Why We Should Be Optimistic About the Future

0:00:03 – Hi everyone, welcome to the A6NZ podcast.
0:00:05 I’m Sonal, happy new year.
0:00:08 Today’s episode is on why we should be optimistic
0:00:10 about the future, because it features two
0:00:13 of the most optimistic people together in conversation.
0:00:15 A6NZ co-founder, Mark Andreessen,
0:00:17 is interviewed by Kevin Kelly,
0:00:20 founding executive editor of Wired Magazine and more.
0:00:22 The conversation originally took place
0:00:25 at our most recent annual innovation conference,
0:00:28 the A6NZ Summit, and it was also previously released
0:00:29 on YouTube if you’d like to check it out there
0:00:30 as well.
0:00:33 – Good afternoon.
0:00:37 Thank you, Mark, for answering some questions.
0:00:38 I have a bunch of questions,
0:00:41 which I hope that we can talk about.
0:00:44 These all have to do about the future, where we’re going.
0:00:46 I want to start with a question about the past.
0:00:47 You know, a generation ago,
0:00:50 a lot of smart people didn’t think
0:00:52 the internet was gonna work,
0:00:55 and therefore they were unprepared for its benefits.
0:01:00 What are we smart people today not prepared for?
0:01:02 – Yeah, so you may remember actually generating,
0:01:04 it wasn’t even just that a lot of people thought
0:01:05 that the internet was gonna work,
0:01:06 a lot of smart people didn’t think that.
0:01:07 In fact, the inventor.
0:01:09 (laughing)
0:01:10 I can’t resist.
0:01:11 I can’t resist on the story.
0:01:12 They actually, the inventor of Ethernet,
0:01:15 which is a foundational technology for the internet,
0:01:17 spent the ’90s actually predicting the internet
0:01:18 would crash, would collapse,
0:01:20 and what we call it would be the gigalapse,
0:01:23 would take down the internet by like 1996, 1997.
0:01:25 He wrote a column at the time for a magazine
0:01:27 called Info World, and he said that if he was wrong,
0:01:28 by, I think it was like,
0:01:30 if the internet hadn’t collapsed by 1997,
0:01:32 he would eat his column.
0:01:36 And to his enormous credit in 1998,
0:01:37 he actually went on stage at a conference,
0:01:40 he actually ripped his column out of the magazine,
0:01:41 he put it in a blender with water,
0:01:43 blended it up, and he drank it on stage.
0:01:46 So it’s one of the more shining examples
0:01:49 of intellectual honesty I’ve ever seen.
0:01:51 As it turns out, he was wrong.
0:01:52 It turns out the internet did work.
0:01:53 So I think the big thing,
0:01:55 I’ve been thinking about this a lot,
0:01:57 you know, it feels to a lot of people
0:01:59 like things are getting strange.
0:02:01 And maybe I’m the only one who feels that way,
0:02:03 but if you read the news,
0:02:04 or just track things happening in the world,
0:02:06 just things feel kind of weird and different
0:02:07 over the last few years.
0:02:08 I actually think there’s like,
0:02:10 there’s that actual generational thing that’s happening,
0:02:12 and you alluded to the generational component.
0:02:14 Like it did take 25 years to get everybody online.
0:02:16 And like we’re not quite there yet,
0:02:18 but we’re getting very close.
0:02:19 Like I think the most exciting thing happened
0:02:21 in the world right now is Mukesh Ambani,
0:02:22 who’s the richest man in India,
0:02:24 has this program called Geo,
0:02:26 where he is literally providing internet access
0:02:28 to the 500 million lowest income Indians,
0:02:33 like literally it’s like free for six months.
0:02:34 And then it’s like a dollar a month.
0:02:35 It’s like the most amazing thing.
0:02:36 And it’s like, it’s working incredibly well.
0:02:39 And so we are very, very close to every,
0:02:42 at least every adult on the planet being internet connected.
0:02:43 But it took 25 years to get there.
0:02:46 And so for me, it’s like, okay, so then what?
0:02:48 One interpretation of that is, okay, we’re done.
0:02:49 We did it.
0:02:50 The other interpretation of that is actually,
0:02:52 okay, that’s just the beginning point.
0:02:53 – Right.
0:02:54 – That’s like the beginning point of what?
0:02:55 – Right.
0:02:56 – And I think it’s the beginning point of like, okay,
0:02:58 like what if you actually interconnect
0:02:59 with everybody on the planet?
0:03:00 Like what, you know, there’s like the metaphor
0:03:02 of the global mind of the global brain.
0:03:04 Like what if you actually connected everybody together
0:03:07 and let everybody find out what everybody else was thinking?
0:03:08 It’s one of those things that people think
0:03:09 sounds good.
0:03:10 And then they encounter it face to face
0:03:11 and they’re like, I don’t know.
0:03:12 – Right, right.
0:03:14 That was like, during my time,
0:03:16 that wired people were kind of concerned
0:03:17 about the digital divide.
0:03:19 And I said, the digital divide is going to cure itself.
0:03:21 The thing you should be worried about
0:03:24 is what happens when everybody is online?
0:03:26 So you think we’re not prepared
0:03:28 for what will happen when everybody is online?
0:03:30 – No, and I think we’re not prepared.
0:03:31 And then I think it’s going to be very exciting.
0:03:33 I mean, I think we’re already seeing that in many ways.
0:03:37 I think the, and then I think we’ve kind of figured out
0:03:38 collectively that it’s going to be different.
0:03:40 And so the initial impulse to say things
0:03:41 are going to get much worse.
0:03:42 And I don’t think that’s right.
0:03:43 I think things are going to get very different.
0:03:46 I think things will be much more positive.
0:03:48 And we’ll talk a lot about that today, hopefully.
0:03:50 But things are definitely going to be different.
0:03:53 – I think one lens that I’ve been trying to put on lately
0:03:55 is kind of think about it through a cultural lens.
0:03:57 Right, sort of what happens to culture
0:04:00 because culture, you know, Ben just wrote this book
0:04:02 about culture being kind of the foundation of the behavior.
0:04:04 And I think that’s really true certainly in companies,
0:04:06 but I think it’s also true in countries and globally.
0:04:08 And it feels like the internet’s impact on culture
0:04:12 is just beginning in the sense of like a world
0:04:13 in which culture is based on the internet,
0:04:15 which is what I think is happening.
0:04:16 It’s just at the very start, right?
0:04:17 ‘Cause it had to get universal
0:04:19 before it could set the culture,
0:04:20 but that’s actually happening now.
0:04:21 – Okay.
0:04:23 And at the same time, a generation ago,
0:04:25 well, there was a few people
0:04:28 who actually did think the internet was going to work,
0:04:32 but they were also, like myself, expecting VR
0:04:36 and conversational AI to happen tomorrow.
0:04:39 So what are we expecting to happen now
0:04:40 that it’s not gonna happen?
0:04:42 – Yep, so I object to the question.
0:04:44 (laughing)
0:04:45 Your Honor.
0:04:48 So this is one of those things in our business
0:04:49 that you deal with a lot,
0:04:50 which is ’cause you find yourself,
0:04:52 these entrepreneurs come in and they pitch an idea
0:04:53 and you kind of feel like you should draw judgment
0:04:54 on whether the idea is gonna work or not.
0:04:57 And it’s something I’m really leery of doing anymore.
0:04:59 And the reason for that,
0:05:01 and I think you know this from all of your reading,
0:05:04 every successful technology that I’m aware of,
0:05:05 the things that are like all of a sudden,
0:05:08 like the next big thing, like the iPhone in 2007,
0:05:09 or just as an example,
0:05:12 they all have this like incredible 25 or 40
0:05:14 or 50 year backstory to them.
0:05:16 And you sometimes have to go back and excavate, right?
0:05:17 Because you haven’t heard a lot of the backstory
0:05:20 as the previous efforts failed, right?
0:05:21 But if you go back and look,
0:05:24 like there’s often a multi-generational run-up,
0:05:26 and so I’ll just give you a few of my favorite examples.
