← Tom Bilyeu's Impact Theory
Tom Bilyeu's Impact Theory
How AI Will Transform Medicine, Entertainment, and Your Everyday Life | Marc Andreessen - PT 1
01:03:05
Loading summary
Transcript87 lines
- [00:01]
Host
Nine out of the 10 largest banks get it. They get that. VantageScore, the modern credit score, is the leader in predictive power, improving mortgage default predictions and saving lenders billions. Better predictions, better for your business. With VantageScore, there's nothing like the American Express Platinum card.
- [00:18]
Marc Andreessen
Find out your welcome offer after you apply, which could be as high as 175,000 points.
- [00:24]
Host
Learn more and find out your offer@americanexpress.com
- [00:28]
Marc Andreessen
explorer Platinum terms apply. I think the AI censorship wars are going to be a thousand times more intense and a thousand times more important.
- [00:35]
Host
My guest today is someone who doesn't just keep up with innovation, he creates it. The incredible Marc Andreessen. Trust me, when someone like Marc, who spent his entire career betting on the future, says this is the next major disruption, you need to listen.
- [00:50]
Marc Andreessen
From a political standpoint, we should hope that we have rapid technology progress, because if we have rapid technology progress, we'll have rapid economic growth. Do people care? And are people going to be willing to stand up for this? And I think that's what's required.
- [01:01]
Host
It's going to displace a lot of jobs. Some of those people will redistribute themselves by acquiring new skills. Other people will not. This isn't something to think about tomorrow. You've got to be prepared today. So let's dive right in. I bring you Marc Hendriessen. Marc Andreessen, welcome to the podcast.
- [01:21]
Marc Andreessen
Awesome. Thank you for having me.
- [01:23]
Host
My pleasure. Now, you've had insane amount of success betting on where industries are going. So let me ask you, what is the most radical disruption that you see coming in the near future with AI?
- [01:36]
Marc Andreessen
You know, I just say, like, we're convinced AI is one of those sort of moments of fundamental change. And, you know, in our, in the tech industry, you know, these come along every couple of decades, but they're not frequent. And, you know, this one is up there with the microprocessor and the computer and the Internet for sure, and maybe bigger. And so for, for us in the tech industry, this is a. This is a, I think, a very, very, very profound, powerful moment. And of course, you're already seeing, you know, a lot of, A lot of the, A lot of the effects that already playing out. But, you know, this technology, this technology is going to change a lot of things, and it's going to be, I think, very, very exciting.
- [02:16]
Host
And so for people that don't know you have a fundamentally optimistic view of AI, of technology in general, do you have, like, from an investment strategy, do you guys have A thesis on what industry you think is going to be most advantaged by AI that you're trying to get into.
- [02:34]
Marc Andreessen
Yeah, there are many. So we're involved in many. I would say there's some obvious slam dunk ones. And so I would say healthcare is a slam dunk one. I actually just happened to have lunch with Demis Hassabas who just won the Nobel Prize in chemistry. Chemistry for his work on protein folding.
- [02:50]
Host
Not a bad lunch date.
- [02:51]
Marc Andreessen
Yeah, exactly. And he was knighted this year also. So he's also a sirdemus. But he and his colleagues basically have this transformative approach that they believe is going to lead to dramatic breakthroughs in the development of medicine in the years ahead. Powered by AI. So healthcare is an obvious one. Entertainment is one that I think it's going to be extremely exciting what happens from here. And again, that's already starting to play out and you know, you're already seeing like just sort of incredible creativity being applied to that. And so, you know, maybe you could kind of maybe bookend it by saying those because it's kind of the most serious one and the most fun one. But then look, there's, there's, there's lots, lots of other stuff. Probably the single biggest question I'm asking right now is robotics. You know, there's been the promise of, you know, kind of robotics, you know, kind of saturating our society and you know, everybody having, you know, robot robots in the home and you know, everybody having, you know, robots to do, you know, to do everything, manual labor and you know, wash the dishes and pack the suitcase and clean the toilet and you know, can, you know, conceivably everything, you know, the manual labor, you know, kind of free people from manual labor. And that's, you know, been a promise, you know, going back, you know, in science fiction it's been a promise for you know like 120 years. And you know, until recently we were no closer than we were maybe back then. But you know, you're starting to see very dramatic I think breakthroughs and I think you know, sort of like you had sort of drones that now work like autonomous drones are like now a standard thing. Self flying, self piloting drones. You now have self driving cars that are now a thing and now work really well. And I think it may be, you know, humanoid robots and all kinds of other forms of robots. We have, we have two of, we have two Chinese robot dogs at home. What? Yeah, yeah, yeah, yeah, yeah. We have our.
- [04:33]
Host
You actually have them at your house?
- [04:35]
Marc Andreessen
Yeah, yeah, yeah. So there's a. Yeah, so everybody's probably seen all the demos. Remember, there's this company, Boston Dynamics, that has all these, they always have these great demos. You see these videos of these robot dogs running around. But you know, they cost like fifty thousand, a hundred thousand dollars. And then that company never really brought them to market and so it never worked outside of it as a demo. But there are now Chinese companies that have these things down to $1,500. Yeah. And they like, they're great then. They're great. They run around. They actually, they run actually quite quickly. They can outrun you. They do flips. They stand on their hind legs, they climb stairs. They can. There's a version of it that has wheels that they can go like 30 miles an hour. Then that one can also climb stairs. It locks the wheels and it's perfectly fine climbing stairs. So, you know, those are really starting to work. And then, you know, humanoids are coming fast. And you know, Elon just had his demo day for the, you know, Tesla Optimus robots.
- [05:25]
Host
Unreal.
- [05:27]
Marc Andreessen
Yeah. And so those are, those are starting to work. You know, it's not quite there yet. Like those were still teleoperated. There's still people in the background with VR headsets that are kind of steering those and guiding those and helping those. But that's also how you train these robots is you kind of have, they kind of watch what people do and then you train. So I think we might be like actually reasonably close in robotics, which would actually have a, you know, would have a very big impact. And so, yeah, maybe you could call out those three categories as obvious ones to, to, to focus on what kind
- [05:51]
Host
of timeline do you have for robotics? When are we going to start having that first round of people buying them and having them in their home? I know Elon's pegged it at 20 to 30 grand when.
