then all of a sudden you can start to apply
this like just these decades of skills
that we’ve built up in how to engineer things
and be able to do things like
engineer new pharmaceuticals from scratch.
– And it all runs on basically ultimately Moore’s Law.
Moore’s Law has been foundational to this year.
It’s almost hard to imagine anything we have
in the modern world today without Moore’s Law.
Do you think Moore’s Law has another 30 years run?
Is it limited?
Is it finite?
Will it go on forever?
We’ll define Moore’s Law in the broadest sense
of computers getting cheaper by half every couple of years.
So what’s your take on Moore’s Law?
– Yeah, so the traditional definition is computer
in the form of the chip, and then specifically a chip.
So Moore’s Law has always been expressed
as kind of unit one of chip.
And that could be right, that could be a CPU
or it could be a graphics card
or it could be a graphics chip or a memory chip.
And then specifically what you were doing was
you were able to put more transistors on that chip
for the same cost.
And actually for a long time as you did that,
you were actually able to reduce the power requirement
for per transistor, which was this kind of added benefit.
And so chips kind of got simultaneously,
they got faster, they got cheaper,
and they got more power efficient.
And that was kind of a cornucopia effect that generated,
as you said, most of what you see today
in the computer industry.
So the bad news is that that in that form
seems to be coming to something of an end,
which is we’re too good at it.
We’ve hit basically, we being the semiconductor industry
broadly, the tech industry have kind of hit
the limits of fundamental physics.
Like we’re now down at the sort of deep atomic level.
And it’s becoming much harder to make,
there’s still progress,
it’s becoming much harder to make progress
at the per chip level.
The good news is that the industry starting
10 or 15 years ago, the computer industry broadly
refocused off of what you do with a chip
to what you do with a large number of chips, right?
So kind of the old model of a chip was
you make the chip more powerful
’cause you’re trying to scale up
what you can do in the chip.
The new model is you use thousands of chips in parallel
and you have this kind of approach to scaling out.
And of course the full form of that
is what’s now known as the cloud.
And so we now have a 15 year head of steam going
to basically be able to get good
at using lots of chips to do things.
And that’s why you see the continued ability to, right?
To accelerate, you know, many, many things
that you deal with are getting still much faster
as if they’re still on the Moore’s Law.
The experiences you’re having are getting faster.
So we think number one,
like the rise of scale at architectures
is a really big deal.
Like, you know, in modern clouds as a developer,
you don’t really care about what the power
of any particular chip is.
You just like light up some more of them
and they don’t cost much.
So there’s that.
The other thing is chips are now specializing.
And in particular, you’ve got the rise
of these new dedicated chips
for things like neural networks
where there’s another level of opportunity to optimize.
And then the other kicker is the programmers.
Software, people like me get to step up.
In the old days, when computers were expensive,
programmers were really good at optimizing
every single step of a software program.
Programmers got out of that habit,
probably starting 30 years ago,
where it didn’t matter as much anymore
because Moore’s Law was working so well.
And so software today is just like massively inefficient.
There’s actually, I forget the name,
there’s something called Worth’s Law,
which is, it was written at the time,
I don’t know if it still holds,
but it was, somebody did benchmarks
of you take Microsoft Office 2000 on a PC from 2000
and you take Microsoft Office 2007 and a PC from 2007
and every function you could do,
you could now do in twice the time, right?
So like literally like,
the old adage in tech in the 90s was
when Andy Grove was running Intel
and Bill Gates was running Microsoft,
it was Andy Gibb in the form of Moore’s Law
and then Bill take it away in the form of software bloat.
And so, and Worth’s Law literally
is a mathematical proof of that.
And so like it’s become prime time again
for software programmers to get really good at optimization,
which is like what’s happening in AI world
and also in the cryptocurrency world.
And so with those different approaches,
it feels like we’ve got,
it feels like decades of advances ahead
that aren’t purely dependent on classic Moore’s Law.
– And because if we take the long-term,
like thinking of a 100 year span to have prosperity
like we’ve seen would kind of require that computer power
sort of get cheaper every year because of a dent
that it’s hard to imagine a world like that.
So is your confidence that we could kind of keep this going
based on just sort of human ingenuity
or do you think that there’s some basic principles
of science that suggest that we’re just at the beginning
of what we can discover?
– Well, so Gordon Moore who invented Moore’s Law
as co-founder of Intel,
he always said Moore’s Law is what was interpreted
as a prophecy and he always said it was not a prophecy.
It was a goal, right?
And it was a goal of basically what you could do
if you focused intensely,
if you focused the entire industry intensely
on a set of engineering optimizations, right?
Over a long period of time.
So he used to say it’s just like,
there’s nothing inevitable about it.
It’s a consequence of thousands and then tens of thousands
and then millions of engineers like working to actually
deliver on these kind of semi-arbitrary goals.
And so I think the answer to that is
we have many, many areas of improvement.
As I said, the problem is we don’t have the one that we had,
which is this transistor doubling kind of effect.
But we’ve got many, I mean,
there’s far more engineers working on all this stuff today
than we’re working on it in 1965.
When he invented Moore’s Law,
we’re in 1995 when everybody bought a PC.
Like we have a lot of mind power going into this.
We’ve got a lot of different technological options.
We’ve got a lot of, you know, incredibly impressive work
happening all over the world.
The other thing is you can’t, you know,
you never like, you know,
one of those things like the transistor was not obvious
and then they invented that.
And then this integrated microchip was like not obvious
and then they invented that.
And so you don’t quite know, you know,
there are lots of technical proposals
for how to get to the next level of Moore’s Law.
You know, so there’s all kinds of theories
around optical computing
and then in the long run, biological computing.
– Quantum computing.
– Quantum computing, exactly.
And so over the course of the next like 20 years,
like, we’ll put it this way.
This is one of the world’s largest prizes, right?
If you’re the engineer who figures out
how to reaccelerate Moore’s Law
or how to shift computing onto a new substrate
like biology, that is the thing to do.
And so that’s the prize.
And that historically has been pretty motivating.
– Right.
So taking this kind of theme of marching forward
progressively, we have 4G, we’re talking about 5G.
So far 5G seems to be faster 4G
with a lot of hype added to it.
There’s a technical specification for 5G,
which is really awesome.
You know, 100 gigabytes, two millisecond latency,
almost impossible.
Are you counting on that for the next decade
that we’re gonna have actual,
what they promise with 5G?
– Yes, I think there’s pretty good odds we will.
And the reason is because 5G has become
a national geopolitical battle.
Like it’s actually a very interesting twist.
It’s become actually a primary, like, you know,
if the Cold War between the US and the USSR
was like defined by the space race,
like at least the sort of nascent Cold War with China
is actually, a lot of it is around 5G, interestingly enough.
I mean, it could have been around a lot of things,
but it happens that it’s around 5G.
And so you now have nation states
that very, very badly need to win,
two big nation states in particular.
And so I think there’s gonna be a lot of, you know,
so we’re gonna start with the payoff from the space races,
like all the products that got, you know,
spun off from that, satellites and GPS
and everything else.
The other thing on 5G, you know,
people sometimes say 5G will lead to applications
they haven’t even thought of yet.
And I think that’s kind of true.
But I look at it a little bit differently.
It’s just a little bit like the most law conversation
we’re having, which is,
I look at a little bit as a math kind of question,
which is there’s sort of three classic rules
for how networks scale
and how network scaling turns into value or usefulness.
And there’s sort of historically,
there’s what’s called Sarnoff’s law,
which was based on broadcast TV,
which is the value of a network is equivalent
to the number of nodes, right?
So it scales with N, right?
So a TV network with 10 million viewers
is twice as valuable as a TV network of 5 million viewers.
That’s kind of the obvious one.
Then there was Metcalf’s law,
which is basically the value of the network
is on the number of connections between two points.
And that’s like how email works, right?
It just emails a person to person.
And that’s correlates to N squared.
So the value of the network rises exponentially
with N squared.
And then there’s this thing called Reed’s law,
which is called the group forming law,
which is the value of the network is proportional
to the number of groups and subgroups
that conform inside the network,
which turns out to be two to the N.
And if you wanna have fun in your plane flights home,
it’s like, just go on Excel and like chart,
N, N squared and two to the Nth, right?
And two to the Nth just goes like straight vertical.
Like you can’t even put them on the same chart.
And two to the Nth is like what’s now happening
was like social networks, right?
So like Facebook groups and all these other things,
like WhatsApp groups and all these other things
people do with social networks and games
and all these other things.
And so those are like the three ways
in which network growth pays off.
And like all three are working now
based on broadband, wired broadband.
They’re all working,
you see it happening very much with mobile.
The introduction of 5G,
the way I think about it is it’s gonna turbo charge
those three networks in particular,
that last one or those last two.
And so it’s gonna add a lot more N.
There’s just gonna be a lot more devices on the network.
There’re gonna be a lot more things
that those devices can do.
They’re gonna be a lot more point to point connections
that make sense to have.
There’s gonna be a lot more groups
that form a lot more economic activity that happens.
– Something that was again,
we were expecting to happen, but didn’t
was in the world of what’s sometimes called
the sharing economy.