0:05:29 So iPhone hit big in 2007.
0:05:30 I for years went around saying,
0:05:32 well, IBM is, there was a 20 year project,
0:05:35 IBM shipped the first smartphone in 1997 called the Simon.
0:05:36 I thought that was true.
0:05:37 It actually turns out it’s not true.
0:05:41 I found the other day, RadioShack had a smartphone in 1982
0:05:43 with their, they literally had a phone version
0:05:45 of their TRS-80 mini computer.
0:05:47 They sold about four of them.
0:05:49 But it was a thing, right?
0:05:52 So that had a 25 year fuse on it.
0:05:54 Video conferencing, video conferencing goes back
0:05:57 at least to the mid ’60s, to the World’s Fair.
0:06:01 The telegraph was invented in the 1870s,
0:06:03 and then sat on a shelf for 100 years
0:06:05 before the Japanese turned it into an industry.
0:06:07 And then another favorite is fiber optics.
0:06:10 Nominally, or you can kind of stretch,
0:06:13 you could say fiber optics were invented in the 1840s.
0:06:18 Paris had a optical telegraph network under the city.
0:06:19 You could actually do, you could actually do under,
0:06:21 you could do telegraphy in the 1840s in Paris.
0:06:23 And it was literally, they were shining flashes of light
0:06:24 through glass tubes.
0:06:27 So there’s this like this incredibly rich back story
0:06:29 to all these things.
0:06:31 And so as a consequence, it’s actually less a question
0:06:32 of like, what’s the new idea?
0:06:33 It turns out the idea is probably already out there
0:06:34 somewhere. – Right, okay.
0:06:35 – And then it’s less the question of like,
0:06:36 is it going to work?
0:06:37 It’s more of the question of like,
0:06:38 when is it gonna work? – Right.
0:06:40 – And I pushed it so far, and people in our office
0:06:41 have heard this.
0:06:42 I pushed all the way to the point where I just think
0:06:43 we should assume that whatever
0:06:45 we’re being pitched is going to work.
0:06:46 (laughing)
0:06:48 It’s just a question of timing.
0:06:50 Then of course, timing turns out to be the hard part,
0:06:52 but it at least focuses the conversation.
0:06:53 – Right, right, right.
0:06:55 So it was the same idea of kind of looking
0:06:57 at the history of things.
0:07:00 One wonders, who really made all the money
0:07:02 when electricity came along?
0:07:03 It probably wasn’t the people
0:07:06 necessarily generating electricity.
0:07:08 Who do you think is gonna make the money
0:07:10 when AI comes along?
0:07:12 Is it the AI providers?
0:07:15 Is it the AI service?
0:07:18 Is it the algorithmic writers?
0:07:20 Who’s gonna be making money in AI?
0:07:22 – Yeah, so we think that there’s two
0:07:24 obvious business models, and probably others,
0:07:25 but the two obvious.
0:07:26 One is to be sort of a horizontal platform provider,
0:07:28 infrastructure provider, you know,
0:07:29 for AI kind of analogous to the operating system
0:07:32 or the database or the cloud.
0:07:33 You know, the other opportunity is kind of in,
0:07:34 would say in the verticals,
0:07:35 and so the applications of AI.
0:07:38 And there’s certainly a lot of those.
0:07:39 So that’s the general answer.
0:07:41 I think that the deeper answer is
0:07:41 there’s an underlying question
0:07:43 that I think is an even bigger question about AI
0:07:45 that reflects directly on this,
0:07:49 which is, is AI a feature or an architecture?
0:07:52 Is AI a feature?
0:07:54 We see this with pitches we get now,
0:07:55 which is just like we get the pitch,
0:07:58 and it’s like, here are the five things my product does,
0:08:00 right, and bullet points one, two, three, four, five,
0:08:02 and then oh yeah, number six is AI.
0:08:04 Right, and so you go, it’s always number six, right,
0:08:05 ’cause it’s the bullet that was added
0:08:06 after they created the rest of the deck.
0:08:08 (laughing)
0:08:10 And so it’s like, okay, if AI is a feature,
0:08:11 then that’s actually correct,
0:08:13 which is like every, basically everything
0:08:15 is just gonna kind of have AI sprinkled on it.
0:08:17 There’ll be AI features kind of in every product.
0:08:18 That’s possible.
0:08:21 We are more believers in the other scenario,
0:08:24 that AI is a platform and is an architecture.
0:08:26 If, in the same sense that like the mainframe
0:08:28 was architecture, or the mini computer was an architecture,
0:08:31 the PC, the internet, the cloud have been architectures,
0:08:33 we think there’s very good odds that AI
0:08:34 is the next one of those.
0:08:37 And if that’s the case, then it means that basically,
0:08:38 when there’s an architecture shift in our business,
0:08:40 it means basically everything above the architecture
0:08:41 gets rebuilt from scratch.
0:08:43 Because the fundamental assumptions
0:08:45 about what you’re building change, right?
0:08:47 And so you’re no longer building a website,
0:08:48 you’re no longer building a mobile app,
0:08:49 you’re no longer building any of those things
0:08:51 you’re building instead an AI engine
0:08:52 that is just like, in the ideal case,
0:08:53 is just giving you the answer
0:08:55 to whatever the question is.
0:08:56 And if that’s the case,
0:08:59 then basically all applications will change,
0:09:01 along with that all infrastructure will change.
0:09:03 Basically the entire industry will turn over again,
0:09:04 the same way that it did with the internet,
0:09:06 and the same way it did with mobile and cloud.
0:09:08 And so if that’s the case, then it’s just,
0:09:10 it’s going to be like an absolutely explosive period
0:09:12 of growth for this entire industry.
0:09:14 – ‘Cause it means then that all the incumbents,
0:09:17 suppose the incumbents really aren’t incumbent at all.
0:09:19 – Yeah, the products just won’t be relevant anymore.
0:09:20 I mean, I’ll just give you an example.
0:09:22 There are lots and lots of sort of business applications,
0:09:23 just business apps as an example.
0:09:25 There’s lots of business apps,
0:09:27 where you basically, you type data into a form
0:09:28 and then it stores the data
0:09:29 and then later on you run reports
0:09:30 against the data and get charts.
0:09:32 And that’s been the model of business software
0:09:34 for 50 years in different versions.
0:09:36 What if that’s just not needed anymore?
0:09:38 Like what if in the future what you’ll do
0:09:40 is you’ll just give your AI and your business access
0:09:42 to all, email, all phone calls, all everything,
0:09:44 all business records, all financials in the company
0:09:46 and just let the AI give you the answer
0:09:47 to whatever the question was.
0:09:49 And you just don’t go through any of the other steps.
0:09:51 Google’s a good example of this.
0:09:52 Like they’re pushing hard on this.
0:09:54 Like the consumer version of this, right, is search, right?
0:09:58 So search has been, it’s been the 10 blue links
0:09:59 for 25 years now.
0:10:02 What Google’s, they talk about this publicly,
0:10:04 what they’re pushing towards this is just like,
0:10:05 no, it should be that answer.
0:10:06 Which is what they’re trying to do
0:10:07 with their voice UIs.
0:10:10 And so that concept might really generalize out, right?
0:10:11 And then everything gets rebuilt.
0:10:12 – Right.
0:10:16 So one of the new interfaces to AI
0:10:18 that people are talking about is voice
0:10:20 as the new interface.
0:10:24 What are we likely to get wrong about voice?
0:10:26 – Yeah, so I think the thing that,
0:10:27 if we’re gonna get something wrong about voice,
0:10:29 I think it’s gonna be that it would be a one-to-one
0:10:31 replacement for existing user interaction models
0:10:34 so that it would be like a replacement for keyboard
0:10:36 or that it’d be a replacement for the mouse or for touch.
0:10:39 Probably not, ’cause it’s a different modality, right?
0:10:42 It’s, you know, we know exactly what to keep.