- [06:01]
Marc Andreessen
Yeah, so the, the big breakthrough. So, so self piloting drones were a very big breakthrough. And you know, the dominant ones on those in the global market are this company dji, which is this big, you know, company in China, you know, but those now work really well. And then there's American companies. We have a, we have American companies that have, you know, I think even better technology that aren't quite the same size yet, but are really good. And so, and that's a big deal. Like, so you can, you can have, you know, we have drones now that can like fly between tree branches. They can fly, you know, indoors. You know, they can fly, you know, completely autonomously through like by the way underground tunnels and so those work really well. And then like I said like self driving cars, you know, the Waymo cars now are great and you know, people who use them have fantastic experiences. And then the Tesla self driving capability is getting really good. And, and so like, so I go through those to say those are both robots, you know, flying robot, driving robot and so walking robot all of a sudden is not so crazy. Exact timing, I don't know, you know, I, you know, swag five years but you know, could be two, could be eight, I don't know, optimistically three or four. The promise. There's many, many possible form factors for these things, right. Designs. The theory of humanoid robots, which I believe is the great thing about humanoid robots is there's just, there's so much of the physical world that assumes that there's a person present, right? So person standing in an assembly line, person driving a car, person driving a tractor, person, you know, picking, you know, you know, picking, picking, you know, vegetables in a field. There's just all these systems, you know, that we assume there's a person. And so if you build a robot in the shape of a person, in theory it can just kind of, you know, it can kind of, you know, fill in and do all that work. And so that, you know, that, that should be a very big market. And obviously people, you know, should be very comfortable with that. You know, they'll, they'll, you know, they'll, they'll dovetail, you know, really well into kind of normal society. But I also think there'll be a lot of other, you know, you can, you can package these things up however you want. And so there will be lots of other, you know, kinds of, you know, there already are obviously lots of robots in the world, but there will be, you know, more and more of different kinds of kinds.
- [08:07]
Host
And what are the hard parts? What are the hurdles they still have to overcome that's going to cause it to be three, four, possibly eight years from now?
- [08:15]
Marc Andreessen
Yeah, so there's basically, I would say three big categories. So there's the physical sort of controls, the actual physical kind of body and its ability to kind of control itself. And that's where if you look at Elon's demonstration the other night, you can kind of see how fast that stuff's moving because if you watch his progression of the other companies doing it, they're getting much better. And so that's just moving right along. Then there's battery power is probably still a fundamental limit because it's a question of how Long can you actually power one of these things before it has to recharge or do a battery swap? And that's still a bit of an issue. And it's hard to make progress on batteries, but a lot of people are working on it. And then software is the big challenge, I think, and where we would get more involved and know. So this sort of. This, all the software. And so think about it like, these robots have sensors. They've got visual sensors, They've actually got, like the robot dogs have what's called lidar, which is sort of the light version of radar, which is the same thing that's in the Waymo cars. And so, you know, they've got sensors, they can kind of, you know, gather. They've got sound, you know, they can gather input, you know, from kind of all around them. Actually, they can gather input from their environment better than human can because they can see 360 degrees and, you know, they can do depth sensing and so forth in ways that we can't. So they get all the raw data, but then it's a process. You have to actually process that data. You have to form it into a model of the world. The robot has to have a plan for what it does, right? And then it has to understand the consequences of the plan, right? And so I'm setting the coffee down on the table. I can't set it down on somebody's hand, so I have to set it down near the hand, but not on the hand. I have to keep it level because if I tip it, I'm going to scald somebody. And then, by the way, while I'm the robot, while I'm setting the coffee down, the person has moved, right? And so I have to adapt to it, right? Or, you know, same thing, walking through a crowd. Like, I can't, you know, you can't have robots running into people. And so you have to have to,
- [10:07]
Host
you know, how they're approaching, how they're approaching that problem. So if I think about when I saw the robots interacting with the people at the party, is there an underlying goal for the robot to be likable? And is it like, hey, get to know people, try to charm them. What. What is the plan that they're giving to the robot that it's moving towards?
- [10:27]
Marc Andreessen
Yeah. So I mean, in general, if you're a company, in general, you want basically completely benign, right? So if you're a company, you want. Because if it's actually, it lines up nicely with the profit incentive. You know, you want friendly, approachable, you know, Think your products that make people happy, products that make people comfortable, you know, products that aren't threatening or intimidating and aren't, you know, aren't hurting people. And so you put a really, really big focus on fitting into the environment. You put a really big focus on avoiding anything that would ever, you know, harm a human being. You know, you put a very big focus on, you know, the robot should, you know, happily, you know, you know, just, you know, should happily, you know, whatever, step into traffic or whatever if it's going to save somebody's life. And so, you know, you want that and then, yeah, I think, you know, generally you want it to be, you know, sort of approachable, safe, harmless, you know, are kind of terms that get used a lot, you know, you know, friendly. Now, look, this is the other thing is there used to be this, like really hard challenge, which is, how are you going to control these things? How are you going to talk to them? Are they going to. You know, if you watch Star wars, they communicate in beeps and boops. You know, if you watch Star Trek and you're watching Commander Data, he's talking in English. Up until two years ago, we thought it would have to be beeps and boops, but now we have large language models and we have these voice AI interfaces. Like OpenAI just released their advanced voice mode. And it's a full. It's like talking to the starship computer on the Starship Enterprise, or just talking to a person. And so all of a sudden you can give these robots voices. They can talk, they can listen, they can explain quantum physics to you. They can sing you a lullaby, they can forecast the presidential election. They can now do what you want. And so that's the other part of it is that you're going to really be able to talk and interact with them. The first one I saw, the Boston Dynamics guys did this hysterical demo where they wired up one of these early language models a couple years ago to their robot dog. And they gave it a super plummy English butler voice. So it's like this mechanical robot dog stomping around, but it's talking to you. It's like you're Bruce Wayne and it's Alfred or something. It's the robot dog. What do you see? And it does the very plummy exit. Oh, I see a lovely pile of rocks. And so, yeah, you're going to, by the way, there's going to be enormous creativity. There's this startup we're not involved in, but I like the guys a lot called Curio in Redwood City that basically has a plushie. So they have a stuffed animal and it's basically designed for little kids and it's a voice UI and it's backended by a large language model and it doesn't move, it's just a plushie with a voice box. But it will happily sit and tell kids jokes and teach them all about, you know, whatever they want to learn about and talk to them about whatever's on their mind. And they have it, you know, really elegantly wired up where the parent, the parent can both control how the toy actually like what it's willing to talk about. So you can, as a parent, you can like define, you know, the topics that are like go zones versus no go zones. So you could kind of say, you know, let it talk to the kid about, you know, science but not politics, for example. And then you get, as a parent, you get a real time transcript of the interaction. So like your kid's up in the bedroom talking to the, talking to the thing and you actually get to see, see, see the conversation, right? And so, and it's funny when you, when you watch this with like little kids, they just think this is like the most natural, normal thing in the world, right? I, I've talked in the past. I, I have a nine year old and I brought home when ChatGPT first shipped, you know, two years ago, I guess he was seven. And so I, he has a laptop that he does, is, is some of his, his school stuff on. And so I set up ChatGPT on his laptop and I sat him down. I was so proud of myself because I'm like, I'm like, I don't know, it's like I'm, I'm, you know, I'm coming down from the mountain to deliver like the gift of fire to my child. Like, I'm giving him like the super technology that's going to be with him his whole life that's going to answer any question, all this work. And this is like the most amazing gift of technology I could give him. And I showed him ChatGPT and I said, you type in any question you want and then it answers the question. And he looked at me and he said so. And I was like, what do you mean? This is like the breakthrough. This is the thing. This is like the thing for 80 years we've all been working on and it finally works. And he's like, what else would you use a computer for?