There was a, after Airbnb and Uber,
there was a stampede of companies
that were gonna be Uber for X.
And then X was everything in the world.
Very few of them have succeeded.
Again, there was an expectation.
We see more of them, but we haven’t.
So is that whole idea kind of at a dead end?
Is it just, we’re in a very slow disruption.
It’s gonna take a while.
Like the generational requirements
we were talking about technology earlier.
Or is something else?
So what do you think happened there?
And what are we looking at?
– Once again, I object to the question.
– Okay.
– Throw the gavel.
So I look at it a little bit differently,
which is the, this is something we try hard to do
in our place.
It is very tempting.
And we do have this conversation all the time
at our place of like, okay, what about the trend?
What about the theme?
Right, what about the variations on the theme?
Kind of as you said, and this is something happened.
When something wins back, you always get this kind of,
we describe it as kind of the Hollywood model
of, it’s like, what’s your new movie about?
It’s Pretty Woman Meets the Rock, right?
Or whatever.
And so in the Valley, it’s super for X,
or most recently, superhuman for X,
which I’m very excited about,
is one of the big new trends
after another one of our companies.
So, but I don’t think it’s really that.
That’s not how the great ideas arrive.
They don’t look like that.
They look like very specific,
they look at very specific theories,
not general theories.
So they tend to be very specific
to the details of the market involved.
One of the things that I think we’ve learned
about ride sharing, why ride sharing works so well,
I mean, it worked well for many reasons.
One of the reasons it works so well as an idea
is because as long as the driver’s good,
as long as they’re rated at a certain level,
it doesn’t really matter who the driver is.
So like one of the classic examples
was Uber for cleaning your house or your apartment.
And it just, it turns out you just don’t want
a different person over every week
to clean your house like it’s a problem.
And so there’s a lot of these kinds of,
I would say, simple, you know,
sort of the simple applications to the idea
that don’t necessarily work.
Now, by trying all those ideas,
you kind of map the idea space
and you start to get a better sense
of like what your overall structure is.
And I think what’s happening now
is you’re starting to see another set of companies
coming out the other end
that have kind of fully internalized that lesson
and have figured out new models that work.
And so my favorite example,
one of my companies called Honor,
so Honors, you might think loosely,
might think of it as kind of Uber for senior care,
for in-home care for seniors.
It’s a loose model.
Actually, it turns out it’s a very loose model
for a couple of reasons.
One is it’s really deeply not a fungible service.
Like if you have an aging parent,
you actually very much don’t want somebody different
to show up all the time.
You want the same person.
And so in that case, for example,
Honor actually has a full-time employment relationship,
salaried employment relationships with the workers.
Right, which of course is very different
than the Uber and Lyft model.
It actually turns out the matching problem
is much more complicated.
Right, because when you’re matching human beings
in somebody’s home, there’s like 20 variables
that you need to match on
so that everybody’s comfortable with the experience.
As an example, in some cases, you literally need people
with the physical strength to be able to lift people
when you’re caring for them.
You do want to be able to do this kind of
multi-dimensional mapping, and that model’s really working.
And so I think we’re gonna see a whole set of these.
Like I think there’s a big kind of vista of exploration
that’s gonna happen from here.
And I would suspect there will be dozens
of hundreds of new models that people figure out.
– So speaking of new models,
do you ever think about new models
for the VC industry itself,
and how you would apply the principles of innovation
and disruption to what you do in general?
So as you look out 30 years,
what kinds of innovations would you expect
in the basic business that you’re in?
– Yeah, so there’s something very timeless about Venture,
which is there’s actually a new book out called,
literally called VC.
It actually tells the story that has been kind of hard
to get at for a long time in a really clear way,
which is the modern venture model is actually,
one of the historical precedents for it
was actually how whaling expeditions got financed
in the 1600s, so coming up on 500 years ago.
So whaling of, it was literally like, okay,
you’re gonna have like a ship with a captain and a crew
that’s gonna go out and try to like bring back a whale.
Right, and so it’s like a problem number one is like,
only two thirds of the ships are gonna come back, right?
So like high failure rate.
Two is like, okay, what the ship is really matters,
who the captain is really matters,
how do you know who good captain is,
and then what’s a good crew,
and are they gonna be willing to follow the captain?
And then there’s all these like strategy questions,
like do you want the captain who knows
where all the whales have been caught recently,
so they go there, or do you want the captain that says,
no, that’s Gary’s gonna be over fish,
do you wanna go someplace else?
And so literally all the whaling voyages,
like in the colonies 500 years ago,
got financed with basically angel syndicates,
basically venture capital effectively.
And then literally the term carry,
which is sort of how VCs get paid,
the so-called carried interest,
which is like the 25% that you make,
or the profits that you share,
the term carry actually was the percentage of the whale
that the ship carried.
It was literally physical carry.
It was literally that part of the whale,
like that’s where that term came from.
And so there’s a timelessness to the art of trying
to figure out how to finance these kind of expeditions
into the unknown that is likely to endure.
The big question for me is,
how will the shape of the companies,
or let’s say the ventures themselves change, right?
And so today there’s like a well-known understood template
for kind of the prototypical Silicon Valley venture
investment, and it’s like a company in a certain place.
It’s a C corporation, it’s domicile in the US,
it’s financed a certain way,
and to certain types of employees,
a certain relationship with its employees, and so forth.
30 years from now, are we gonna be financing companies here,
or anywhere, or in two places, 50 places, 500 places,
are the companies still gonna have physical place,
or are they gonna be fully virtual?
Are they gonna be companies,
or are they all gonna be blockchains, right?
Are they gonna have actual employment relationships,
or are they gonna have basically developers
and center through cryptocurrency?
That’s a real model.
And so I think the big question is like,
we don’t even know what the shape of companies
is gonna look like, or ventures is gonna look like
in 30 years.
So if I could figure that out,
then I could answer what venture looks like.
Without that, I think it’s hard to say.
– Okay, so we were tempted to do a little bit
of long-term thinking, and long-term thinking
is sort of rare and often ignored,
whereas civilizations demanded as being necessary.
So do you have any suggestions
about how long-term thinking could be applied
in Silicon Valley, and whether you have even
any suggestions to the people in this room
about how they could use long-term thinking?
– Yeah, so the thing I’ve always found about long,
I think long-term thinking is of course central.
Essentially one of the things about the valley
that I find outsiders miss the most,
which is it feels like it’s all moving so fast,
and yet like any of the important companies
and any of the important products
take like a decade or more to build.
And so it’s like everything important
basically takes a long time.
And so a lot of it actually feels quite slow.
And so long-term orientation is absolutely necessary,
and I think we probably all agree
there’s not enough of it in the world.
The thing about long-term thinking I’ve found is like,
it’s really easy if you know the thing is gonna work.
Like, boy, that’s completely straightforward.
Like let’s go on a 10-year journey to a place
where we know it’s gonna be great.
The problem is it’s long-term thinking
crossed with uncertainty, right?
And quite possibly fatality,
like the thing may just simply not work
for any of a thousand reasons.
And so that’s the issue.
And so I think the issue is less around long-term thinking.
I think the issue is more about how to deal with risk
and how to deal with uncertainty
and how to make really big consequential decisions
in the face of literally an unknowable future landscape.
And for there, I mean,
this is kind of the one kind of secret weapon of venture.
It’s like venture is the worst of all asset classes
in a lot of ways in that it’s like it’s a liquid
and it’s like incredibly volatile
and it’s like hit and miss in this kind of crazy way.
The one thing that venture really has going for it
as an asset class is we have the concept
of the portfolio kind of wired into the model
in which you just kind of assume,
in top-end venture, you just kind of assume,
fundamentally it’s half the company’s
gonna work half of them aren’t.
Right, and then the classic, right?
The classic, the cliche is like the ones that work,
then you have to work enough so that they pay
for the ones that don’t to make the whole enterprise work.
And so if you can adapt yourself
from the mentality of will this thing work, right,
to will this portfolio of things basically pay off, right?
Will enough things work
that they’ll actually pay for the portfolio?
Then at that point, you can start to make risk
a somewhat tractable thing to contemplate.
It’s still hard to divorce yourself emotionally from it,
right, ’cause it’s just like it’s still like absolutely,
you know, it’s just terrible
when any of the individual things don’t work,
but at least you have a conception of framework
for you to be able to make 10 long run bets
and being able to get to the other side.
Now, the response that I have to get to that is oh,
that’s great if you’re a VC, the problem is you’re a portfolio,
you know, you’re a founder or a CEO,
like you don’t get that, right?
You have the much harder version of the problem,
which is you’re on a one-way journey,
like you’re the captain of the whaling ship.
Yeah, there’s all those other captains over there,
but like, you know, they’re on their own, you’re on your own.
Even there though, you know,
the best run companies tend to run experiments,
they tend to run multiple experiments against their goals,
and they certainly run those experiments sequentially
as they kind of, you know, try to figure out what works
in a lot of cases, they run experiments in parallel,
as they’re trying to test different things.
And so I also think this kind of mentality
of sort of portfolio risk also applies
to how you run a company,
which is you want to basically,
you want to have a great deal of conviction
about where you’re trying to head,
but you want to have a lot of flexibility inherent
in how you’re going to get there, right,
and what the tactics are,
and then you want to be able to run
a lot of experiments against that,
and you can kind of diversify your risk
of any one theory by doing that.