0:10:43 After all this time, we know what the keyboard is for,
0:10:46 we know what touch is for and for voice
0:10:48 to displace those, seems like a stretch.
0:10:53 On the other hand, to the previous question,
0:10:55 there has been this turning point reached,
0:10:58 it feels like in AI applied to language
0:11:01 and from there to voice, right, to text and to speech.
0:11:03 Which is, it feels to us in the technology
0:11:05 like the natural language processing methods
0:11:06 that people have been working on for, again,
0:11:08 for 50 years, computer scientists have been working
0:11:11 on getting computers to understand basically speech.
0:11:13 And what we’re seeing now is in the technology
0:11:14 is that that now has started to work
0:11:15 in the same way that machine vision
0:11:18 started to work about seven years ago.
0:11:20 And so if that’s the case, then all of a sudden
0:11:23 the conversational UIs are about to get much better.
0:11:25 And again, and then you couple that with, okay,
0:11:26 what are you actually trying to achieve
0:11:29 when you talk to a computer?
0:11:30 Are you actually trying to like, you know,
0:11:31 are you trying to write a document?
0:11:32 Are you trying to read an email?
0:11:34 Are you trying to like do all these other things
0:11:34 that you do today?
0:11:36 Or are you fundamentally gonna be doing something different?
0:11:38 ‘Cause the machine’s gonna be so much smarter.
0:11:40 And I think that’s a very interesting open question.
0:11:43 – When I think about the AR mirror world,
0:11:45 I find it very hard to imagine it without it
0:11:48 having a voice component where we can understand
0:11:50 what you’re saying besides what you’re looking at,
0:11:52 that is that an essential part of the AR world?
0:11:54 – Yeah, I think actually I’d go so far as to say
0:11:56 it may be the case that voice actually is the key
0:11:59 to the AR world, like voice may be the thing.
0:12:01 Voice may actually be the foundation of the whole thing.
0:12:03 You know, for, this is kind of a cliche at this point,
0:12:05 but like the Apple AirPods,
0:12:06 I think were a fundamental breakthrough.
0:12:08 Like it’s again one of these funny things where it’s like,
0:12:10 okay, wireless headphones, okay, cool.
0:12:12 Like wireless headphones where there’s, you know,
0:12:14 there’s not even a wire connecting the two things.
0:12:15 Cool, okay, it seems like more of the same,
0:12:18 but you know, if you want, the experience you can have now
0:12:21 is like you can wear one of these things basically all day
0:12:22 and you can talk to it all day.
0:12:24 And you know, they’re getting, you know,
0:12:25 the new versions are getting better.
0:12:27 You know, and Siri and Google Now and Croton
0:12:30 and all these things are getting really good really fast.
0:12:33 And so it may be that we have just this constant
0:12:35 ongoing running dialogue.
0:12:37 This is kind of, you know, basically the machine
0:12:38 talking to our ear.
0:12:40 And then, you know, the visual overlay of AR
0:12:42 will obviously be important and valuable,
0:12:43 but it might be, it might,
0:12:45 the visual overlay might be supportive
0:12:47 on top of the voice experience.
0:12:51 – And we could very quickly have universal language
0:12:52 translation speaking over the years.
0:12:55 And I think people underestimate the change
0:12:56 that that would bring about in the world.
0:12:59 You’d have millions of people who are highly skilled
0:13:02 in everything except the skill of English.
0:13:05 Now being able to participate in a global economy.
0:13:08 We were talking about unexpected and unexpected things.
0:13:13 Biology, which is a million times as complicated as digital.
0:13:16 We’re now talking about a biotech revolution.
0:13:20 Are we misunderstanding what biotechnology actually is?
0:13:22 – Yeah, so that’s the big bet that we’ve made
0:13:26 with our bio effort that we started a few years back.
0:13:29 We think biological science is at a turning point
0:13:31 at the scientific level and we think it’s at a turning point
0:13:35 from basically being a process of discovery
0:13:38 of how biology works to being able to engineer biology.
0:13:40 And up to including literally being able
0:13:42 to program biology, being able to actually basically
0:13:45 be able to use electrical engineering and computer science
0:13:47 and these mechanical engineering and these kind of fields
0:13:49 for engineering things and be able to apply
0:13:52 those kinds of skills to biology.
0:13:55 If we’re right about that, then the whole concept
0:13:58 of kind of how bio and biotech work might be
0:14:00 on the verge of really changing.
0:14:03 The most obvious application that would be in pharmaceuticals,
0:14:05 there’s this concept of drug discovery.
0:14:06 It’s always the word discovery.
0:14:08 It’s always like, discovery sounds great.
0:14:09 It’s like, it’s optimistic.
0:14:13 It’s like, ooh, discovering things is fantastic.
0:14:16 The problem is, discovering, they literally call it that
0:14:18 ’cause they literally have to run all these experiments
0:14:19 and try to discover the drug that works.
0:14:22 Like try to kind of reverse engineer back from nature.
0:14:25 And the problem is sometimes they discover it
0:14:26 and sometimes they don’t.
0:14:29 So the example we always give is we talk about
0:14:32 with computers, we’ve been on this kind of 50 year track
0:14:34 of what’s called Moore’s Law.
0:14:35 We’re at chips beginning faster and cheaper
0:14:37 every year for a long time.
0:14:39 In biology, in drug discovery,
0:14:41 there’s what they call e-rooms law,
0:14:43 which is more spelled backwards, e-room.
0:14:46 And it’s the cost of discovering a new drug.
0:14:48 And it’s exactly the wrong direction.
0:14:51 It’s up and to the right, billions of dollars now.
0:14:54 And so if you could actually engineer biology,
0:14:56 then all of a sudden you can start to apply
0:14:57 this like just these decades of skills
0:14:59 that we’ve built up in how to engineer things
0:15:00 and be able to do things like
0:15:02 engineer new pharmaceuticals from scratch.
0:15:05 – And it all runs on basically ultimately Moore’s Law.
0:15:08 Moore’s Law has been foundational to this year.
0:15:10 It’s almost hard to imagine anything we have
0:15:13 in the modern world today without Moore’s Law.
0:15:18 Do you think Moore’s Law has another 30 years run?
0:15:19 Is it limited?
0:15:20 Is it finite?
0:15:21 Will it go on forever?
0:15:23 We’ll define Moore’s Law in the broadest sense
0:15:28 of computers getting cheaper by half every couple of years.
0:15:30 So what’s your take on Moore’s Law?
0:15:32 – Yeah, so the traditional definition is computer
0:15:35 in the form of the chip, and then specifically a chip.
0:15:36 So Moore’s Law has always been expressed
0:15:39 as kind of unit one of chip.
0:15:40 And that could be right, that could be a CPU
0:15:41 or it could be a graphics card
0:15:44 or it could be a graphics chip or a memory chip.
0:15:45 And then specifically what you were doing was
0:15:48 you were able to put more transistors on that chip
0:15:49 for the same cost.
0:15:51 And actually for a long time as you did that,
0:15:53 you were actually able to reduce the power requirement
0:15:56 for per transistor, which was this kind of added benefit.
0:15:58 And so chips kind of got simultaneously,
0:16:00 they got faster, they got cheaper,
0:16:01 and they got more power efficient.
0:16:04 And that was kind of a cornucopia effect that generated,
0:16:06 as you said, most of what you see today
0:16:08 in the computer industry.
0:16:11 So the bad news is that that in that form
0:16:13 seems to be coming to something of an end,
0:16:15 which is we’re too good at it.
0:16:17 We’ve hit basically, we being the semiconductor industry
0:16:19 broadly, the tech industry have kind of hit
0:16:20 the limits of fundamental physics.
0:16:23 Like we’re now down at the sort of deep atomic level.
0:16:25 And it’s becoming much harder to make,
0:16:26 there’s still progress,
0:16:27 it’s becoming much harder to make progress
0:16:29 at the per chip level.