- [14:20]
Host
So funny.
- [14:21]
Marc Andreessen
Obviously it answers your questions, right? And so I think kids are going to, kids are I mean, it's already happening. Kids are going to pick this up incredibly fast. It's going to be super normal. Anyway, so long answer to your question. But we have a chance to design, you know, we can design technology to be as friendly and helpful and accommodating and supportive as we can possibly imagine. And I think that the commercial products will all get built that way for sure.
- [14:43]
Host
Yeah. To me, that's where the biggest disruption is going to be. When I think about AI, I think I'm as optimistic as you in terms of the things that it will do for us. It's intellect. You're going to be able to throw, you know, God knows how many new PhD level people and maybe one day even more at all these incredible problems. All right, that's going to be utterly fantastic. But then I think about your dog becomes a robot dog becomes furry and fluffy and wonderful, but it also talks to your kids and helps raise them and you have this lens into it. And then all of a sudden it's. Well, it's not just the dog, it's. I've got an AI girlfriend. She's not really a girlfriend. Not like that. Well, but then I, you know, I've been talking to her for three years, and now robot body comes online and I want to put that AI into the robot body. And all of a sudden I. I think that there's going to be a pretty fascinating. To try to keep it positive here, a fascinating schism that'll happen in society. So five years ago, I wrote a comic book about this, about what I think is going to happen. And I think there's going to be a bifurcation in society. And I, I really think this is actually going to happen. How big and how dramatic that that remains to be seen. But I think you're going to get a subset of society that says, nope, not doing this. It's like the opening line in Dune that thou shall not make a mind, an artificial mind mirroring human intelligence or whatever the exact line is. And I think people will eschew AI. They will eschew neuralink and things like that, and there'll be sort of this new puritanical vein of humanity. And then you're going to get other people like me that embrace the technology. I may not be an early adopter of neuralink, but if it truly gets safe and it allows me to upgrade my abilities, man, I will do that in a heartbeat. And so then it becomes a question of how much friction will there be between those two sides. But those seem inevitable. Do you think I'm crazy about that, or do you see that same inevitability? And if so, how does it play out? Hold tight. We're going to take a quick break.
- [16:47]
Grainger Advertiser
If you work in university maintenance, Grainger considers you an MVP because your playbook ensures your arena is always ready for tip off. And Grainger is your trusted partner, offering the products you need, all in one place, from H VAC and plumbing supplies to lighting and more, and all delivered with plenty of time left on the clock. So your team always gets the win. Call 1-800-GRAINGER visit grainger.com or just stop by Grainger for the ones who get it done. If you work in university maintenance, Grainger considers you an MVP because your playbook ensures your arena is always ready for tip off. And Grainger is your trusted partner, offering the products you need, all in one place, from H VAC and plumbing supplies to lighting and more, and all delivered with plenty of time left on the clock, so your team always gets the win. Call 1-800-GRAINGER visit grainger.com or just stop by Grainger for the ones who get it done.
- [17:49]
Host
All right, let's pick up where we left off.
- [17:51]
Marc Andreessen
I mean, I think it's certainly a plausible scenario. I think it's certainly logical. It certainly can play out that way. I guess my model of human behavior is different. So I'm skeptical that that is what will happen. And I would just start by saying that there is a schism of like that in our society today. And they are the Amish. Yeah, and I actually grew up, you know, they were Amish near where I grew up. And, you know, so the good news with the Amish is they have a defined quality of life. You know, a whole value system sort of, you know, involves, you know, rejecting technology for some, by the way, for some very deeply thought through reasons. And, you know, they're, you know, by all accounts, you know, in many cases very happy. And by the way, by the way, they're also very fertile. You know, so they're, you know, they're having lots of kids. And so, you know, there's actually, I think, quite a bit to admire about what they do. You know, look, having said that, I would just say two things. One is they're a very, very, very, very, very small percent of the population. And so there's not a lot of people who volunteer to become Amish. And then the other thing that happens, if you track them in detail, what actually happens is they don't reject, just adopted on a lag. Right. And Basically, the lag is about 30 years. And there have been a bunch of articles this, over the last decade where, for example, they're now adopting PCs, a personal computer. Really? Yeah, yeah. Well, because it's.
- [19:07]
Host
So I thought they were still without electricity.
- [19:09]
Marc Andreessen
No, no, no, no, they've got electricity. I mean, you know, they try to control it, but they definitely. This is a great example. They definitely have it. Right. And then they now have landline telephones. So there's a point where you just, you know, things just get to be practically. So, you know, the PC. So the PC thing, apparently the articles that I've read, basically what it is is the personal computer. Personal computer. Like, you know, they run these small businesses. They'll. They'll have like a, you know, they'll do like hand car furniture, for example. That's like, you know, these amazing things. Well, it's just a lot easier to run a furniture store if you've got a personal computer to do the ledger and the inventory on it. Right. And it's just. And at a certain point, they figure out a theory under which that's okay, they still don't connect it to the Internet, you know, but they. That they do that, you know, they have the personal computer, by the way, that, you know, and then you just kind of say, inevitably the next step is they're going to want to sell their furniture online. And so it's just a matter of time until they figure a way to bring in an Internet connection. Right. And so one of the really, really fascinating things about AI is it went from being something that was sort of speculative and weird three years ago to something that is now actually quite common, already in use. And this is quite a profound and powerful thing that I think we'll probably talk a lot about today, which is, number one, it's AI is already in wide use. And so the number of users on systems like ChatGPT and Midjourney and whatever are already in the hundreds of millions and are growing very fast. And lots and lots of people are using these things and they use them in their everyday life. They use them for work. They may or may not admit to their boss they're using them for work, but they're definitely using them for work. Students are using them in school, if you've got teenage kids, any classroom in America now is grappling with this question of is the kid bringing an essay that ChatGPT wrote? But they're helping with homework and they're doing all kinds of stuff. And the usage numbers on these services reflect Already broad based adoption. And then there's a really powerful thing underneath that that's really important. Which is the most powerful AI systems in the world are the ones that you get on the Internet for free or maximum 20 bucks a month. And very specifically, I have the capability, if I want to, I could go spend a million dollars to just have like the best AI. I could go spend a million dollars a year. If I go spend a million dollars a year today, I do not get a better AI than you get when you sign up for ChatGPT. It's literally not available. I can't do it. The best AI in the world is the thing it's on ChatGPT or by the way, Google Gemini or Microsoft Bing or Anthropic, Claude or xx, Grok, the XAI one or Mistral, which is one of our companies, or Llama for Meta. There's seven of these now that are available either for free or at most for 20 bucks a month. And they're the best in the world. It's actually quite striking, shocking, which is a lot of people have the mental model of, oh, well, the best technology must be basically hoarded by a few people who are then going to lord it over the rest of us and are going to make all the money on it. Right. It's kind of the, you know, the kind of, you know, kind of the fear on these things. The reality is like this technology is democratizing faster than the computer did, faster than the Internet did. It's available to everybody right out of the chute, by the way it's getting built. You know, Apple's building it into the iPhone. It's just, you know, now it's just Apple intelligence is going to be a standard feature of the iPhone. And so this technology basically has gone from not present in our society to like almost universal in one step. And I just, you know, it may be that people choose to voluntarily give it up, but I, in my life, I have not yet seen people who sort of voluntarily renounce something that they get used to. So yeah, it would be a first if it happened.