– And that’s what governments are in some senses,
they have a portfolio of different kind of prospects
about the future, bits, I mean, some senses.
– So you think that’s an optimistic view of what governments do.
– Yeah, so, I mean, that’s what they’re,
and they’re adverse to risk, unfortunately.
– Well, the problem, the problem,
the problem governments have with risk
is like the end of one, right?
So there’s only one government per, right?
We only get to run, you know,
I mean, ex-federalism, which has been a huge advantage,
I think, for the U.S., but like, you know,
the U.S. national government only gets to run one scenario.
– Right.
– And running experiments in the population
is not necessarily well-received.
– Right, ’cause you can’t tolerate failure.
– Yeah, right, yeah, failure has real consequences, so.
– So there’s currently not the only introspection
about government, but also about capitalism,
and capitalism so far has depended on growth,
and growth is something that VCs pay attention to,
but we’re now wondering if what’s the minimum amount
of growth that you might need to have prosperity?
Can you have prosperity with low growth?
Can you have prosperity with fixed growth?
Do you have any insights about that
at the civilizational scale?
– Yeah, so I think, and actually I don’t even say
that the issue is even more intense these days,
’cause there’s now very prominent people
in public life arguing that growth is bad, right?
And in fact, it’s, that in fact is ruinous and destructive,
and that the right goal might actually be
to have no growth or to actually go into negative growth,
then especially in the very common view
in the environmental movement.
So I’m a very strong proponent, a very strong believer
that growth is absolutely necessary,
and I’ll come back to the environmental thing in a second
’cause it’s a very interesting case of this.
I think growth is absolutely necessary,
and I think the reason growth is absolutely necessary
is because you can fundamentally have two different
mindset views of how the world works, right?
One is positive sum, which is rising tide,
lifts all boats, we can all do better together,
and the other is zero sum, right?
Where for me to win, somebody else must lose,
and vice versa.
And the reason I think economic growth is so important
at core is because if there is fast economic growth,
then we have positive sum politics,
and we start to have all these discussions
about all these things that we can do as a society,
and if we have zero sum growth,
if we have a flat growth or no growth or negative growth,
all of a sudden the politics becomes sharply zero sum.
And the most, you just kind of see this
if you kind of track kind of the political climate,
you just, basically it’s the wake of every recession, right?
It’s just that in the wake of every economic recession,
the politics just go like seriously negative
on in terms of thinking about the world’s zero sum.
And then when you get a zero sum outlook in politics,
that’s when you get like anti-immigration,
that’s when you get anti-trade,
that’s when you get anti-tech.
If the world’s not growing, then all that’s left to do
is to fight over what we already have.
And so my view is like, you need to have economic growth.
You need to have economic growth for all of the reasons
that I would say right wingers like economic growth,
which is you wanna have higher levels of material
prosperity, more opportunity, more job creation,
all those things.
You wanna have economic growth for the purpose
of having like sane politics,
like a productive political conversation.
And then I think the kicker is you also want
economic growth actually for many of the things
that left wing people want.
One of the best books this year,
new books this year has got Andrew McAfee,
I was writing a book called I think More From Less.
And it’s actually a story of a really remarkable thing
that a lot of people are missing about what’s happening
with the environment, which is globally carbon emissions
are rising and resource utilization is rising.
In the US, carbon emissions and resource utilization
are actually falling.
And so in the US, we have figured out to grow our economy
while reducing our use of natural resources,
which is a completely unexpected twist, right,
to the plot of what kind of,
if you listen to environmentalists in the 60s and 70s,
like nobody predicted that.
And it turns out, he talks about this in the book,
but it turns out basically what happens is economies,
when economies advance to a certain point,
they get really, really good at doing more with less, right?
They get really, really good at efficiency.
And they get really good at energy efficiency,
they get really used in environmental resources,
they get really good at recycling
in lots of different ways.
And then they get really good at what’s called
dematerialization, which is what is happening
with digital technology, right?
Which is basically taking things that used to require
atoms and turning them into bits,
which inherently consumes less resources.
And so what you actually want,
like my view on environmental issues is like,
you’ve got a global problem,
which is you have too many people in too many countries
stuck in kind of mid the industrial revolution,
they’ve got to grow to get to the point
where they’re in a fully digital economy,
like we are precisely so that they can start
to have declining resource utilization, right?
I mean, the classic example is energy.
Like, you know, the big problem with energy emissions globally,
like a huge problem with emissions and with health
from emissions is literally people burning wood,
like in their houses, right, to be able to heat and cook.
And what you want to do is you want to go to like
hyper-efficient solar or ideally nuclear, right?
You want to go to these like super advanced forms
of technology.
So you want that, and then by the way,
if you want like a big social safety net,
you know, in all the social programs,
you want to pay for that stuff.
You also want economic growth because that generates taxes
that pays for that stuff.
And so like growth is the single kind of biggest
form of magic that we have, right?
To be able to like actually make progress
and hold the whole thing together.
– And to your point about the developing countries,
I think the idea of leapfrogging technology is a myth.
It doesn’t really work.
You actually have to,
if you want to have a high tech infrastructure,
you actually need the intermediate roads, clean water.
You can’t skip over that.
And so they all need to be built out in order to
have that prosperity at the end.
So, you know, it seems like you don’t worry about much.
I don’t worry about much.
But one thing I do worry about is cyber conflict,
cyber war, partly because I think we have no consensus
about what’s allowable.
Does this worry you at all?
– So I think there’s a lot of unknownness to it.
I think people are trying to figure this out,
but it’s a complex issue to grapple with.
I will make an optimistic argument,
which is gonna sound a little strange.
If you kind of project forward what’s happening
with generally a cyber, with information,
you know, operations of different kinds,
but also with drones, you know, UAVs.
And then also with, you know, unmanned fighter jets, right?
Unmanned, you know, ships increasingly being built.
You know, there’ll be unmanned submarines at some point.
If you project this stuff forward,
you start to get this very interesting potential world
in which basically the way I think about it
is like all human conflict between peoples
or between nation states up until now
has been basically throwing people at each other, right?
Throwing soldiers at each other
and like letting them make the decision of who to shoot
and like hoping they don’t get shot,
like with very serious repercussions
of all those individual human decisions.
You do have the prospect of basically a new world
of both offense and defense.
It’s like completely motorized,
completely mechanized, completely software driven
and technology driven.
And a lot of people, it’s just immediately like,
oh my God, that’s horrible.
You know, Terminator, like, you know, Skynet,
like, you know, this is just the worst thing ever.
There’s a novel called “Kill Decision.”
If you want the dystopian theory,
there’s a novel called “Kill Decision.”
– By Daniel Suarez.
– Daniel Suarez that extrapolates the drones forward
and it’ll keep you up late at night.
But the optimistic view would be like, boy,
isn’t it good that there aren’t human beings involved?
Isn’t it good?
Like if the machines are shooting at each other,
like isn’t that good?
Isn’t that better than if they’re shooting at us?
And by the way, and by the way,
I would go so far as to say like,
I don’t know that I’m in favor of like the machines
making like kill decisions, like decisions on who to shoot.
But like the one thing I know is humans do that very badly.
Very, very, very badly.
I’m the opposite of pro war.
I don’t want to see any of this stuff actually play out.
But if it has to play out, maybe having it be software
and machines is going to be actually a better outcome.
– Right.
I mean, it’s this kind of weird that we don’t allow,
we don’t want machines to kill humans.
We want other humans to kill humans.
– We want 18 year olds.
We want to take 18 year olds out of their homes, right?
And we want to put a gun in their hand
and send them someplace and tell them to decide who to shoot.
Like that that is going to go down in history
is having been a good idea.
Just strikes me as like unlikely.
– So we have only time for one last question, which is,
I’m usually, I claim to be the most optimistic person
in the room, but with you sitting across from me,
I don’t think that may be true.
What is your optimism based on?
– So my optimism, okay, so get cosmic for a second.
– Why not?
– I guess we’re here.
– It’s the last question.
– It’s the last question.
– So the science fiction authors always talk about
what’s called the singularity.
This kind of singularity, so the singularity is basically
what happens when the machines get so smart
that all of a sudden everything goes into exponential mode
and all of a sudden the entire world changes.
So my reading history is actually,
we actually were in the singularity already
and that it actually started 300 years ago.
And if you look at basically, if you look at basically
any chart of human welfare over time,
and you can look at, Child Mortality’s an obvious one,
but there’s many, many, many others,
and you just look at progress on that metric.
Just look at Child Mortality as an example
and it’s just basically flat, flat, flat, flat, flat, flat,
flat for like 50,000 years.
And this is the famous, Thomas Hobbes,
life is nasty, brutish and short, right?
It was just like the thing,
like everything was terrible everywhere,
all the time, forever, the end,
until 300 years ago and all of a sudden
there’s this knee in the curve.
And then all the indicators of human welfare,
not uniformly across the planet,
but in societies that were making progress.
Societies that were making progress first,
all of a sudden, all those indicators
of human welfare went up into the right, right?