0:16:31 The good news is that the industry starting
0:16:34 10 or 15 years ago, the computer industry broadly
0:16:36 refocused off of what you do with a chip
0:16:39 to what you do with a large number of chips, right?
0:16:41 So kind of the old model of a chip was
0:16:42 you make the chip more powerful
0:16:42 ’cause you’re trying to scale up
0:16:44 what you can do in the chip.
0:16:46 The new model is you use thousands of chips in parallel
0:16:48 and you have this kind of approach to scaling out.
0:16:49 And of course the full form of that
0:16:51 is what’s now known as the cloud.
0:16:53 And so we now have a 15 year head of steam going
0:16:55 to basically be able to get good
0:16:58 at using lots of chips to do things.
0:17:00 And that’s why you see the continued ability to, right?
0:17:03 To accelerate, you know, many, many things
0:17:05 that you deal with are getting still much faster
0:17:06 as if they’re still on the Moore’s Law.
0:17:08 The experiences you’re having are getting faster.
0:17:09 So we think number one,
0:17:10 like the rise of scale at architectures
0:17:11 is a really big deal.
0:17:13 Like, you know, in modern clouds as a developer,
0:17:15 you don’t really care about what the power
0:17:15 of any particular chip is.
0:17:17 You just like light up some more of them
0:17:18 and they don’t cost much.
0:17:20 So there’s that.
0:17:22 The other thing is chips are now specializing.
0:17:23 And in particular, you’ve got the rise
0:17:24 of these new dedicated chips
0:17:25 for things like neural networks
0:17:28 where there’s another level of opportunity to optimize.
0:17:31 And then the other kicker is the programmers.
0:17:34 Software, people like me get to step up.
0:17:38 In the old days, when computers were expensive,
0:17:39 programmers were really good at optimizing
0:17:42 every single step of a software program.
0:17:44 Programmers got out of that habit,
0:17:45 probably starting 30 years ago,
0:17:47 where it didn’t matter as much anymore
0:17:49 because Moore’s Law was working so well.
0:17:52 And so software today is just like massively inefficient.
0:17:53 There’s actually, I forget the name,
0:17:55 there’s something called Worth’s Law,
0:17:58 which is, it was written at the time,
0:17:59 I don’t know if it still holds,
0:18:00 but it was, somebody did benchmarks
0:18:04 of you take Microsoft Office 2000 on a PC from 2000
0:18:07 and you take Microsoft Office 2007 and a PC from 2007
0:18:09 and every function you could do,
0:18:12 you could now do in twice the time, right?
0:18:13 So like literally like,
0:18:16 the old adage in tech in the 90s was
0:18:18 when Andy Grove was running Intel
0:18:19 and Bill Gates was running Microsoft,
0:18:21 it was Andy Gibb in the form of Moore’s Law
0:18:25 and then Bill take it away in the form of software bloat.
0:18:28 And so, and Worth’s Law literally
0:18:31 is a mathematical proof of that.
0:18:33 And so like it’s become prime time again
0:18:35 for software programmers to get really good at optimization,
0:18:37 which is like what’s happening in AI world
0:18:39 and also in the cryptocurrency world.
0:18:40 And so with those different approaches,
0:18:42 it feels like we’ve got,
0:18:43 it feels like decades of advances ahead
0:18:46 that aren’t purely dependent on classic Moore’s Law.
0:18:48 – And because if we take the long-term,
0:18:52 like thinking of a 100 year span to have prosperity
0:18:55 like we’ve seen would kind of require that computer power
0:18:59 sort of get cheaper every year because of a dent
0:19:01 that it’s hard to imagine a world like that.
0:19:04 So is your confidence that we could kind of keep this going
0:19:07 based on just sort of human ingenuity
0:19:11 or do you think that there’s some basic principles
0:19:14 of science that suggest that we’re just at the beginning
0:19:15 of what we can discover?
0:19:17 – Well, so Gordon Moore who invented Moore’s Law
0:19:18 as co-founder of Intel,
0:19:20 he always said Moore’s Law is what was interpreted
0:19:23 as a prophecy and he always said it was not a prophecy.
0:19:25 It was a goal, right?
0:19:27 And it was a goal of basically what you could do
0:19:28 if you focused intensely,
0:19:30 if you focused the entire industry intensely
0:19:32 on a set of engineering optimizations, right?
0:19:34 Over a long period of time.
0:19:35 So he used to say it’s just like,
0:19:37 there’s nothing inevitable about it.
0:19:39 It’s a consequence of thousands and then tens of thousands
0:19:41 and then millions of engineers like working to actually
0:19:44 deliver on these kind of semi-arbitrary goals.
0:19:47 And so I think the answer to that is
0:19:49 we have many, many areas of improvement.
0:19:52 As I said, the problem is we don’t have the one that we had,
0:19:54 which is this transistor doubling kind of effect.
0:19:56 But we’ve got many, I mean,
0:19:58 there’s far more engineers working on all this stuff today
0:20:00 than we’re working on it in 1965.
0:20:01 When he invented Moore’s Law,
0:20:03 we’re in 1995 when everybody bought a PC.
0:20:08 Like we have a lot of mind power going into this.
0:20:12 We’ve got a lot of different technological options.
0:20:14 We’ve got a lot of, you know, incredibly impressive work
0:20:15 happening all over the world.
0:20:16 The other thing is you can’t, you know,
0:20:19 you never like, you know,
0:20:21 one of those things like the transistor was not obvious
0:20:22 and then they invented that.
0:20:25 And then this integrated microchip was like not obvious
0:20:26 and then they invented that.
0:20:27 And so you don’t quite know, you know,
0:20:28 there are lots of technical proposals
0:20:30 for how to get to the next level of Moore’s Law.
0:20:31 You know, so there’s all kinds of theories
0:20:32 around optical computing
0:20:33 and then in the long run, biological computing.
0:20:34 – Quantum computing.
0:20:36 – Quantum computing, exactly.
0:20:38 And so over the course of the next like 20 years,
0:20:40 like, we’ll put it this way.
0:20:42 This is one of the world’s largest prizes, right?
0:20:45 If you’re the engineer who figures out
0:20:46 how to reaccelerate Moore’s Law
0:20:48 or how to shift computing onto a new substrate
0:20:49 like biology, that is the thing to do.
0:20:52 And so that’s the prize.
0:20:54 And that historically has been pretty motivating.
0:20:54 – Right.
0:20:57 So taking this kind of theme of marching forward
0:21:02 progressively, we have 4G, we’re talking about 5G.
0:21:05 So far 5G seems to be faster 4G
0:21:06 with a lot of hype added to it.
0:21:09 There’s a technical specification for 5G,
0:21:11 which is really awesome.
0:21:14 You know, 100 gigabytes, two millisecond latency,
0:21:16 almost impossible.
0:21:19 Are you counting on that for the next decade
0:21:21 that we’re gonna have actual,
0:21:23 what they promise with 5G?
0:21:24 – Yes, I think there’s pretty good odds we will.
0:21:27 And the reason is because 5G has become
0:21:29 a national geopolitical battle.
0:21:31 Like it’s actually a very interesting twist.
0:21:33 It’s become actually a primary, like, you know,
0:21:35 if the Cold War between the US and the USSR
0:21:37 was like defined by the space race,
0:21:40 like at least the sort of nascent Cold War with China
0:21:43 is actually, a lot of it is around 5G, interestingly enough.
0:21:45 I mean, it could have been around a lot of things,
0:21:46 but it happens that it’s around 5G.
0:21:49 And so you now have nation states
0:21:51 that very, very badly need to win,
0:21:54 two big nation states in particular.
0:21:57 And so I think there’s gonna be a lot of, you know,
0:21:58 so we’re gonna start with the payoff from the space races,
0:22:00 like all the products that got, you know,
0:22:02 spun off from that, satellites and GPS
0:22:02 and everything else.
0:22:05 The other thing on 5G, you know,
0:22:07 people sometimes say 5G will lead to applications
0:22:08 they haven’t even thought of yet.
0:22:09 And I think that’s kind of true.