- [22:24]
Host
All right, I hear that. And you're the right person for me to have this conversation. I love when dogs bark the loudest because they're on a leash. So you're going to be my leash. I'm going to paint a scenario knowing that you're going to pull me back from the brink because I'm fundamentally a techno optimist and I'm definitely somebody that will embrace this technology as fast as humanly possible. We're deploying it here in my company as rapidly as we can. I will literally, if it's proven safe, get neuralink the whole nine. So here's what I think plays out. This is as close to the sort of realistic mess that I think we'll go through. The long arc of history bends towards justice, but history does not care about any single generation. And I think that the thing we will all have to get very politically comfortable with is the fact that, yes, AI is going to displace jobs wildly as we move towards something absolutely wonderful and spectacular, but it's going to displace a lot of jobs. Some of those people will redistribute themselves by acquiring new skills. Other people will not, and it won't be a great time for them. And their families will rally around them as the material wealth is unlocked, as spending power becomes more abundant, all of that, the younger people that are more intellectually nimble will take advantage of that to care for people. But there's going to be this conflict on the left and the right as to, hey, shouldn't we just give these people UBI or whatever to take care of the people that are going to struggle? Because they are going to struggle. And if people don't have a mental defense, if they don't have a narrative that they can understand about how we weather that storm, I think they'll make very bizarre economic choices. As you were talking, you're talking about deflation. And people ought to wonder how on earth, given all the technological advances we've had over the last 300 years, how is inflation still going up? This seems crazy. And the reason that inflation goes up despite the massive deflation that technology brings is that the government gobbles it up by printing money. And, oh boy, do I have a personal bone to pick. I have no idea your take on the economy and how it intersects. So I'll plant my flag and let you react. I think that the. You need only look at the M2 money supply chart to see. I mean, it's just, just absolutely outrageous how much more money has been poured into the system completely artificially, just generated out of thin air. And that, that is the inflation. When we say inflation, that's what we're talking about, the inflation of the money supply. In doing that, the government doesn't have to get your vote on something they will. I refer to, and I do not want to put words in your mouth, but I refer to that as the government steals from you and then they force you to play the stock market as One stand in for investments in order to beat inflation caused by them printing money and stealing from you. And I think that's deranging. And I think that the government has a moral obligation to give people a non inflatable currency in which people can at least park their wealth so that the average person who does not want to play the stock market can just save. Like a guy that is a janitor and he's just trying to get by and, and take care of his family should be able to sock away money and not have its value eroded over time through very conscious and poor, in my opinion, policies. Curious to get your take on that.
- [25:47]
Marc Andreessen
If I tell you that in any given time you could have more or less technology change and then that change would show up in economic statistics the way that economists measure it, as with productivity growth, which is a thing they measure, it's actually a number. And so you know, if you have, if a society has, you know, 1% productivity growth, that's super low. If they have 4% productivity growth per year, that's like super high. Let's call that the super rain. And if you could ever get to 8 or 10% productivity growth, you'd have cornucopia technological utopia. It'd be amazing. Everything would get like super cheap and abundant super fast. But like that modern societies go somewhere between 1 and 4%. Would you say that we live in a time today in which productivity is growth and therefore technological change is running high or low?
- [26:29]
Host
I think we are about to unleash a ton of that productivity. But right now I think that the government is siphoning off so much of that productivity that you get this schism between the young and the old. So the old, I think are doing very well and the young are getting absolutely clobbered and so they don't feel it. But if AI does what we think it's going to do, then yes, I think that we will finally be able to unlock a lot of that.
- [26:55]
Marc Andreessen
But just take the distributional part of it out just because we'll come back to that. But just take the distributional part out. But just talk just about just the rate of technology change. Do we live right now in a time of great technology change or low technology change?
- [27:08]
Host
The only great technology change is in AI so low.
- [27:12]
Marc Andreessen
Okay, and then you'll probably get the next answer right, which is did we have faster technology change between 1930 and 1970 than we do today or slower?
- [27:22]
Host
Much faster.
- [27:22]
Marc Andreessen
Much faster. Yeah, so those are the correct answers. And so the metric on what's happened, and this is actually quite important, is that productivity growth and therefore technological change in the economy was much faster in the decades that preceded the 1970s. Actually, by the way, the turning point was the year I was born. It was 1971.
- [27:39]
Host
In 1970, WTF happened?