And then all corresponded, by the way,
to economic growth.
But it was also right, it was the enlightenment,
it was the rise of democracy,
it was the rise of markets,
it was the rise of rationality, the scientific method,
by the way, human rights, free speech, free thought, right?
And they all kind of catalyzed, right,
around 300 years ago and they’ve been making their way
to the world in sort of increasing concentric circles
kind of ever since.
And so we have, I would argue like we have the answers,
like we actually don’t need new discoveries
to have the future be much better,
we actually know how to do it,
is to apply basically those systems.
And basically, contrary to the sort of constant temptation
from all kinds of people to try to compromise
on these things or subvert these things,
basically double down on these systems
that we know work, right?
So double down on economic growth,
double down on human rights,
double down on markets, on capitalism,
double down on the scientific method.
Fixed science, like we got as far as we did with science
actually being pretty seriously screwed up right now
with the replication crisis, like so we should fix that.
And then science will all of a sudden
start to work much better.
Technology, right, use of technological tools.
So we literally have the systems, like we know how to do this,
we know how to make the planet much better in every respect.
And so what we just need to do is keep doing that.
And then what I try to do when I read the news
is notwithstanding everything that’s going on
is basically try to look through whatever’s happened
in the moment, try to look underneath and kind of say,
okay, are those fundamental systems actually still working?
Like is the world getting more democratic or less, right?
Is free speech spreading or receding, right?
Are markets expanding or falling, right?
Are more and more people able to participate
in a modern market economy or not?
And, you know, those indicators generally
are all still up and to the right.
– So let’s go out and make the world better.
Thank you.
– Yeah, good, good.
Thank you everybody.
(audience applauding)
I’m Sonal, happy new year.
Today’s episode is on why we should be optimistic
about the future, because it features two
of the most optimistic people together in conversation.
A6NZ co-founder, Mark Andreessen,
is interviewed by Kevin Kelly,
founding executive editor of Wired Magazine and more.
The conversation originally took place
at our most recent annual innovation conference,
the A6NZ Summit, and it was also previously released
on YouTube if you’d like to check it out there
as well.
– Good afternoon.
Thank you, Mark, for answering some questions.
I have a bunch of questions,
which I hope that we can talk about.
These all have to do about the future, where we’re going.
I want to start with a question about the past.
You know, a generation ago,
a lot of smart people didn’t think
the internet was gonna work,
and therefore they were unprepared for its benefits.
What are we smart people today not prepared for?
– Yeah, so you may remember actually generating,
it wasn’t even just that a lot of people thought
that the internet was gonna work,
a lot of smart people didn’t think that.
In fact, the inventor.
(laughing)
I can’t resist.
I can’t resist on the story.
They actually, the inventor of Ethernet,
which is a foundational technology for the internet,
spent the ’90s actually predicting the internet
would crash, would collapse,
and what we call it would be the gigalapse,
would take down the internet by like 1996, 1997.
He wrote a column at the time for a magazine
called Info World, and he said that if he was wrong,
by, I think it was like,
if the internet hadn’t collapsed by 1997,
he would eat his column.
And to his enormous credit in 1998,
he actually went on stage at a conference,
he actually ripped his column out of the magazine,
he put it in a blender with water,
blended it up, and he drank it on stage.
So it’s one of the more shining examples
of intellectual honesty I’ve ever seen.
As it turns out, he was wrong.
It turns out the internet did work.
So I think the big thing,
I’ve been thinking about this a lot,
you know, it feels to a lot of people
like things are getting strange.
And maybe I’m the only one who feels that way,
but if you read the news,
or just track things happening in the world,
just things feel kind of weird and different
over the last few years.
I actually think there’s like,
there’s that actual generational thing that’s happening,
and you alluded to the generational component.
Like it did take 25 years to get everybody online.
And like we’re not quite there yet,
but we’re getting very close.
Like I think the most exciting thing happened
in the world right now is Mukesh Ambani,
who’s the richest man in India,
has this program called Geo,
where he is literally providing internet access
to the 500 million lowest income Indians,
like literally it’s like free for six months.
And then it’s like a dollar a month.
It’s like the most amazing thing.
And it’s like, it’s working incredibly well.
And so we are very, very close to every,
at least every adult on the planet being internet connected.
But it took 25 years to get there.
And so for me, it’s like, okay, so then what?
One interpretation of that is, okay, we’re done.
We did it.
The other interpretation of that is actually,
okay, that’s just the beginning point.
– Right.
– That’s like the beginning point of what?
– Right.
– And I think it’s the beginning point of like, okay,
like what if you actually interconnect
with everybody on the planet?
Like what, you know, there’s like the metaphor
of the global mind of the global brain.
Like what if you actually connected everybody together
and let everybody find out what everybody else was thinking?
It’s one of those things that people think
sounds good.
And then they encounter it face to face
and they’re like, I don’t know.
– Right, right.
That was like, during my time,
that wired people were kind of concerned
about the digital divide.
And I said, the digital divide is going to cure itself.
The thing you should be worried about
is what happens when everybody is online?
So you think we’re not prepared
for what will happen when everybody is online?
– No, and I think we’re not prepared.
And then I think it’s going to be very exciting.
I mean, I think we’re already seeing that in many ways.
I think the, and then I think we’ve kind of figured out
collectively that it’s going to be different.
And so the initial impulse to say things
are going to get much worse.
And I don’t think that’s right.
I think things are going to get very different.
I think things will be much more positive.
And we’ll talk a lot about that today, hopefully.
But things are definitely going to be different.
– I think one lens that I’ve been trying to put on lately
is kind of think about it through a cultural lens.
Right, sort of what happens to culture
because culture, you know, Ben just wrote this book
about culture being kind of the foundation of the behavior.
And I think that’s really true certainly in companies,
but I think it’s also true in countries and globally.
And it feels like the internet’s impact on culture
is just beginning in the sense of like a world
in which culture is based on the internet,
which is what I think is happening.
It’s just at the very start, right?
‘Cause it had to get universal
before it could set the culture,
but that’s actually happening now.
– Okay.
And at the same time, a generation ago,
well, there was a few people
who actually did think the internet was going to work,
but they were also, like myself, expecting VR
and conversational AI to happen tomorrow.
So what are we expecting to happen now
that it’s not gonna happen?
– Yep, so I object to the question.
(laughing)
Your Honor.
So this is one of those things in our business
that you deal with a lot,
which is ’cause you find yourself,
these entrepreneurs come in and they pitch an idea
and you kind of feel like you should draw judgment
on whether the idea is gonna work or not.
And it’s something I’m really leery of doing anymore.
And the reason for that,
and I think you know this from all of your reading,
every successful technology that I’m aware of,
the things that are like all of a sudden,
like the next big thing, like the iPhone in 2007,
or just as an example,
they all have this like incredible 25 or 40
or 50 year backstory to them.
And you sometimes have to go back and excavate, right?
Because you haven’t heard a lot of the backstory
as the previous efforts failed, right?
But if you go back and look,
like there’s often a multi-generational run-up,
and so I’ll just give you a few of my favorite examples.
So iPhone hit big in 2007.
I for years went around saying,
well, IBM is, there was a 20 year project,
IBM shipped the first smartphone in 1997 called the Simon.
I thought that was true.
It actually turns out it’s not true.
I found the other day, RadioShack had a smartphone in 1982
with their, they literally had a phone version
of their TRS-80 mini computer.
They sold about four of them.
But it was a thing, right?
So that had a 25 year fuse on it.
Video conferencing, video conferencing goes back
at least to the mid ’60s, to the World’s Fair.
The telegraph was invented in the 1870s,
and then sat on a shelf for 100 years
before the Japanese turned it into an industry.
And then another favorite is fiber optics.
Nominally, or you can kind of stretch,
you could say fiber optics were invented in the 1840s.
Paris had a optical telegraph network under the city.
You could actually do, you could actually do under,
you could do telegraphy in the 1840s in Paris.
And it was literally, they were shining flashes of light
through glass tubes.
So there’s this like this incredibly rich back story
to all these things.
And so as a consequence, it’s actually less a question
of like, what’s the new idea?
It turns out the idea is probably already out there
somewhere. – Right, okay.
– And then it’s less the question of like,
is it going to work?
It’s more of the question of like,
when is it gonna work? – Right.
– And I pushed it so far, and people in our office
have heard this.
I pushed all the way to the point where I just think
we should assume that whatever
we’re being pitched is going to work.
(laughing)
It’s just a question of timing.
Then of course, timing turns out to be the hard part,
but it at least focuses the conversation.
– Right, right, right.
So it was the same idea of kind of looking
at the history of things.
One wonders, who really made all the money
when electricity came along?
It probably wasn’t the people
necessarily generating electricity.
Who do you think is gonna make the money
when AI comes along?
Is it the AI providers?
Is it the AI service?
Is it the algorithmic writers?
Who’s gonna be making money in AI?
– Yeah, so we think that there’s two
obvious business models, and probably others,
but the two obvious.
One is to be sort of a horizontal platform provider,
infrastructure provider, you know,
for AI kind of analogous to the operating system
or the database or the cloud.
You know, the other opportunity is kind of in,
would say in the verticals,
and so the applications of AI.