0:22:11 But I look at it a little bit differently.
0:22:13 It’s just a little bit like the most law conversation
0:22:14 we’re having, which is,
0:22:15 I look at a little bit as a math kind of question,
0:22:18 which is there’s sort of three classic rules
0:22:20 for how networks scale
0:22:23 and how network scaling turns into value or usefulness.
0:22:24 And there’s sort of historically,
0:22:26 there’s what’s called Sarnoff’s law,
0:22:28 which was based on broadcast TV,
0:22:30 which is the value of a network is equivalent
0:22:31 to the number of nodes, right?
0:22:33 So it scales with N, right?
0:22:36 So a TV network with 10 million viewers
0:22:38 is twice as valuable as a TV network of 5 million viewers.
0:22:39 That’s kind of the obvious one.
0:22:40 Then there was Metcalf’s law,
0:22:42 which is basically the value of the network
0:22:44 is on the number of connections between two points.
0:22:47 And that’s like how email works, right?
0:22:49 It just emails a person to person.
0:22:51 And that’s correlates to N squared.
0:22:52 So the value of the network rises exponentially
0:22:54 with N squared.
0:22:55 And then there’s this thing called Reed’s law,
0:22:57 which is called the group forming law,
0:22:59 which is the value of the network is proportional
0:23:00 to the number of groups and subgroups
0:23:02 that conform inside the network,
0:23:04 which turns out to be two to the N.
0:23:06 And if you wanna have fun in your plane flights home,
0:23:09 it’s like, just go on Excel and like chart,
0:23:11 N, N squared and two to the Nth, right?
0:23:13 And two to the Nth just goes like straight vertical.
0:23:15 Like you can’t even put them on the same chart.
0:23:17 And two to the Nth is like what’s now happening
0:23:18 was like social networks, right?
0:23:20 So like Facebook groups and all these other things,
0:23:21 like WhatsApp groups and all these other things
0:23:23 people do with social networks and games
0:23:24 and all these other things.
0:23:26 And so those are like the three ways
0:23:27 in which network growth pays off.
0:23:29 And like all three are working now
0:23:32 based on broadband, wired broadband.
0:23:32 They’re all working,
0:23:35 you see it happening very much with mobile.
0:23:36 The introduction of 5G,
0:23:37 the way I think about it is it’s gonna turbo charge
0:23:39 those three networks in particular,
0:23:42 that last one or those last two.
0:23:43 And so it’s gonna add a lot more N.
0:23:45 There’s just gonna be a lot more devices on the network.
0:23:47 There’re gonna be a lot more things
0:23:48 that those devices can do.
0:23:49 They’re gonna be a lot more point to point connections
0:23:51 that make sense to have.
0:23:52 There’s gonna be a lot more groups
0:23:54 that form a lot more economic activity that happens.
0:23:56 – Something that was again,
0:23:58 we were expecting to happen, but didn’t
0:24:01 was in the world of what’s sometimes called
0:24:02 the sharing economy.
0:24:05 There was a, after Airbnb and Uber,
0:24:08 there was a stampede of companies
0:24:09 that were gonna be Uber for X.
0:24:12 And then X was everything in the world.
0:24:13 Very few of them have succeeded.
0:24:16 Again, there was an expectation.
0:24:18 We see more of them, but we haven’t.
0:24:21 So is that whole idea kind of at a dead end?
0:24:24 Is it just, we’re in a very slow disruption.
0:24:25 It’s gonna take a while.
0:24:28 Like the generational requirements
0:24:31 we were talking about technology earlier.
0:24:32 Or is something else?
0:24:35 So what do you think happened there?
0:24:36 And what are we looking at?
0:24:38 – Once again, I object to the question.
0:24:39 – Okay.
0:24:41 – Throw the gavel.
0:24:43 So I look at it a little bit differently,
0:24:45 which is the, this is something we try hard to do
0:24:47 in our place.
0:24:48 It is very tempting.
0:24:50 And we do have this conversation all the time
0:24:51 at our place of like, okay, what about the trend?
0:24:52 What about the theme?
0:24:54 Right, what about the variations on the theme?
0:24:56 Kind of as you said, and this is something happened.
0:24:58 When something wins back, you always get this kind of,
0:25:00 we describe it as kind of the Hollywood model
0:25:03 of, it’s like, what’s your new movie about?
0:25:05 It’s Pretty Woman Meets the Rock, right?
0:25:06 Or whatever.
0:25:10 And so in the Valley, it’s super for X,
0:25:11 or most recently, superhuman for X,
0:25:12 which I’m very excited about,
0:25:13 is one of the big new trends
0:25:16 after another one of our companies.
0:25:20 So, but I don’t think it’s really that.
0:25:22 That’s not how the great ideas arrive.
0:25:23 They don’t look like that.
0:25:25 They look like very specific,
0:25:27 they look at very specific theories,
0:25:28 not general theories.
0:25:29 So they tend to be very specific
0:25:32 to the details of the market involved.
0:25:33 One of the things that I think we’ve learned
0:25:36 about ride sharing, why ride sharing works so well,
0:25:37 I mean, it worked well for many reasons.
0:25:38 One of the reasons it works so well as an idea
0:25:39 is because as long as the driver’s good,
0:25:41 as long as they’re rated at a certain level,
0:25:42 it doesn’t really matter who the driver is.
0:25:44 So like one of the classic examples
0:25:46 was Uber for cleaning your house or your apartment.
0:25:48 And it just, it turns out you just don’t want
0:25:49 a different person over every week
0:25:52 to clean your house like it’s a problem.
0:25:54 And so there’s a lot of these kinds of,
0:25:56 I would say, simple, you know,
0:25:58 sort of the simple applications to the idea
0:25:59 that don’t necessarily work.
0:26:01 Now, by trying all those ideas,
0:26:02 you kind of map the idea space
0:26:03 and you start to get a better sense
0:26:04 of like what your overall structure is.
0:26:06 And I think what’s happening now
0:26:08 is you’re starting to see another set of companies
0:26:09 coming out the other end
0:26:10 that have kind of fully internalized that lesson
0:26:12 and have figured out new models that work.
0:26:13 And so my favorite example,
0:26:15 one of my companies called Honor,
0:26:17 so Honors, you might think loosely,
0:26:19 might think of it as kind of Uber for senior care,
0:26:22 for in-home care for seniors.
0:26:24 It’s a loose model.
0:26:25 Actually, it turns out it’s a very loose model
0:26:26 for a couple of reasons.
0:26:29 One is it’s really deeply not a fungible service.
0:26:30 Like if you have an aging parent,
0:26:32 you actually very much don’t want somebody different
0:26:34 to show up all the time.
0:26:35 You want the same person.
0:26:36 And so in that case, for example,
0:26:39 Honor actually has a full-time employment relationship,
0:26:42 salaried employment relationships with the workers.
0:26:43 Right, which of course is very different
0:26:45 than the Uber and Lyft model.
0:26:47 It actually turns out the matching problem
0:26:48 is much more complicated.
0:26:50 Right, because when you’re matching human beings
0:26:52 in somebody’s home, there’s like 20 variables
0:26:53 that you need to match on
0:26:55 so that everybody’s comfortable with the experience.
0:26:58 As an example, in some cases, you literally need people
0:27:00 with the physical strength to be able to lift people
0:27:00 when you’re caring for them.
0:27:02 You do want to be able to do this kind of
0:27:04 multi-dimensional mapping, and that model’s really working.
0:27:07 And so I think we’re gonna see a whole set of these.
0:27:09 Like I think there’s a big kind of vista of exploration
0:27:11 that’s gonna happen from here.
0:27:12 And I would suspect there will be dozens
0:27:14 of hundreds of new models that people figure out.
0:27:16 – So speaking of new models,
0:27:18 do you ever think about new models
0:27:19 for the VC industry itself,
0:27:22 and how you would apply the principles of innovation
0:27:26 and disruption to what you do in general?