- [27:41]
Marc Andreessen
It was you, Mark. Yeah, so there's a website called WTF happened in1971.com and it's like literally hundreds of charts of basically this discontinuous change on all kinds of economic and social markers that kind of kicked in the year I was born. I do believe it is entirely my fault, I will confess to that. But yeah, one of the things that happened was right around that time productivity growth downshifted. It was running at like 2, 3, 4% and then it's sort of been 1 to 2% ever since. And it ebbs and flows a little bit with the economic cycle, but like it's been quite low for the last 60 years. Part of it dovetails to the political thing you were saying. There's a lot of questions as to why it's been so low. There's actually economists talk about something called the productivity paradox because it was really weird because the computer emerged in the 1970s. And so all the economists in the 1970s said the computer is going to lead to. It's going to lead to cornucopia, it's going to lead to enormous productivity growth. Of course it is. You got Moore's Law and it's just like it's all this software and all this inventory just in time, manufacturing. And you're going to have, by the way, robots. And so you're going to have this for sure. You're going to have a massive takeoff in productivity growth. And actually what happened was productivity growth growth actually downshifted. And so all of our expectations for how society works are actually geared towards low productivity growth and low economic growth. From a historical standpoint, the importance of that is really key to the next thing that you said, which is the psychological effect of being in a low growth environment is zero sum politics. Logically, right? Because if we're in a high growth environment, if technology productivity growth is running at 4% or God willing, someday more, and if economic growth is running at 4% or more, the economy will be doing so well, it will be spewing money in all directions. Everything will be going crazy. Everything will be. Every business will be flush, every consumer will feel fantastic. Jobs are being created all over the place. Everybody's kids for sure are going to live better lives than their parents did. It's going to be great. By the way, the 1990s were that there was this kind of five year stretch in the 1990s where economic growth really took off. And you probably remember it was fantastic. Everybody felt fucking awesome. And so this is one of the kind of weird, this is why a lot of the fears around the impact of technology I think are really misguided when it comes to all these economic and political topics, which is from a political standpoint, we should hope that we have rapid technology progress, because if we have rapid technology progress, we'll have rapid economic growth. If we have rapid economic growth, we'll have positive some politics, right? For me to be in a high growth environment, for me to be better off, I can go be better off. I can go exercise my skills and talents and get new jobs and switch jobs and switch careers and do all kinds of things. And I have a path and a future for myself and my children that does not require taking away from other people. In a low growth environment. All of the economics and all the politics go zero sum because the only way for me to do better is I have to take away from you or to your point, the government. Exactly, I completely agree with you. Or what happens is the government just inflates and they inflate because they want to basically buy votes. They want to basically spend on programs and they want to buy votes. And so this is sort of what I would say, which is like if you want zero sum, smash mouth destructive politics with the government playing a bigger and bigger role, you want low technological development, you want a slow pace of technological development. If you want, you want positive some politics where people are thrilled and excited about the future and about their own opportunity and they don't have to feel like they have to take away from somebody else and they don't need handouts from the government because they're doing so well. You want rapid productivity growth. Right? And so you said, I'm saying like it's the opposite of the fear that everybody thinks that they have. I have many other thoughts on your question, but yeah, let me pause there and see which part you wanted to get to. Oh, inflation. So inflation, yeah. So look, but I would just say two things on inflation. It's actually pretty interesting. So there's an overall concept of inflation which as you said is growth of the money supply. But the way that that plays out in the economy is. And they actually analyze it this way, it's basically the way they think about it is it's the basket of overall prices of everything in the economy. And the government agency that calculates the rate of inflation uses a basket of equivalent products over time to try to get a sense of what's actually happening with prices. And so there's both the money supply aspect of inflation and the government printing press and all that. And that's totally true. But what's actually happened inside that is actually because of differences in technology and regulation. You actually have really, I think, historically unprecedented difference in how different industries are actually inflating or deflating. And there's a chart that we can maybe post for your listeners that basically shows three really big important sectors of the economy, which are healthcare, education and housing, where the, the prices are skyrocketing, which, and by the way, everybody feels this, right? This is just like, okay, you want to go buy a starter home, or you want to get good healthcare, or you want to get your kid in a good school, the prices are going crazy. I mean, you see this in housing prices, of course. Another version of this is the higher education. A four year college degree at a private university now costs $400,000 and is on its way to a million dollars. Right?
- [32:41]
Host
That's crazy.
- [32:43]
Marc Andreessen
Completely crazy. Completely crazy. So the price of higher ed is just skyrocketing. The price of higher education, bachelor's degrees, master's degrees, rising far faster than the rate of inflation. Same thing is health care costs are rising faster than the rate of inflation, and housing prices are rising faster than the rate of inflation. But then you have all these other sectors, and these are sectors like video games, entertainment, consumer electronics, by the way, food, cars, which is good retail, consumer products. Generally those prices are crashing. And so the things that you can buy today versus 20, 30, 40 years ago for the same dollar in those categories, I just take obvious examples. Music, obviously, music. To buy music, 30 years ago, you had to go spend $15 to buy a CD and get 10 songs out of which you maybe wanted two of the songs. Today, $10 buys you Spotify for a month and you have 10 million songs on demand and you can listen to it 24, 7 and it's fantastic. And so the price of music has crashed. And so the price of housing, education and health care has skyrocketed. The prices of everything else is crashing. What explains that? Well, the prices for everything, that's crashing. Number one, they have rapid technological changes which is driving down prices because of productivity growth. And they're not regulated. Nobody in the government is price fixing music. Whereas housing, education and healthcare are incredibly highly regulated and centrally controlled by the government. And they have fixed supply dictated by the government, and they have very slow rate of technological adoption. It's almost impossible to get new technology into the healthcare system, into the education system, or into health housing. Like robots are not building houses. Like it's not happening right? Like it's just not happening.
- [34:24]
Host
More to come. We'll be back in a bit with
- [34:27]
Vrbal Advertiser
Vrbal's last minute deals. You can save over $50 on your spring getaway. So whether it's a mountain escape with friends, a family week at the beach or sightseeing in a new city, there's still time to get great discounts. Book your next day now. Average savings $72 select homes only. When you manage procurement for multiple facilities, every order matters. But when it's for a hospital system, they matter even more. Grainger gets it and knows there's no time for managing multiple suppliers and no room for shipping delays. That's why Grainger Granger offers millions of products in fast, dependable delivery so you can keep your facility stocked, safe and running smoothly. Call 1-800-GRAINGER Click grainger.com or just stop by Grainger for the ones who get it done.
- [35:28]
Host
We're back. Let's dive right in.
- [35:31]
Marc Andreessen
And so what we have actually in the economy is a diverse. I call these sort of the slow sectors versus the fast sectors. The sectors for which prices are skyrocketing because of slow technology change and too much government regulation. And the sectors where prices are crashing because of rapid technological advances and lack of government regulation. And when you chart these out, you could just extrapolate the lines. And so where this is happening is within a decade, if the current trends continue within a decade, a four year college degree is going to cost a million dollars. Right? And a flat screen TV that covers your entire wall is going to cost $100. And at some point you might want to ask the question like, isn't that backwards? Isn't what we all. This is where I get very emotional about this because, okay, define the American dream. The American dream. And by the way, for that you could probably substitute the dream in many other countries, but let's just say the American dream. The American dream. I want to buy a house for my family. I want to be able to send my kids to a great school and then I want my family to be able to get great healthcare. Those are the three high order bits. And those are the things where we have wired, our system is wired right now to drive the prices of those things to the moon, right? And then, good news, iPhones and cars and digital Music are plentiful, but they're not healthcare, education and housing. And this is the other thing that's driving inflation. Because then what happens is the fast sectors of the economy with prices are crashing, they're shrinking as a percentage of the economy because prices are falling so fast. And then because prices are growing so fast for healthcare, education and housing, they're becoming larger and larger parts of the economy. And so the economy writ large and people's pocketbooks and how you spend your money, it's being eaten by these sectors that have slow technology growth and therefore rapidly rising prices. By the way, once again, if you want to fix this problem, what's the way to fix this problem? You inject a lot more technology into those three sectors. You would want completely automated AI driven healthcare, you would want AI education, every kid having an AI tutorial teacher, and you would want robots building houses. If you wedged full modern technology into those three sectors, you could crash prices, which would also crash inflation and would cause everybody to be far better off. Once again, it's this thing where you think you don't want the technology to change. You actually very, very, very much want the technology to change. If we don't get the technology change, our politics for the next 30 years are going to be so crazily vicious because we're all going to be fighting over this shrinking piece. We're just going to hate how we have to live. So let me pause that.
- [38:05]
Host
Do you think the benefits of AI will be so overwhelming that there's just no way for politicians to hide the ball? Or will there be enough narrative and story and being able to leverage the resentment that exists right now to continue to forestall that, continue to grow government, keep it strong, keep it big.