And there’s certainly a lot of those.
So that’s the general answer.
I think that the deeper answer is
there’s an underlying question
that I think is an even bigger question about AI
that reflects directly on this,
which is, is AI a feature or an architecture?
Is AI a feature?
We see this with pitches we get now,
which is just like we get the pitch,
and it’s like, here are the five things my product does,
right, and bullet points one, two, three, four, five,
and then oh yeah, number six is AI.
Right, and so you go, it’s always number six, right,
’cause it’s the bullet that was added
after they created the rest of the deck.
(laughing)
And so it’s like, okay, if AI is a feature,
then that’s actually correct,
which is like every, basically everything
is just gonna kind of have AI sprinkled on it.
There’ll be AI features kind of in every product.
That’s possible.
We are more believers in the other scenario,
that AI is a platform and is an architecture.
If, in the same sense that like the mainframe
was architecture, or the mini computer was an architecture,
the PC, the internet, the cloud have been architectures,
we think there’s very good odds that AI
is the next one of those.
And if that’s the case, then it means that basically,
when there’s an architecture shift in our business,
it means basically everything above the architecture
gets rebuilt from scratch.
Because the fundamental assumptions
about what you’re building change, right?
And so you’re no longer building a website,
you’re no longer building a mobile app,
you’re no longer building any of those things
you’re building instead an AI engine
that is just like, in the ideal case,
is just giving you the answer
to whatever the question is.
And if that’s the case,
then basically all applications will change,
along with that all infrastructure will change.
Basically the entire industry will turn over again,
the same way that it did with the internet,
and the same way it did with mobile and cloud.
And so if that’s the case, then it’s just,
it’s going to be like an absolutely explosive period
of growth for this entire industry.
– ‘Cause it means then that all the incumbents,
suppose the incumbents really aren’t incumbent at all.
– Yeah, the products just won’t be relevant anymore.
I mean, I’ll just give you an example.
There are lots and lots of sort of business applications,
just business apps as an example.
There’s lots of business apps,
where you basically, you type data into a form
and then it stores the data
and then later on you run reports
against the data and get charts.
And that’s been the model of business software
for 50 years in different versions.
What if that’s just not needed anymore?
Like what if in the future what you’ll do
is you’ll just give your AI and your business access
to all, email, all phone calls, all everything,
all business records, all financials in the company
and just let the AI give you the answer
to whatever the question was.
And you just don’t go through any of the other steps.
Google’s a good example of this.
Like they’re pushing hard on this.
Like the consumer version of this, right, is search, right?
So search has been, it’s been the 10 blue links
for 25 years now.
What Google’s, they talk about this publicly,
what they’re pushing towards this is just like,
no, it should be that answer.
Which is what they’re trying to do
with their voice UIs.
And so that concept might really generalize out, right?
And then everything gets rebuilt.
– Right.
So one of the new interfaces to AI
that people are talking about is voice
as the new interface.
What are we likely to get wrong about voice?
– Yeah, so I think the thing that,
if we’re gonna get something wrong about voice,
I think it’s gonna be that it would be a one-to-one
replacement for existing user interaction models
so that it would be like a replacement for keyboard
or that it’d be a replacement for the mouse or for touch.
Probably not, ’cause it’s a different modality, right?
It’s, you know, we know exactly what to keep.
After all this time, we know what the keyboard is for,
we know what touch is for and for voice
to displace those, seems like a stretch.
On the other hand, to the previous question,
there has been this turning point reached,
it feels like in AI applied to language
and from there to voice, right, to text and to speech.
Which is, it feels to us in the technology
like the natural language processing methods
that people have been working on for, again,
for 50 years, computer scientists have been working
on getting computers to understand basically speech.
And what we’re seeing now is in the technology
is that that now has started to work
in the same way that machine vision
started to work about seven years ago.
And so if that’s the case, then all of a sudden
the conversational UIs are about to get much better.
And again, and then you couple that with, okay,
what are you actually trying to achieve
when you talk to a computer?
Are you actually trying to like, you know,
are you trying to write a document?
Are you trying to read an email?
Are you trying to like do all these other things
that you do today?
Or are you fundamentally gonna be doing something different?
‘Cause the machine’s gonna be so much smarter.
And I think that’s a very interesting open question.
– When I think about the AR mirror world,
I find it very hard to imagine it without it
having a voice component where we can understand
what you’re saying besides what you’re looking at,
that is that an essential part of the AR world?
– Yeah, I think actually I’d go so far as to say
it may be the case that voice actually is the key
to the AR world, like voice may be the thing.
Voice may actually be the foundation of the whole thing.
You know, for, this is kind of a cliche at this point,
but like the Apple AirPods,
I think were a fundamental breakthrough.
Like it’s again one of these funny things where it’s like,
okay, wireless headphones, okay, cool.
Like wireless headphones where there’s, you know,
there’s not even a wire connecting the two things.
Cool, okay, it seems like more of the same,
but you know, if you want, the experience you can have now
is like you can wear one of these things basically all day
and you can talk to it all day.
And you know, they’re getting, you know,
the new versions are getting better.
You know, and Siri and Google Now and Croton
and all these things are getting really good really fast.
And so it may be that we have just this constant
ongoing running dialogue.
This is kind of, you know, basically the machine
talking to our ear.
And then, you know, the visual overlay of AR
will obviously be important and valuable,
but it might be, it might,
the visual overlay might be supportive
on top of the voice experience.
– And we could very quickly have universal language
translation speaking over the years.
And I think people underestimate the change
that that would bring about in the world.
You’d have millions of people who are highly skilled
in everything except the skill of English.
Now being able to participate in a global economy.
We were talking about unexpected and unexpected things.
Biology, which is a million times as complicated as digital.
We’re now talking about a biotech revolution.
Are we misunderstanding what biotechnology actually is?
– Yeah, so that’s the big bet that we’ve made
with our bio effort that we started a few years back.
We think biological science is at a turning point
at the scientific level and we think it’s at a turning point
from basically being a process of discovery
of how biology works to being able to engineer biology.
And up to including literally being able
to program biology, being able to actually basically
be able to use electrical engineering and computer science
and these mechanical engineering and these kind of fields
for engineering things and be able to apply
those kinds of skills to biology.
If we’re right about that, then the whole concept
of kind of how bio and biotech work might be
on the verge of really changing.
The most obvious application that would be in pharmaceuticals,
there’s this concept of drug discovery.
It’s always the word discovery.
It’s always like, discovery sounds great.
It’s like, it’s optimistic.
It’s like, ooh, discovering things is fantastic.
The problem is, discovering, they literally call it that
’cause they literally have to run all these experiments
and try to discover the drug that works.
Like try to kind of reverse engineer back from nature.
And the problem is sometimes they discover it
and sometimes they don’t.
So the example we always give is we talk about
with computers, we’ve been on this kind of 50 year track
of what’s called Moore’s Law.
We’re at chips beginning faster and cheaper
every year for a long time.
In biology, in drug discovery,
there’s what they call e-rooms law,
which is more spelled backwards, e-room.
And it’s the cost of discovering a new drug.
And it’s exactly the wrong direction.
It’s up and to the right, billions of dollars now.
And so if you could actually engineer biology,
then all of a sudden you can start to apply
this like just these decades of skills
that we’ve built up in how to engineer things
and be able to do things like
engineer new pharmaceuticals from scratch.
– And it all runs on basically ultimately Moore’s Law.
Moore’s Law has been foundational to this year.
It’s almost hard to imagine anything we have
in the modern world today without Moore’s Law.
Do you think Moore’s Law has another 30 years run?
Is it limited?
Is it finite?
Will it go on forever?
We’ll define Moore’s Law in the broadest sense
of computers getting cheaper by half every couple of years.
So what’s your take on Moore’s Law?
– Yeah, so the traditional definition is computer
in the form of the chip, and then specifically a chip.
So Moore’s Law has always been expressed
as kind of unit one of chip.
And that could be right, that could be a CPU
or it could be a graphics card
or it could be a graphics chip or a memory chip.
And then specifically what you were doing was
you were able to put more transistors on that chip
for the same cost.
And actually for a long time as you did that,
you were actually able to reduce the power requirement
for per transistor, which was this kind of added benefit.
And so chips kind of got simultaneously,
they got faster, they got cheaper,
and they got more power efficient.
And that was kind of a cornucopia effect that generated,
as you said, most of what you see today
in the computer industry.
So the bad news is that that in that form
seems to be coming to something of an end,
which is we’re too good at it.
We’ve hit basically, we being the semiconductor industry
broadly, the tech industry have kind of hit
the limits of fundamental physics.
Like we’re now down at the sort of deep atomic level.
And it’s becoming much harder to make,
there’s still progress,
it’s becoming much harder to make progress
at the per chip level.
The good news is that the industry starting
10 or 15 years ago, the computer industry broadly
refocused off of what you do with a chip
to what you do with a large number of chips, right?
So kind of the old model of a chip was
you make the chip more powerful
’cause you’re trying to scale up
what you can do in the chip.
The new model is you use thousands of chips in parallel
and you have this kind of approach to scaling out.
And of course the full form of that
is what’s now known as the cloud.