0:27:28 So as you look out 30 years,
0:27:30 what kinds of innovations would you expect
0:27:33 in the basic business that you’re in?
0:27:36 – Yeah, so there’s something very timeless about Venture,
0:27:37 which is there’s actually a new book out called,
0:27:39 literally called VC.
0:27:41 It actually tells the story that has been kind of hard
0:27:43 to get at for a long time in a really clear way,
0:27:45 which is the modern venture model is actually,
0:27:46 one of the historical precedents for it
0:27:50 was actually how whaling expeditions got financed
0:27:53 in the 1600s, so coming up on 500 years ago.
0:27:58 So whaling of, it was literally like, okay,
0:28:00 you’re gonna have like a ship with a captain and a crew
0:28:02 that’s gonna go out and try to like bring back a whale.
0:28:03 Right, and so it’s like a problem number one is like,
0:28:05 only two thirds of the ships are gonna come back, right?
0:28:07 So like high failure rate.
0:28:10 Two is like, okay, what the ship is really matters,
0:28:11 who the captain is really matters,
0:28:13 how do you know who good captain is,
0:28:15 and then what’s a good crew,
0:28:18 and are they gonna be willing to follow the captain?
0:28:20 And then there’s all these like strategy questions,
0:28:21 like do you want the captain who knows
0:28:23 where all the whales have been caught recently,
0:28:25 so they go there, or do you want the captain that says,
0:28:27 no, that’s Gary’s gonna be over fish,
0:28:29 do you wanna go someplace else?
0:28:31 And so literally all the whaling voyages,
0:28:34 like in the colonies 500 years ago,
0:28:37 got financed with basically angel syndicates,
0:28:39 basically venture capital effectively.
0:28:41 And then literally the term carry,
0:28:42 which is sort of how VCs get paid,
0:28:44 the so-called carried interest,
0:28:46 which is like the 25% that you make,
0:28:48 or the profits that you share,
0:28:52 the term carry actually was the percentage of the whale
0:28:54 that the ship carried.
0:28:55 It was literally physical carry.
0:28:56 It was literally that part of the whale,
0:28:58 like that’s where that term came from.
0:29:01 And so there’s a timelessness to the art of trying
0:29:04 to figure out how to finance these kind of expeditions
0:29:08 into the unknown that is likely to endure.
0:29:09 The big question for me is,
0:29:11 how will the shape of the companies,
0:29:14 or let’s say the ventures themselves change, right?
0:29:17 And so today there’s like a well-known understood template
0:29:20 for kind of the prototypical Silicon Valley venture
0:29:23 investment, and it’s like a company in a certain place.
0:29:25 It’s a C corporation, it’s domicile in the US,
0:29:26 it’s financed a certain way,
0:29:28 and to certain types of employees,
0:29:30 a certain relationship with its employees, and so forth.
0:29:35 30 years from now, are we gonna be financing companies here,
0:29:40 or anywhere, or in two places, 50 places, 500 places,
0:29:42 are the companies still gonna have physical place,
0:29:44 or are they gonna be fully virtual?
0:29:45 Are they gonna be companies,
0:29:47 or are they all gonna be blockchains, right?
0:29:49 Are they gonna have actual employment relationships,
0:29:51 or are they gonna have basically developers
0:29:52 and center through cryptocurrency?
0:29:53 That’s a real model.
0:29:56 And so I think the big question is like,
0:29:57 we don’t even know what the shape of companies
0:29:59 is gonna look like, or ventures is gonna look like
0:30:00 in 30 years.
0:30:01 So if I could figure that out,
0:30:03 then I could answer what venture looks like.
0:30:05 Without that, I think it’s hard to say.
0:30:08 – Okay, so we were tempted to do a little bit
0:30:10 of long-term thinking, and long-term thinking
0:30:15 is sort of rare and often ignored,
0:30:20 whereas civilizations demanded as being necessary.
0:30:22 So do you have any suggestions
0:30:25 about how long-term thinking could be applied
0:30:28 in Silicon Valley, and whether you have even
0:30:30 any suggestions to the people in this room
0:30:34 about how they could use long-term thinking?
0:30:36 – Yeah, so the thing I’ve always found about long,
0:30:39 I think long-term thinking is of course central.
0:30:40 Essentially one of the things about the valley
0:30:42 that I find outsiders miss the most,
0:30:45 which is it feels like it’s all moving so fast,
0:30:47 and yet like any of the important companies
0:30:48 and any of the important products
0:30:49 take like a decade or more to build.
0:30:51 And so it’s like everything important
0:30:52 basically takes a long time.
0:30:54 And so a lot of it actually feels quite slow.
0:30:58 And so long-term orientation is absolutely necessary,
0:30:59 and I think we probably all agree
0:31:01 there’s not enough of it in the world.
0:31:03 The thing about long-term thinking I’ve found is like,
0:31:06 it’s really easy if you know the thing is gonna work.
0:31:10 Like, boy, that’s completely straightforward.
0:31:12 Like let’s go on a 10-year journey to a place
0:31:14 where we know it’s gonna be great.
0:31:15 The problem is it’s long-term thinking
0:31:17 crossed with uncertainty, right?
0:31:18 And quite possibly fatality,
0:31:20 like the thing may just simply not work
0:31:22 for any of a thousand reasons.
0:31:23 And so that’s the issue.
0:31:25 And so I think the issue is less around long-term thinking.
0:31:27 I think the issue is more about how to deal with risk
0:31:28 and how to deal with uncertainty
0:31:31 and how to make really big consequential decisions
0:31:35 in the face of literally an unknowable future landscape.
0:31:37 And for there, I mean,
0:31:39 this is kind of the one kind of secret weapon of venture.
0:31:41 It’s like venture is the worst of all asset classes
0:31:44 in a lot of ways in that it’s like it’s a liquid
0:31:45 and it’s like incredibly volatile
0:31:48 and it’s like hit and miss in this kind of crazy way.
0:31:50 The one thing that venture really has going for it
0:31:52 as an asset class is we have the concept
0:31:55 of the portfolio kind of wired into the model
0:31:56 in which you just kind of assume,
0:31:58 in top-end venture, you just kind of assume,
0:31:59 fundamentally it’s half the company’s
0:32:01 gonna work half of them aren’t.
0:32:02 Right, and then the classic, right?
0:32:03 The classic, the cliche is like the ones that work,
0:32:05 then you have to work enough so that they pay
0:32:09 for the ones that don’t to make the whole enterprise work.
0:32:11 And so if you can adapt yourself
0:32:14 from the mentality of will this thing work, right,
0:32:19 to will this portfolio of things basically pay off, right?
0:32:20 Will enough things work
0:32:21 that they’ll actually pay for the portfolio?
0:32:22 Then at that point, you can start to make risk
0:32:26 a somewhat tractable thing to contemplate.
0:32:28 It’s still hard to divorce yourself emotionally from it,
0:32:30 right, ’cause it’s just like it’s still like absolutely,
0:32:31 you know, it’s just terrible
0:32:33 when any of the individual things don’t work,
0:32:35 but at least you have a conception of framework
0:32:37 for you to be able to make 10 long run bets
0:32:39 and being able to get to the other side.
0:32:41 Now, the response that I have to get to that is oh,
0:32:43 that’s great if you’re a VC, the problem is you’re a portfolio,
0:32:45 you know, you’re a founder or a CEO,
0:32:46 like you don’t get that, right?
0:32:48 You have the much harder version of the problem,
0:32:50 which is you’re on a one-way journey,
0:32:52 like you’re the captain of the whaling ship.
0:32:53 Yeah, there’s all those other captains over there,
0:32:57 but like, you know, they’re on their own, you’re on your own.
0:32:58 Even there though, you know,
0:33:01 the best run companies tend to run experiments,
0:33:05 they tend to run multiple experiments against their goals,
0:33:07 and they certainly run those experiments sequentially
0:33:09 as they kind of, you know, try to figure out what works
0:33:11 in a lot of cases, they run experiments in parallel,
0:33:12 as they’re trying to test different things.