- [38:26]
Marc Andreessen
Yes. Let me give you a micro answer and a macro answer. So the micro answer. So did you see the dock worker strike that just happened? Yeah. So the dock workers just went on strike and they demanded this huge raise. They demanded a huge raise. They demanded no more technology at the docks. They have this actually this dichotomy of an argument. They say our jobs are so backbreaking and arduous and physically harmful to our workers that we need to be appreciated a lot more. And we want you to completely ban the introduction of automation that would basically automate those jobs so that our workers don't have to do them. They kind of make both sides of this argument like at the same time, because they're completely contradictory. But that's not their responsibility to resolve it. But the dock workers go on strike they were literally asking for no more new technology at the docks to preserve the jobs. It turned out through that I discovered I just had never looked at that industry before. It turns out there are 25,000 dock workers in the US except that's not right. There's actually 50,000 dock workers in the US there's 25,000 dock workers actually work on the docks. And then there's 25,000 dock workers who don't work, who just sit at home and collect paychecks because of prior union agreements banning automation. What? Yes. Whoa. Yes. Because in previous bargaining rounds, they cut deals where if there were introduction of like, for example, machine cranes to unload containers from ships, that those jobs would not go away. And so those jobs have not gone away. There's nothing for them.
- [39:49]
Host
That's crazy. That is malpractice.
- [39:51]
Marc Andreessen
Well, so this is the thing. So this is the thing. Okay, so this is the classic thing on all these things. Is that good or bad? Well, it depends who you are. In political science, there's this concept of concentrated benefits and diffuse harms. And so for those 50,000 dock workers, this is great. For the rest of us, it just makes everything we buy more expensive because it makes working the docks more expensive because it's got all this deadweight loss on chips, which is a big part of the cost of all the food we buy is more expensive as a consequence of these kinds of arrangements. But you and I pay another 5 cents every time we go to the supermarket as a consequence of this, versus the 50,000 people who are organized in a union and are able to negotiate on their behalf concentrated benefits to the dock workers, diffuse harms to the rest of the economy. And every time you get a special interest group in the economy pleading for this kind of employment protection, that's what's happening. They're basically trying to create an employment cartel that benefits the people in the cartel at the expense of everybody else. Here's the macro version of that. 30% of the jobs in the United States today require some form of occupational licensing. You can't just get the job. You have to have some form of certification that you're qualified for the job. This has been pushed to extraordinary lengths in the United. In California, you need, I think it's, it's now it's like 900 plus hours of professional training to be a hairdresser, right? Yes. Correct.
- [41:18]
Host
You need what?
- [41:19]
Marc Andreessen
Yes. You cannot just like start cutting people's hair for money. No, no, no, no, no, no, no. That's illegal. You need to have a Whatever. Cosmetology certificate. If you get the cosmetology certificate, you have to go to hairdressing school to do that. By the way, you have to get admitted to hairdressing school. It has to be a certified hairdressing school. By the way, guess who controls how many hairdressing schools there can be? Is the, you know, the current. Oh, this is my favorite part. Let me give you my favorite example of this. So the university system. So, so federal student loans. There's. There's federal student loans for you to go to college. For you to go to college. You basically can't. Normal people, you can't afford to go to college if you can't get federal student loans. So you can't be a university or college or university in the US without having access to the federal student loan program. It's not possible. But to be a college or university that is able to give out federal student loans, they have to be accredited. Guess who accredits colleges? Colleges and universities. The existing colleges and universities.
- [42:11]
Host
Yeah, saw that one coming.
- [42:12]
Marc Andreessen
Guess how many new colleges and universities they're accredited. Like zero. Right. And so 30% of jobs in the country right now require some form of license or accreditation. By the way, this is all, you know, by the way, this is all doctors. And by the way, I think that's good. You probably want doctors to be accredited, but it's also, you know, nurses, nurse practitioners. It's, you know, and then it's. And then it's. It's not just lawyers, it's also paralegals. And then it's not, you know, it's. It's, you know, it's all kind of general contractors. You know, it's like. And then, and then on and on and on, including, depending on which state you're in, including, you know, hairdressers and many other jobs where you would not think this is required, by the way. Or another version of this is teacher. You know, to be a teacher in a lot of places in the U.S. now, you need an education degree. Right. Is there any evidence that teachers with an education degree are better teachers than teachers without an education degree? I don't think so. By the way, the education schools are completely bananas. Crazy. They're the most crazy of, like the academic departments at these crazy universities. Right. But again, it's a cartel structure. Of course, K through 12 education is not just a cartel. It's a government monopoly. Right. So you have to get actually hired into the, into the. Well, actually, this is the other great part. You have higher ed is like this. There's K12 is like this. And there's other branches of the federal workforce and state workforce that are like this, or actually police and police like this. You have quite a few people in the economy today who both have their government employees. They have civil service protections because they're government employees, which means in practice they can't be fired. But they're also members of what are called public sector unions. Right. So they both have to get hired by the government with whatever criteria they set and they have to get admitted into the public sector union. And they have the employment protections of both. Right. Of both the civil service and the public sector unions. Right. And this is why, by the way, you can't fire bad teachers, can't get fired. Right. Because you hit all these things. So just. So the point of that, the point of that is AI cannot not cause change that quickly in this system. AI cannot become a lawyer, it's not legally allowed to. It can't become a doctor, it can't replace the dock worker, it can't cut your hair, it can't build your house, it's not legally allowed to. Right. And so a very large, it goes actually to the Gulliver thing. A very large percentage of the economy as we experience, it literally cannot be automated. It's illegal to do so. That's so ridiculous. Yeah. And so I actually think what's going to happen is the economic impact of AI is actually going to be very muted compared to what people are fearing or hoping or fearing because it's literally not legal to do that.
- [44:41]
Host
It's crazy. So if everything that you just walked us through is true in terms of when you have high growth, everybody's feeling good. More technology equals more growth. AI is poised to bring that growth. But you have this trepidation and so people. Well, not, it's not just that, but you have trepidation around it. So the fact that the government tends towards this justify its existence, create a new regulatory body, slow things down, everything just grinds to a halt. For people that don't know the story of Gulliver's Travels, you have this guy that encounters these tiny Lilliputians and despite him being, you know, whatever, a thousand times bigger than they are, they. They just end up tying him down with all these tiny little strings. And that's a analogy that Elon certainly has used a lot. What do you think about his idea of going in and creating a, an efficiency program inside the government to try to free up some of these strings so that the economy can get going again.