And so we now have a 15 year head of steam going
to basically be able to get good
at using lots of chips to do things.
And that’s why you see the continued ability to, right?
To accelerate, you know, many, many things
that you deal with are getting still much faster
as if they’re still on the Moore’s Law.
The experiences you’re having are getting faster.
So we think number one,
like the rise of scale at architectures
is a really big deal.
Like, you know, in modern clouds as a developer,
you don’t really care about what the power
of any particular chip is.
You just like light up some more of them
and they don’t cost much.
So there’s that.
The other thing is chips are now specializing.
And in particular, you’ve got the rise
of these new dedicated chips
for things like neural networks
where there’s another level of opportunity to optimize.
And then the other kicker is the programmers.
Software, people like me get to step up.
In the old days, when computers were expensive,
programmers were really good at optimizing
every single step of a software program.
Programmers got out of that habit,
probably starting 30 years ago,
where it didn’t matter as much anymore
because Moore’s Law was working so well.
And so software today is just like massively inefficient.
There’s actually, I forget the name,
there’s something called Worth’s Law,
which is, it was written at the time,
I don’t know if it still holds,
but it was, somebody did benchmarks
of you take Microsoft Office 2000 on a PC from 2000
and you take Microsoft Office 2007 and a PC from 2007
and every function you could do,
you could now do in twice the time, right?
So like literally like,
the old adage in tech in the 90s was
when Andy Grove was running Intel
and Bill Gates was running Microsoft,
it was Andy Gibb in the form of Moore’s Law
and then Bill take it away in the form of software bloat.
And so, and Worth’s Law literally
is a mathematical proof of that.
And so like it’s become prime time again
for software programmers to get really good at optimization,
which is like what’s happening in AI world
and also in the cryptocurrency world.
And so with those different approaches,
it feels like we’ve got,
it feels like decades of advances ahead
that aren’t purely dependent on classic Moore’s Law.
– And because if we take the long-term,
like thinking of a 100 year span to have prosperity
like we’ve seen would kind of require that computer power
sort of get cheaper every year because of a dent
that it’s hard to imagine a world like that.
So is your confidence that we could kind of keep this going
based on just sort of human ingenuity
or do you think that there’s some basic principles
of science that suggest that we’re just at the beginning
of what we can discover?
– Well, so Gordon Moore who invented Moore’s Law
as co-founder of Intel,
he always said Moore’s Law is what was interpreted
as a prophecy and he always said it was not a prophecy.
It was a goal, right?
And it was a goal of basically what you could do
if you focused intensely,
if you focused the entire industry intensely
on a set of engineering optimizations, right?
Over a long period of time.
So he used to say it’s just like,
there’s nothing inevitable about it.
It’s a consequence of thousands and then tens of thousands
and then millions of engineers like working to actually
deliver on these kind of semi-arbitrary goals.
And so I think the answer to that is
we have many, many areas of improvement.
As I said, the problem is we don’t have the one that we had,
which is this transistor doubling kind of effect.
But we’ve got many, I mean,
there’s far more engineers working on all this stuff today
than we’re working on it in 1965.
When he invented Moore’s Law,
we’re in 1995 when everybody bought a PC.
Like we have a lot of mind power going into this.
We’ve got a lot of different technological options.
We’ve got a lot of, you know, incredibly impressive work
happening all over the world.
The other thing is you can’t, you know,
you never like, you know,
one of those things like the transistor was not obvious
and then they invented that.
And then this integrated microchip was like not obvious
and then they invented that.
And so you don’t quite know, you know,
there are lots of technical proposals
for how to get to the next level of Moore’s Law.
You know, so there’s all kinds of theories
around optical computing
and then in the long run, biological computing.
– Quantum computing.
– Quantum computing, exactly.
And so over the course of the next like 20 years,
like, we’ll put it this way.
This is one of the world’s largest prizes, right?
If you’re the engineer who figures out
how to reaccelerate Moore’s Law
or how to shift computing onto a new substrate
like biology, that is the thing to do.
And so that’s the prize.
And that historically has been pretty motivating.
– Right.
So taking this kind of theme of marching forward
progressively, we have 4G, we’re talking about 5G.
So far 5G seems to be faster 4G
with a lot of hype added to it.
There’s a technical specification for 5G,
which is really awesome.
You know, 100 gigabytes, two millisecond latency,
almost impossible.
Are you counting on that for the next decade
that we’re gonna have actual,
what they promise with 5G?
– Yes, I think there’s pretty good odds we will.
And the reason is because 5G has become
a national geopolitical battle.
Like it’s actually a very interesting twist.
It’s become actually a primary, like, you know,
if the Cold War between the US and the USSR
was like defined by the space race,
like at least the sort of nascent Cold War with China
is actually, a lot of it is around 5G, interestingly enough.
I mean, it could have been around a lot of things,
but it happens that it’s around 5G.
And so you now have nation states
that very, very badly need to win,
two big nation states in particular.
And so I think there’s gonna be a lot of, you know,
so we’re gonna start with the payoff from the space races,
like all the products that got, you know,
spun off from that, satellites and GPS
and everything else.
The other thing on 5G, you know,
people sometimes say 5G will lead to applications
they haven’t even thought of yet.
And I think that’s kind of true.
But I look at it a little bit differently.
It’s just a little bit like the most law conversation
we’re having, which is,
I look at a little bit as a math kind of question,
which is there’s sort of three classic rules
for how networks scale
and how network scaling turns into value or usefulness.
And there’s sort of historically,
there’s what’s called Sarnoff’s law,
which was based on broadcast TV,
which is the value of a network is equivalent
to the number of nodes, right?
So it scales with N, right?
So a TV network with 10 million viewers
is twice as valuable as a TV network of 5 million viewers.
That’s kind of the obvious one.
Then there was Metcalf’s law,
which is basically the value of the network
is on the number of connections between two points.
And that’s like how email works, right?
It just emails a person to person.
And that’s correlates to N squared.
So the value of the network rises exponentially
with N squared.
And then there’s this thing called Reed’s law,
which is called the group forming law,
which is the value of the network is proportional
to the number of groups and subgroups
that conform inside the network,
which turns out to be two to the N.
And if you wanna have fun in your plane flights home,
it’s like, just go on Excel and like chart,
N, N squared and two to the Nth, right?
And two to the Nth just goes like straight vertical.
Like you can’t even put them on the same chart.
And two to the Nth is like what’s now happening
was like social networks, right?
So like Facebook groups and all these other things,
like WhatsApp groups and all these other things
people do with social networks and games
and all these other things.
And so those are like the three ways
in which network growth pays off.
And like all three are working now
based on broadband, wired broadband.
They’re all working,
you see it happening very much with mobile.
The introduction of 5G,
the way I think about it is it’s gonna turbo charge
those three networks in particular,
that last one or those last two.
And so it’s gonna add a lot more N.
There’s just gonna be a lot more devices on the network.
There’re gonna be a lot more things
that those devices can do.
They’re gonna be a lot more point to point connections
that make sense to have.
There’s gonna be a lot more groups
that form a lot more economic activity that happens.
– Something that was again,
we were expecting to happen, but didn’t
was in the world of what’s sometimes called
the sharing economy.
There was a, after Airbnb and Uber,
there was a stampede of companies
that were gonna be Uber for X.
And then X was everything in the world.
Very few of them have succeeded.
Again, there was an expectation.
We see more of them, but we haven’t.
So is that whole idea kind of at a dead end?
Is it just, we’re in a very slow disruption.
It’s gonna take a while.
Like the generational requirements
we were talking about technology earlier.
Or is something else?
So what do you think happened there?
And what are we looking at?
– Once again, I object to the question.
– Okay.
– Throw the gavel.
So I look at it a little bit differently,
which is the, this is something we try hard to do
in our place.
It is very tempting.
And we do have this conversation all the time
at our place of like, okay, what about the trend?
What about the theme?
Right, what about the variations on the theme?
Kind of as you said, and this is something happened.
When something wins back, you always get this kind of,
we describe it as kind of the Hollywood model
of, it’s like, what’s your new movie about?
It’s Pretty Woman Meets the Rock, right?
Or whatever.
And so in the Valley, it’s super for X,
or most recently, superhuman for X,
which I’m very excited about,
is one of the big new trends
after another one of our companies.
So, but I don’t think it’s really that.
That’s not how the great ideas arrive.
They don’t look like that.
They look like very specific,
they look at very specific theories,
not general theories.
So they tend to be very specific
to the details of the market involved.
One of the things that I think we’ve learned
about ride sharing, why ride sharing works so well,
I mean, it worked well for many reasons.
One of the reasons it works so well as an idea
is because as long as the driver’s good,
as long as they’re rated at a certain level,
it doesn’t really matter who the driver is.
So like one of the classic examples
was Uber for cleaning your house or your apartment.
And it just, it turns out you just don’t want
a different person over every week
to clean your house like it’s a problem.
And so there’s a lot of these kinds of,
I would say, simple, you know,
sort of the simple applications to the idea
that don’t necessarily work.
Now, by trying all those ideas,
you kind of map the idea space
and you start to get a better sense
of like what your overall structure is.
And I think what’s happening now
is you’re starting to see another set of companies
coming out the other end
that have kind of fully internalized that lesson
and have figured out new models that work.