0:33:14 And so I also think this kind of mentality
0:33:15 of sort of portfolio risk also applies
0:33:16 to how you run a company,
0:33:18 which is you want to basically,
0:33:20 you want to have a great deal of conviction
0:33:21 about where you’re trying to head,
0:33:23 but you want to have a lot of flexibility inherent
0:33:24 in how you’re going to get there, right,
0:33:25 and what the tactics are,
0:33:26 and then you want to be able to run
0:33:27 a lot of experiments against that,
0:33:29 and you can kind of diversify your risk
0:33:31 of any one theory by doing that.
0:33:33 – And that’s what governments are in some senses,
0:33:37 they have a portfolio of different kind of prospects
0:33:40 about the future, bits, I mean, some senses.
0:33:43 – So you think that’s an optimistic view of what governments do.
0:33:46 – Yeah, so, I mean, that’s what they’re,
0:33:49 and they’re adverse to risk, unfortunately.
0:33:50 – Well, the problem, the problem,
0:33:51 the problem governments have with risk
0:33:53 is like the end of one, right?
0:33:56 So there’s only one government per, right?
0:33:57 We only get to run, you know,
0:33:59 I mean, ex-federalism, which has been a huge advantage,
0:34:00 I think, for the U.S., but like, you know,
0:34:02 the U.S. national government only gets to run one scenario.
0:34:03 – Right.
0:34:04 – And running experiments in the population
0:34:06 is not necessarily well-received.
0:34:07 – Right, ’cause you can’t tolerate failure.
0:34:11 – Yeah, right, yeah, failure has real consequences, so.
0:34:14 – So there’s currently not the only introspection
0:34:16 about government, but also about capitalism,
0:34:21 and capitalism so far has depended on growth,
0:34:24 and growth is something that VCs pay attention to,
0:34:30 but we’re now wondering if what’s the minimum amount
0:34:32 of growth that you might need to have prosperity?
0:34:33 Can you have prosperity with low growth?
0:34:36 Can you have prosperity with fixed growth?
0:34:40 Do you have any insights about that
0:34:42 at the civilizational scale?
0:34:43 – Yeah, so I think, and actually I don’t even say
0:34:45 that the issue is even more intense these days,
0:34:47 ’cause there’s now very prominent people
0:34:49 in public life arguing that growth is bad, right?
0:34:53 And in fact, it’s, that in fact is ruinous and destructive,
0:34:54 and that the right goal might actually be
0:34:56 to have no growth or to actually go into negative growth,
0:34:58 then especially in the very common view
0:34:59 in the environmental movement.
0:35:02 So I’m a very strong proponent, a very strong believer
0:35:04 that growth is absolutely necessary,
0:35:06 and I’ll come back to the environmental thing in a second
0:35:08 ’cause it’s a very interesting case of this.
0:35:10 I think growth is absolutely necessary,
0:35:11 and I think the reason growth is absolutely necessary
0:35:13 is because you can fundamentally have two different
0:35:15 mindset views of how the world works, right?
0:35:18 One is positive sum, which is rising tide,
0:35:20 lifts all boats, we can all do better together,
0:35:24 and the other is zero sum, right?
0:35:25 Where for me to win, somebody else must lose,
0:35:26 and vice versa.
0:35:29 And the reason I think economic growth is so important
0:35:32 at core is because if there is fast economic growth,
0:35:34 then we have positive sum politics,
0:35:36 and we start to have all these discussions
0:35:38 about all these things that we can do as a society,
0:35:40 and if we have zero sum growth,
0:35:44 if we have a flat growth or no growth or negative growth,
0:35:48 all of a sudden the politics becomes sharply zero sum.
0:35:50 And the most, you just kind of see this
0:35:53 if you kind of track kind of the political climate,
0:35:55 you just, basically it’s the wake of every recession, right?
0:35:57 It’s just that in the wake of every economic recession,
0:36:00 the politics just go like seriously negative
0:36:03 on in terms of thinking about the world’s zero sum.
0:36:05 And then when you get a zero sum outlook in politics,
0:36:07 that’s when you get like anti-immigration,
0:36:08 that’s when you get anti-trade,
0:36:09 that’s when you get anti-tech.
0:36:11 If the world’s not growing, then all that’s left to do
0:36:14 is to fight over what we already have.
0:36:16 And so my view is like, you need to have economic growth.
0:36:18 You need to have economic growth for all of the reasons
0:36:20 that I would say right wingers like economic growth,
0:36:21 which is you wanna have higher levels of material
0:36:23 prosperity, more opportunity, more job creation,
0:36:25 all those things.
0:36:28 You wanna have economic growth for the purpose
0:36:30 of having like sane politics,
0:36:32 like a productive political conversation.
0:36:34 And then I think the kicker is you also want
0:36:36 economic growth actually for many of the things
0:36:38 that left wing people want.
0:36:39 One of the best books this year,
0:36:41 new books this year has got Andrew McAfee,
0:36:44 I was writing a book called I think More From Less.
0:36:46 And it’s actually a story of a really remarkable thing
0:36:48 that a lot of people are missing about what’s happening
0:36:52 with the environment, which is globally carbon emissions
0:36:55 are rising and resource utilization is rising.
0:36:58 In the US, carbon emissions and resource utilization
0:36:59 are actually falling.
0:37:02 And so in the US, we have figured out to grow our economy
0:37:04 while reducing our use of natural resources,
0:37:07 which is a completely unexpected twist, right,
0:37:08 to the plot of what kind of,
0:37:10 if you listen to environmentalists in the 60s and 70s,
0:37:11 like nobody predicted that.
0:37:14 And it turns out, he talks about this in the book,
0:37:16 but it turns out basically what happens is economies,
0:37:18 when economies advance to a certain point,
0:37:20 they get really, really good at doing more with less, right?
0:37:23 They get really, really good at efficiency.
0:37:25 And they get really good at energy efficiency,
0:37:27 they get really used in environmental resources,
0:37:29 they get really good at recycling
0:37:30 in lots of different ways.
0:37:31 And then they get really good at what’s called
0:37:33 dematerialization, which is what is happening
0:37:35 with digital technology, right?
0:37:37 Which is basically taking things that used to require
0:37:38 atoms and turning them into bits,
0:37:41 which inherently consumes less resources.
0:37:42 And so what you actually want,
0:37:44 like my view on environmental issues is like,
0:37:46 you’ve got a global problem,
0:37:48 which is you have too many people in too many countries
0:37:52 stuck in kind of mid the industrial revolution,
0:37:54 they’ve got to grow to get to the point
0:37:55 where they’re in a fully digital economy,
0:37:57 like we are precisely so that they can start
0:38:00 to have declining resource utilization, right?
0:38:01 I mean, the classic example is energy.
0:38:04 Like, you know, the big problem with energy emissions globally,
0:38:06 like a huge problem with emissions and with health
0:38:09 from emissions is literally people burning wood,
0:38:11 like in their houses, right, to be able to heat and cook.
0:38:13 And what you want to do is you want to go to like
0:38:15 hyper-efficient solar or ideally nuclear, right?
0:38:16 You want to go to these like super advanced forms
0:38:18 of technology.
0:38:20 So you want that, and then by the way,
0:38:22 if you want like a big social safety net,
0:38:23 you know, in all the social programs,
0:38:25 you want to pay for that stuff.
0:38:27 You also want economic growth because that generates taxes
0:38:27 that pays for that stuff.
0:38:30 And so like growth is the single kind of biggest
0:38:32 form of magic that we have, right?
0:38:33 To be able to like actually make progress
0:38:35 and hold the whole thing together.
0:38:38 – And to your point about the developing countries,
0:38:40 I think the idea of leapfrogging technology is a myth.
0:38:41 It doesn’t really work.
0:38:42 You actually have to,
0:38:45 if you want to have a high tech infrastructure,
0:38:48 you actually need the intermediate roads, clean water.
0:38:50 You can’t skip over that.