- [45:36]
Marc Andreessen
Yeah, that's right. So I'll give you a couple books if people want to read about this. So one is Supreme Court Justice Neil Gorsuch just wrote a book. I think it's called, like, I forget the name. It's like overlawed or overload or something like that. But he basically lays out the data on the number of laws in the country. And by the way, this is another one of these WTF happened in 1971 things, which is starting in the 1970s, the number of laws and regulations in the US just took off like a rock. It basically what happens, the lawyers took over everything. And by the way, a big part of that is in politics. Basically, almost everybody now who's in elected office is a lawyer. Right. And so basically the lawyers just kind of swept in and took control of everything. And so if you. And also Senator Mike Lee has also done a lot of work on this. And you can just count the number of laws and then you can also count the number of regulations, which if anything is even worse because they're not even laws. They're just like a bureaucrat who's decided something. Right. And the number of regulations is just like skyrocketing. So he goes through it in the book. And then there's another book called Three Felonies a Day. And it goes through in detail that technically, odds are you, I, and every other American citizen are committing at least three felonies every day. There are so many.
- [46:40]
Host
And we just don't know it.
- [46:41]
Marc Andreessen
We don't know it. We don't know it. And the reason is because there are so many penalties, there are so many felonies on the book. On the books. And the felony laws are so sweeping in what they cover now. You know, most of those never get detected or prosecuted. But like, if somebody want, if prosecutors want to come at you, they can figure out ways to. This is what people with lots of experience in legal system always tell you. Like, if the feds want to get you, they're going to figure out a way to do it because you're almost certainly tripping something. And so, yeah, so I completely agree with Elon on the nature of the problem. Like, it's just, it just. Yeah. And again, this is, this is sort of this weird. It's this like, concentrated benefit, diffuse harm thing, which is like each law or regulation in isolation seems like a good idea. And each law or regulation has somebody advocating for it because they're going to benefit from it. And typically there's some level of Self interest, somebody's trying to get something for themselves and then they sort of have a cover story of consumer benefit or something. And then they get these things passed and they operate in Washington and they're in the State House and they get these things passed and each one of them on its own is not a big deal. But you run that process at scale over 60 years and that's when you end up with the Gulliver scenario, which is you're just drowning in lawsuit regulations. And again, I tie back to what I said before. Like that's why the prices of healthcare, education and housing have skyrocketed is because that's where like the laws and regulations in the economy are concentrated.
- [47:59]
Host
All right, so let's talk about then the next four years. So if Elon were to find himself in that position, do you think that we could meaningfully strip away red tape to the point that those, that, that scenario you painted, where those three things we care about so much, where the prices begin to crash? Or is that just unre realistic, full stop? Is it unrealistic in 4 years, how much can we do?
- [48:27]
Marc Andreessen
So it could be done for sure. There is actually a case of it actually happening in the world right now, which sitting here today looks very good, which is Argentina. And so Javier Milei was the new president of Argentina, has passed. I don't know the exact details, but I think his first big reform package, which was a real fight for him to pass, I think it was like, it fundamentally was like, I think it, it took regulations out of I think 800 different sectors of the Argentinian economy in one package. And they have a follow up package they're working on that's like another, it's like 2000 or something. So he's trying to do exactly what you just described. He's trying to just. Basically he's just like Milei's staunch libertarian, anti socialist, anti communist. He has this great line which he used the other day, which I love so much. So Margaret Thatcher had the famous line about socialism, which is, she said the thing about spending other people's money is eventually you run out. Milei has a better term, but he says anybody can be a prostitute with other people's asses.
- [49:28]
Host
That guy is a gangster.
- [49:29]
Marc Andreessen
He's hilarious, which is freaking amazing. Anyway, so he's trying to strip as much regulation out as possible. And the thesis of it is precisely this. It's like, okay, you strip out regulation, you remove government control, you liberate the people. You liberate the people to be able to exchange, go into voluntary trade and exchange, to be able to actually conduct business with each other without the government interfering with it all the time. Time. And then as a consequence, you get far higher rates of economic growth, far higher rates of prosperity. And so this is a big experiment. And of course, Argentina has been a case study for 100 years of doing this the wrong way. And he's now administering a form of shock therapy to basically see if he could do it the right way. And by the way, sitting here today, in very short order, inflation in Argentina has. They've had a persistent inflation problem for a very long time. He's completely nuked inflation. And economic growth has kicked in and job growth has kicked in. Now he is fighting like he has enemies, right? He is fighting like crazy, both in the political system and riots in the streets from people who are trying to stop this. And so anyway, so that goes to our situation, which is, yes, the theory is totally sound, right? Everything that Elon's describing should absolutely happen. This should absolutely be done, by the way, I think basically everybody knows this should be done. Again, concentrated benefits, diffused farms, even people who benefit from some aspect of this are suffering from it in every other area of their lives. Lives, right. And so this is what Mile always points out is the system in aggregate is making everybody poor. It is leading to all these bad, as you said. It's leading, for example, to intergenerational conflict that's just unnecessary and very destructive. And so it's just like, let's just stop this form of self harm. But to do that. The reason I say this, every single regulation has somebody behind it who doesn't want it to go away because it benefits somebody. It benefits the dock workers who are sitting at home. It benefits somebody. All the little cartels and monopolies and oligopolies and little conspiracies in the economy, they are in business because they're protected by the government. And when you strip these regulations away, you expose them to competition. And they really don't like that. And so there will be a backlash from the system, from all of the special interest groups in aggregate will rebel in great numbers. And then, look, the key fight ultimately is the civil service itself, you know, the actual government employees, right? And so, you know, for example, you know, how about a reform where, like, there's actual performance metrics for government employees and low performers.
- [51:50]
Host
Get fired, brother, please, if you want to get me in a cult, start a cult about that. I'm here for that. I'll do what we need. I'LL wear whatever crazy outfit. I am here for that one.
- [52:00]
Marc Andreessen
Yeah, yeah, exactly.
- [52:02]
Host
Let me ask, going back to Melee, are in. Are the layoffs causing any sort of economic downturn? Because one criticism I've heard of Elon Musk is, hey, if you come in and you do this and you slash it, not only is it cruel, but you're going to tank the economy. You're going to have so many people without a job.
- [52:16]
Marc Andreessen
Yeah, yeah. So that. So, yeah. So this, this happens. And by the way, this happened actually in, in the, in the late 70s, early 80s, there was actually a version of this which is inflation in the US actually got completely out of control. And you know, there was. Everything was kind of going sideways, but inflation went crazy. I think inflation spiked at like 15%. And then Paul Volcker, which was super destructive, right. Like really ruinously.
- [52:38]
Host
Yeah.
- [52:39]
Marc Andreessen
Bad. It destroys everything. It destroys savings, it destroys the ability for businesses to plan. It basically damages everything. And the way you crack the back of inflation is you raise interest rates and you deliberately cool the economy in order to bring down the demand for money. And then inflation falls. And so Paul Volcker, who was the chairman of the Federal Reserve, he's a famous guy. He's like the six foot eight giant guy with a cigar. And he was the head of the Federal Reserve and he lived in the undergraduate dorms at I think Georgetown and took the taxi to work.
- [53:08]
Host
Work.
- [53:08]
Marc Andreessen
So he was in contact with regular people every day, even though he was the head of the Federal Reserve in his three piece suit. And whenever he testified to Congress, if you see the old photos, he's just constantly. Just giant clouds of cigar smoke around him all the time. So one of these old school figures. And he raised interest rates in 1981, I think, to 20%.