And so my favorite example,
one of my companies called Honor,
so Honors, you might think loosely,
might think of it as kind of Uber for senior care,
for in-home care for seniors.
It’s a loose model.
Actually, it turns out it’s a very loose model
for a couple of reasons.
One is it’s really deeply not a fungible service.
Like if you have an aging parent,
you actually very much don’t want somebody different
to show up all the time.
You want the same person.
And so in that case, for example,
Honor actually has a full-time employment relationship,
salaried employment relationships with the workers.
Right, which of course is very different
than the Uber and Lyft model.
It actually turns out the matching problem
is much more complicated.
Right, because when you’re matching human beings
in somebody’s home, there’s like 20 variables
that you need to match on
so that everybody’s comfortable with the experience.
As an example, in some cases, you literally need people
with the physical strength to be able to lift people
when you’re caring for them.
You do want to be able to do this kind of
multi-dimensional mapping, and that model’s really working.
And so I think we’re gonna see a whole set of these.
Like I think there’s a big kind of vista of exploration
that’s gonna happen from here.
And I would suspect there will be dozens
of hundreds of new models that people figure out.
– So speaking of new models,
do you ever think about new models
for the VC industry itself,
and how you would apply the principles of innovation
and disruption to what you do in general?
So as you look out 30 years,
what kinds of innovations would you expect
in the basic business that you’re in?
– Yeah, so there’s something very timeless about Venture,
which is there’s actually a new book out called,
literally called VC.
It actually tells the story that has been kind of hard
to get at for a long time in a really clear way,
which is the modern venture model is actually,
one of the historical precedents for it
was actually how whaling expeditions got financed
in the 1600s, so coming up on 500 years ago.
So whaling of, it was literally like, okay,
you’re gonna have like a ship with a captain and a crew
that’s gonna go out and try to like bring back a whale.
Right, and so it’s like a problem number one is like,
only two thirds of the ships are gonna come back, right?
So like high failure rate.
Two is like, okay, what the ship is really matters,
who the captain is really matters,
how do you know who good captain is,
and then what’s a good crew,
and are they gonna be willing to follow the captain?
And then there’s all these like strategy questions,
like do you want the captain who knows
where all the whales have been caught recently,
so they go there, or do you want the captain that says,
no, that’s Gary’s gonna be over fish,
do you wanna go someplace else?
And so literally all the whaling voyages,
like in the colonies 500 years ago,
got financed with basically angel syndicates,
basically venture capital effectively.
And then literally the term carry,
which is sort of how VCs get paid,
the so-called carried interest,
which is like the 25% that you make,
or the profits that you share,
the term carry actually was the percentage of the whale
that the ship carried.
It was literally physical carry.
It was literally that part of the whale,
like that’s where that term came from.
And so there’s a timelessness to the art of trying
to figure out how to finance these kind of expeditions
into the unknown that is likely to endure.
The big question for me is,
how will the shape of the companies,
or let’s say the ventures themselves change, right?
And so today there’s like a well-known understood template
for kind of the prototypical Silicon Valley venture
investment, and it’s like a company in a certain place.
It’s a C corporation, it’s domicile in the US,
it’s financed a certain way,
and to certain types of employees,
a certain relationship with its employees, and so forth.
30 years from now, are we gonna be financing companies here,
or anywhere, or in two places, 50 places, 500 places,
are the companies still gonna have physical place,
or are they gonna be fully virtual?
Are they gonna be companies,
or are they all gonna be blockchains, right?
Are they gonna have actual employment relationships,
or are they gonna have basically developers
and center through cryptocurrency?
That’s a real model.
And so I think the big question is like,
we don’t even know what the shape of companies
is gonna look like, or ventures is gonna look like
in 30 years.
So if I could figure that out,
then I could answer what venture looks like.
Without that, I think it’s hard to say.
– Okay, so we were tempted to do a little bit
of long-term thinking, and long-term thinking
is sort of rare and often ignored,
whereas civilizations demanded as being necessary.
So do you have any suggestions
about how long-term thinking could be applied
in Silicon Valley, and whether you have even
any suggestions to the people in this room
about how they could use long-term thinking?
– Yeah, so the thing I’ve always found about long,
I think long-term thinking is of course central.
Essentially one of the things about the valley
that I find outsiders miss the most,
which is it feels like it’s all moving so fast,
and yet like any of the important companies
and any of the important products
take like a decade or more to build.
And so it’s like everything important
basically takes a long time.
And so a lot of it actually feels quite slow.
And so long-term orientation is absolutely necessary,
and I think we probably all agree
there’s not enough of it in the world.
The thing about long-term thinking I’ve found is like,
it’s really easy if you know the thing is gonna work.
Like, boy, that’s completely straightforward.
Like let’s go on a 10-year journey to a place
where we know it’s gonna be great.
The problem is it’s long-term thinking
crossed with uncertainty, right?
And quite possibly fatality,
like the thing may just simply not work
for any of a thousand reasons.
And so that’s the issue.
And so I think the issue is less around long-term thinking.
I think the issue is more about how to deal with risk
and how to deal with uncertainty
and how to make really big consequential decisions
in the face of literally an unknowable future landscape.
And for there, I mean,
this is kind of the one kind of secret weapon of venture.
It’s like venture is the worst of all asset classes
in a lot of ways in that it’s like it’s a liquid
and it’s like incredibly volatile
and it’s like hit and miss in this kind of crazy way.
The one thing that venture really has going for it
as an asset class is we have the concept
of the portfolio kind of wired into the model
in which you just kind of assume,
in top-end venture, you just kind of assume,
fundamentally it’s half the company’s
gonna work half of them aren’t.
Right, and then the classic, right?
The classic, the cliche is like the ones that work,
then you have to work enough so that they pay
for the ones that don’t to make the whole enterprise work.
And so if you can adapt yourself
from the mentality of will this thing work, right,
to will this portfolio of things basically pay off, right?
Will enough things work
that they’ll actually pay for the portfolio?
Then at that point, you can start to make risk
a somewhat tractable thing to contemplate.
It’s still hard to divorce yourself emotionally from it,
right, ’cause it’s just like it’s still like absolutely,
you know, it’s just terrible
when any of the individual things don’t work,
but at least you have a conception of framework
for you to be able to make 10 long run bets
and being able to get to the other side.
Now, the response that I have to get to that is oh,
that’s great if you’re a VC, the problem is you’re a portfolio,
you know, you’re a founder or a CEO,
like you don’t get that, right?
You have the much harder version of the problem,
which is you’re on a one-way journey,
like you’re the captain of the whaling ship.
Yeah, there’s all those other captains over there,
but like, you know, they’re on their own, you’re on your own.
Even there though, you know,
the best run companies tend to run experiments,
they tend to run multiple experiments against their goals,
and they certainly run those experiments sequentially
as they kind of, you know, try to figure out what works
in a lot of cases, they run experiments in parallel,
as they’re trying to test different things.
And so I also think this kind of mentality
of sort of portfolio risk also applies
to how you run a company,
which is you want to basically,
you want to have a great deal of conviction
about where you’re trying to head,
but you want to have a lot of flexibility inherent
in how you’re going to get there, right,
and what the tactics are,
and then you want to be able to run
a lot of experiments against that,
and you can kind of diversify your risk
of any one theory by doing that.
– And that’s what governments are in some senses,
they have a portfolio of different kind of prospects
about the future, bits, I mean, some senses.
– So you think that’s an optimistic view of what governments do.
– Yeah, so, I mean, that’s what they’re,
and they’re adverse to risk, unfortunately.
– Well, the problem, the problem,
the problem governments have with risk
is like the end of one, right?
So there’s only one government per, right?
We only get to run, you know,
I mean, ex-federalism, which has been a huge advantage,
I think, for the U.S., but like, you know,
the U.S. national government only gets to run one scenario.
– Right.
– And running experiments in the population
is not necessarily well-received.
– Right, ’cause you can’t tolerate failure.
– Yeah, right, yeah, failure has real consequences, so.
– So there’s currently not the only introspection
about government, but also about capitalism,
and capitalism so far has depended on growth,
and growth is something that VCs pay attention to,
but we’re now wondering if what’s the minimum amount
of growth that you might need to have prosperity?
Can you have prosperity with low growth?
Can you have prosperity with fixed growth?
Do you have any insights about that
at the civilizational scale?
– Yeah, so I think, and actually I don’t even say
that the issue is even more intense these days,
’cause there’s now very prominent people
in public life arguing that growth is bad, right?
And in fact, it’s, that in fact is ruinous and destructive,
and that the right goal might actually be
to have no growth or to actually go into negative growth,
then especially in the very common view
in the environmental movement.
So I’m a very strong proponent, a very strong believer
that growth is absolutely necessary,
and I’ll come back to the environmental thing in a second
’cause it’s a very interesting case of this.
I think growth is absolutely necessary,
and I think the reason growth is absolutely necessary
is because you can fundamentally have two different
mindset views of how the world works, right?
One is positive sum, which is rising tide,
lifts all boats, we can all do better together,
and the other is zero sum, right?
Where for me to win, somebody else must lose,
and vice versa.