0:38:52 And so they all need to be built out in order to
0:38:54 have that prosperity at the end.
0:38:58 So, you know, it seems like you don’t worry about much.
0:38:59 I don’t worry about much.
0:39:02 But one thing I do worry about is cyber conflict,
0:39:05 cyber war, partly because I think we have no consensus
0:39:07 about what’s allowable.
0:39:09 Does this worry you at all?
0:39:11 – So I think there’s a lot of unknownness to it.
0:39:13 I think people are trying to figure this out,
0:39:16 but it’s a complex issue to grapple with.
0:39:18 I will make an optimistic argument,
0:39:20 which is gonna sound a little strange.
0:39:24 If you kind of project forward what’s happening
0:39:26 with generally a cyber, with information,
0:39:28 you know, operations of different kinds,
0:39:31 but also with drones, you know, UAVs.
0:39:35 And then also with, you know, unmanned fighter jets, right?
0:39:39 Unmanned, you know, ships increasingly being built.
0:39:42 You know, there’ll be unmanned submarines at some point.
0:39:44 If you project this stuff forward,
0:39:46 you start to get this very interesting potential world
0:39:49 in which basically the way I think about it
0:39:51 is like all human conflict between peoples
0:39:53 or between nation states up until now
0:39:56 has been basically throwing people at each other, right?
0:39:57 Throwing soldiers at each other
0:39:59 and like letting them make the decision of who to shoot
0:40:00 and like hoping they don’t get shot,
0:40:02 like with very serious repercussions
0:40:03 of all those individual human decisions.
0:40:06 You do have the prospect of basically a new world
0:40:07 of both offense and defense.
0:40:08 It’s like completely motorized,
0:40:10 completely mechanized, completely software driven
0:40:11 and technology driven.
0:40:12 And a lot of people, it’s just immediately like,
0:40:14 oh my God, that’s horrible.
0:40:15 You know, Terminator, like, you know, Skynet,
0:40:17 like, you know, this is just the worst thing ever.
0:40:19 There’s a novel called “Kill Decision.”
0:40:21 If you want the dystopian theory,
0:40:22 there’s a novel called “Kill Decision.”
0:40:23 – By Daniel Suarez.
0:40:25 – Daniel Suarez that extrapolates the drones forward
0:40:28 and it’ll keep you up late at night.
0:40:30 But the optimistic view would be like, boy,
0:40:33 isn’t it good that there aren’t human beings involved?
0:40:33 Isn’t it good?
0:40:35 Like if the machines are shooting at each other,
0:40:36 like isn’t that good?
0:40:38 Isn’t that better than if they’re shooting at us?
0:40:39 And by the way, and by the way,
0:40:41 I would go so far as to say like,
0:40:43 I don’t know that I’m in favor of like the machines
0:40:45 making like kill decisions, like decisions on who to shoot.
0:40:48 But like the one thing I know is humans do that very badly.
0:40:50 Very, very, very badly.
0:40:51 I’m the opposite of pro war.
0:40:53 I don’t want to see any of this stuff actually play out.
0:40:55 But if it has to play out, maybe having it be software
0:40:56 and machines is going to be actually a better outcome.
0:40:57 – Right.
0:41:00 I mean, it’s this kind of weird that we don’t allow,
0:41:02 we don’t want machines to kill humans.
0:41:03 We want other humans to kill humans.
0:41:05 – We want 18 year olds.
0:41:07 We want to take 18 year olds out of their homes, right?
0:41:08 And we want to put a gun in their hand
0:41:10 and send them someplace and tell them to decide who to shoot.
0:41:13 Like that that is going to go down in history
0:41:14 is having been a good idea.
0:41:16 Just strikes me as like unlikely.
0:41:19 – So we have only time for one last question, which is,
0:41:22 I’m usually, I claim to be the most optimistic person
0:41:24 in the room, but with you sitting across from me,
0:41:26 I don’t think that may be true.
0:41:29 What is your optimism based on?
0:41:34 – So my optimism, okay, so get cosmic for a second.
0:41:35 – Why not?
0:41:35 – I guess we’re here.
0:41:36 – It’s the last question.
0:41:37 – It’s the last question.
0:41:41 – So the science fiction authors always talk about
0:41:42 what’s called the singularity.
0:41:45 This kind of singularity, so the singularity is basically
0:41:46 what happens when the machines get so smart
0:41:48 that all of a sudden everything goes into exponential mode
0:41:52 and all of a sudden the entire world changes.
0:41:54 So my reading history is actually,
0:41:56 we actually were in the singularity already
0:41:59 and that it actually started 300 years ago.
0:42:03 And if you look at basically, if you look at basically
0:42:05 any chart of human welfare over time,
0:42:07 and you can look at, Child Mortality’s an obvious one,
0:42:10 but there’s many, many, many others,
0:42:12 and you just look at progress on that metric.
0:42:13 Just look at Child Mortality as an example
0:42:15 and it’s just basically flat, flat, flat, flat, flat, flat,
0:42:17 flat for like 50,000 years.
0:42:20 And this is the famous, Thomas Hobbes,
0:42:22 life is nasty, brutish and short, right?
0:42:23 It was just like the thing,
0:42:25 like everything was terrible everywhere,
0:42:28 all the time, forever, the end,
0:42:29 until 300 years ago and all of a sudden
0:42:31 there’s this knee in the curve.
0:42:33 And then all the indicators of human welfare,
0:42:36 not uniformly across the planet,
0:42:38 but in societies that were making progress.
0:42:41 Societies that were making progress first,
0:42:42 all of a sudden, all those indicators
0:42:43 of human welfare went up into the right, right?
0:42:45 And then all corresponded, by the way,
0:42:46 to economic growth.
0:42:48 But it was also right, it was the enlightenment,
0:42:49 it was the rise of democracy,
0:42:50 it was the rise of markets,
0:42:53 it was the rise of rationality, the scientific method,
0:42:57 by the way, human rights, free speech, free thought, right?
0:42:58 And they all kind of catalyzed, right,
0:43:01 around 300 years ago and they’ve been making their way
0:43:04 to the world in sort of increasing concentric circles
0:43:05 kind of ever since.
0:43:08 And so we have, I would argue like we have the answers,
0:43:11 like we actually don’t need new discoveries
0:43:12 to have the future be much better,
0:43:13 we actually know how to do it,
0:43:16 is to apply basically those systems.
0:43:20 And basically, contrary to the sort of constant temptation
0:43:23 from all kinds of people to try to compromise
0:43:25 on these things or subvert these things,
0:43:26 basically double down on these systems
0:43:27 that we know work, right?
0:43:28 So double down on economic growth,
0:43:30 double down on human rights,
0:43:33 double down on markets, on capitalism,
0:43:35 double down on the scientific method.
0:43:37 Fixed science, like we got as far as we did with science
0:43:40 actually being pretty seriously screwed up right now
0:43:43 with the replication crisis, like so we should fix that.
0:43:44 And then science will all of a sudden
0:43:45 start to work much better.
0:43:49 Technology, right, use of technological tools.
0:43:52 So we literally have the systems, like we know how to do this,
0:43:54 we know how to make the planet much better in every respect.
0:43:57 And so what we just need to do is keep doing that.
0:44:00 And then what I try to do when I read the news
0:44:02 is notwithstanding everything that’s going on
0:44:03 is basically try to look through whatever’s happened
0:44:05 in the moment, try to look underneath and kind of say,
0:44:09 okay, are those fundamental systems actually still working?
0:44:11 Like is the world getting more democratic or less, right?
0:44:13 Is free speech spreading or receding, right?
0:44:16 Are markets expanding or falling, right?
0:44:17 Are more and more people able to participate
0:44:19 in a modern market economy or not?
0:44:20 And, you know, those indicators generally
0:44:23 are all still up and to the right.
0:44:25 – So let’s go out and make the world better.
0:44:26 Thank you.
0:44:27 – Yeah, good, good.
0:44:28 Thank you everybody.
0:44:31 (audience applauding)