And the reason I think economic growth is so important
at core is because if there is fast economic growth,
then we have positive sum politics,
and we start to have all these discussions
about all these things that we can do as a society,
and if we have zero sum growth,
if we have a flat growth or no growth or negative growth,
all of a sudden the politics becomes sharply zero sum.
And the most, you just kind of see this
if you kind of track kind of the political climate,
you just, basically it’s the wake of every recession, right?
It’s just that in the wake of every economic recession,
the politics just go like seriously negative
on in terms of thinking about the world’s zero sum.
And then when you get a zero sum outlook in politics,
that’s when you get like anti-immigration,
that’s when you get anti-trade,
that’s when you get anti-tech.
If the world’s not growing, then all that’s left to do
is to fight over what we already have.
And so my view is like, you need to have economic growth.
You need to have economic growth for all of the reasons
that I would say right wingers like economic growth,
which is you wanna have higher levels of material
prosperity, more opportunity, more job creation,
all those things.
You wanna have economic growth for the purpose
of having like sane politics,
like a productive political conversation.
And then I think the kicker is you also want
economic growth actually for many of the things
that left wing people want.
One of the best books this year,
new books this year has got Andrew McAfee,
I was writing a book called I think More From Less.
And it’s actually a story of a really remarkable thing
that a lot of people are missing about what’s happening
with the environment, which is globally carbon emissions
are rising and resource utilization is rising.
In the US, carbon emissions and resource utilization
are actually falling.
And so in the US, we have figured out to grow our economy
while reducing our use of natural resources,
which is a completely unexpected twist, right,
to the plot of what kind of,
if you listen to environmentalists in the 60s and 70s,
like nobody predicted that.
And it turns out, he talks about this in the book,
but it turns out basically what happens is economies,
when economies advance to a certain point,
they get really, really good at doing more with less, right?
They get really, really good at efficiency.
And they get really good at energy efficiency,
they get really used in environmental resources,
they get really good at recycling
in lots of different ways.
And then they get really good at what’s called
dematerialization, which is what is happening
with digital technology, right?
Which is basically taking things that used to require
atoms and turning them into bits,
which inherently consumes less resources.
And so what you actually want,
like my view on environmental issues is like,
you’ve got a global problem,
which is you have too many people in too many countries
stuck in kind of mid the industrial revolution,
they’ve got to grow to get to the point
where they’re in a fully digital economy,
like we are precisely so that they can start
to have declining resource utilization, right?
I mean, the classic example is energy.
Like, you know, the big problem with energy emissions globally,
like a huge problem with emissions and with health
from emissions is literally people burning wood,
like in their houses, right, to be able to heat and cook.
And what you want to do is you want to go to like
hyper-efficient solar or ideally nuclear, right?
You want to go to these like super advanced forms
of technology.
So you want that, and then by the way,
if you want like a big social safety net,
you know, in all the social programs,
you want to pay for that stuff.
You also want economic growth because that generates taxes
that pays for that stuff.
And so like growth is the single kind of biggest
form of magic that we have, right?
To be able to like actually make progress
and hold the whole thing together.
– And to your point about the developing countries,
I think the idea of leapfrogging technology is a myth.
It doesn’t really work.
You actually have to,
if you want to have a high tech infrastructure,
you actually need the intermediate roads, clean water.
You can’t skip over that.
And so they all need to be built out in order to
have that prosperity at the end.
So, you know, it seems like you don’t worry about much.
I don’t worry about much.
But one thing I do worry about is cyber conflict,
cyber war, partly because I think we have no consensus
about what’s allowable.
Does this worry you at all?
– So I think there’s a lot of unknownness to it.
I think people are trying to figure this out,
but it’s a complex issue to grapple with.
I will make an optimistic argument,
which is gonna sound a little strange.
If you kind of project forward what’s happening
with generally a cyber, with information,
you know, operations of different kinds,
but also with drones, you know, UAVs.
And then also with, you know, unmanned fighter jets, right?
Unmanned, you know, ships increasingly being built.
You know, there’ll be unmanned submarines at some point.
If you project this stuff forward,
you start to get this very interesting potential world
in which basically the way I think about it
is like all human conflict between peoples
or between nation states up until now
has been basically throwing people at each other, right?
Throwing soldiers at each other
and like letting them make the decision of who to shoot
and like hoping they don’t get shot,
like with very serious repercussions
of all those individual human decisions.
You do have the prospect of basically a new world
of both offense and defense.
It’s like completely motorized,
completely mechanized, completely software driven
and technology driven.
And a lot of people, it’s just immediately like,
oh my God, that’s horrible.
You know, Terminator, like, you know, Skynet,
like, you know, this is just the worst thing ever.
There’s a novel called “Kill Decision.”
If you want the dystopian theory,
there’s a novel called “Kill Decision.”
– By Daniel Suarez.
– Daniel Suarez that extrapolates the drones forward
and it’ll keep you up late at night.
But the optimistic view would be like, boy,
isn’t it good that there aren’t human beings involved?
Isn’t it good?
Like if the machines are shooting at each other,
like isn’t that good?
Isn’t that better than if they’re shooting at us?
And by the way, and by the way,
I would go so far as to say like,
I don’t know that I’m in favor of like the machines
making like kill decisions, like decisions on who to shoot.
But like the one thing I know is humans do that very badly.
Very, very, very badly.
I’m the opposite of pro war.
I don’t want to see any of this stuff actually play out.
But if it has to play out, maybe having it be software
and machines is going to be actually a better outcome.
– Right.
I mean, it’s this kind of weird that we don’t allow,
we don’t want machines to kill humans.
We want other humans to kill humans.
– We want 18 year olds.
We want to take 18 year olds out of their homes, right?
And we want to put a gun in their hand
and send them someplace and tell them to decide who to shoot.
Like that that is going to go down in history
is having been a good idea.
Just strikes me as like unlikely.
– So we have only time for one last question, which is,
I’m usually, I claim to be the most optimistic person
in the room, but with you sitting across from me,
I don’t think that may be true.
What is your optimism based on?
– So my optimism, okay, so get cosmic for a second.
– Why not?
– I guess we’re here.
– It’s the last question.
– It’s the last question.
– So the science fiction authors always talk about
what’s called the singularity.
This kind of singularity, so the singularity is basically
what happens when the machines get so smart
that all of a sudden everything goes into exponential mode
and all of a sudden the entire world changes.
So my reading history is actually,
we actually were in the singularity already
and that it actually started 300 years ago.
And if you look at basically, if you look at basically
any chart of human welfare over time,
and you can look at, Child Mortality’s an obvious one,
but there’s many, many, many others,
and you just look at progress on that metric.
Just look at Child Mortality as an example
and it’s just basically flat, flat, flat, flat, flat, flat,
flat for like 50,000 years.
And this is the famous, Thomas Hobbes,
life is nasty, brutish and short, right?
It was just like the thing,
like everything was terrible everywhere,
all the time, forever, the end,
until 300 years ago and all of a sudden
there’s this knee in the curve.
And then all the indicators of human welfare,
not uniformly across the planet,
but in societies that were making progress.
Societies that were making progress first,
all of a sudden, all those indicators
of human welfare went up into the right, right?
And then all corresponded, by the way,
to economic growth.
But it was also right, it was the enlightenment,
it was the rise of democracy,
it was the rise of markets,
it was the rise of rationality, the scientific method,
by the way, human rights, free speech, free thought, right?
And they all kind of catalyzed, right,
around 300 years ago and they’ve been making their way
to the world in sort of increasing concentric circles
kind of ever since.
And so we have, I would argue like we have the answers,
like we actually don’t need new discoveries
to have the future be much better,
we actually know how to do it,
is to apply basically those systems.
And basically, contrary to the sort of constant temptation
from all kinds of people to try to compromise
on these things or subvert these things,
basically double down on these systems
that we know work, right?
So double down on economic growth,
double down on human rights,
double down on markets, on capitalism,
double down on the scientific method.
Fixed science, like we got as far as we did with science
actually being pretty seriously screwed up right now
with the replication crisis, like so we should fix that.
And then science will all of a sudden
start to work much better.
Technology, right, use of technological tools.
So we literally have the systems, like we know how to do this,
we know how to make the planet much better in every respect.
And so what we just need to do is keep doing that.
And then what I try to do when I read the news
is notwithstanding everything that’s going on
is basically try to look through whatever’s happened
in the moment, try to look underneath and kind of say,
okay, are those fundamental systems actually still working?
Like is the world getting more democratic or less, right?
Is free speech spreading or receding, right?
Are markets expanding or falling, right?
Are more and more people able to participate
in a modern market economy or not?
And, you know, those indicators generally
are all still up and to the right.
– So let’s go out and make the world better.
Thank you.
– Yeah, good, good.
Thank you everybody.
(audience applauding)
Many skeptics thought the internet would never reach mass adoption, but today it’s shaping global culture, is integral to our lives — and it’s just the beginning.
In this conversation from our 2019 innovation summit, Kevin Kelly (Founding Executive Editor, WIRED magazine) and Marc Andreessen sit down to discuss the evolution of technology, key trends, and why they’re the most optimistic people in the room.
***
The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.
This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.
Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.
Leave a Reply