- [53:29]
Vrbal Advertiser
Whoa.
- [53:29]
Marc Andreessen
Which basically crushed the economy. It basically crushed demand in the economy. It meant that nobody could borrow money, nobody could buy a house, nobody could start a business. It was very devastating in that moment. Moment. But he wrote a book about this and he said at no point when he would walk down the street and people would recognize him, this is in D.C. and he'd be walking down the street or he'd be in the cab. He said nobody was ever mad at him because what they said was inflation is so bad. We know that inflation is bad. We know that you have to do what you're doing at interest rates to do it. We know if you do it, you're going to fix the inflation problem and things are going to go back to being good again. And so we Support, you stick with it. And so he had the people on his side. And Milei has the same thing in Argentina right now. He has very high level of support from the population because they've seen the other experiment for too long. They've been through a society with too much regulation, too much corruption, and too much inflation for a long time. And they're just like, look, the people are behind them. You've seen in the polls and you see it in the voting. They're just like, all right, we're going to try Plan B. And so what you need is you need a politics of Plan B. You need a majority of the population to basically say, look, whatever the pros and cons of the old system were, they're not working, working. And we need fundamental change. And then obviously you need leadership that's going to, you know, that's going to be willing to implement that. But if the people are behind it, then you can actually do that. And so the fact that it's actually. The fact that it worked under Volker and the fact that it's working under Melee is very promising. Like, those are two great examples of how it can work. We don't yet have that, but we could.
- [54:58]
Host
Very, very interesting. When I start thinking about how we build back, we get the economy going, we take off the Gulliver strings. One of the things that I would want to see is, one of the things I think we need to see is a return to prizing freedom of speech. Because if we can't debate these ideas, if people can't get in there and mix it up and say, okay, I think this is the way. No, that's terrible. We should be doing it this way. But, you know, nothing being verboten like actually being able to discuss these ideas, that feels like a critical need. What's your take, especially coming off the heels of talking so much about AI, what's your take on censorship? Where are we culturally? And what's AI's role going to be in either breaking us free from censorship or using that to really tighten down?
- [55:46]
Marc Andreessen
Yep, yep. So I should start with I am classic Gen x. I am 100% pro free speech. Two of them.
- [55:53]
Host
This.
- [55:53]
Marc Andreessen
I am 100% pro free Speech. By the way. The first, you may know this, the First Amendment guarantees the government, at least in theory, is not supposed to censor us, although that's been happening a bit lately.
- [56:03]
Host
Just a smidge.
- [56:04]
Marc Andreessen
Just a smidge. But the government also, there's a case law around the First Amendment that actually defines illegal speech. And there are a bunch of forms of illegal speech. And it's things like child porn and its incitement to violence. It's terrorist recruitment. Right. And so there's actually like carve outs for, you know, that stuff. And so like, my philosophy is basically US Law is actually very good on this. And US Law isn't just US Law. It's also an, you know, this has been litigated culturally in the US as well as legally for 250 years. You know, going back to the Bill of Rights, you know, we and our predecessors in the US went through a long process to get to where the First Amendment is. I think it therefore represents more than just a law. I think it's also a statement of culture and a statement of values. And I've always. True, right. And I've always been an advocate that like the code for Internet freedom of speech should basically be that it should be the First Amendment with only limited carve outs for things that are truly dangerous, you know, truly destructive. Like I, you know, I don't want, you know, I don't want terrorist recruitment anymore than anybody else. But like, you know, should people be able to talk about their politics online without getting censored? 100%. Right. Full range of expression. 100%, of course, like, it's the American way. Of course. And so I'm 100% on that. You know, you, you know, probably as much as I do about the last decade, you know, which I've seen up close, which is, you know, generally things went very bad, bad. You know, the Internet companies, you know, ran into a variety of, you know, externally and self inflicted, you know, situations where there ended up being a pervasive censorship machine for a long time. You know, the most dramatic change of that is Twitter before and after Elon buying it. And we're, by the way, we're a proud member of the syndicate that bought it with Elon. And so, you know, I'm completely thrilled by.
- [57:37]
Host
Thank you for your service, by the way. To me, it's just so better. I cannot. Yeah, I just can't believe that that was controversial. It's crazy.
- [57:45]
Marc Andreessen
Yep. And as you know, it was a big change. Like, it was a. It was an absolutely dramatic change. We're also, by the way, the main investor, outside investor in Substack, which I think has also has done a spectacular job at navigating through this and is, you know, is basically has come out the other side of. And you know, they're a small company, so when pressure gets brought to bear on a small company, it can really have an impact. But the team there has, I think, done a fantastic job navigating to a real freedom of speech position. And as a consequence, Substack has now the full range of views on all kinds of topics in a really good way. So the good news is we have two case studies where this has gone really well. The other ones are more difficult. Here's what I would say is I think the Internet social media censorship wars were the preamble to the AI censorship wars. I think the AI censorship wars are going to be a thousand times more intense and a thousand times more important. Yes, 100%. And the reason for that is, you know, the Internet social media is important because it's what we all say to each other. But AI is going to be, I think, the software layer that controls everything. It's going to be the software layer that basically tells us everything. It's going to be the software layer that teaches our kids. It's going to be the software layer that we talk to every day. And as I think, you know, there's already AI censorship. Like, you know, these, you know, a lot of these LLMs are very slanted. And it's very easy, by the way. It's very easy to see because you can go on them today and you just ask them two questions about two opposing political candidates, and they give you completely different. One candidate, they're like, I'd be happy to tell you all about his positions. And the other candidate, they're like, oh, he's a hate figure. I won't talk about him. And it's like, wait a minute, right? Like, half the country's voting for one, half the country's voting for the other. Who are you as an AI company to basically censor like that? And so, look, the AI censorship, the AI censorship conflict is already underway. The information war around AI is already underway. By the way, the same people who were pushing so hard for social media censorship have now shifted their focus to AI censorship. By the way, a lot of the actual censors themselves who used to work at companies like Twitter now work for the AI companies. So there's been like a direct, you know, just app, you know, lessons learned, and now applying it at a larger scale. And so I think that, yeah, no, look, I think this is going to be a giant fight. I think it's just starting. I think it's, you know, maybe the most important. I think it's maybe the most important political fight in the next 30 years.
- [60:00]
Host
Tell me why.
- [60:01]
Marc Andreessen
Well, because everything is downstream. Everything is Downstream from being able to discuss and argue and be able to, you know, be able to communicate. And so if, if you can't have. If you cannot have open discussions about important topics, you can't get good answers.
- [60:17]
Host
Let me give you an angle on this. I'm. I am pretty sure we will agree about this. The, the thing about AI censorship that scares me isn't just the. That person is a bad person, and so I'm not gonna tell you about them. It is that you can control the entire world through framing. Just how you frame something and everything has a frame. And when you have humans with a desire to convert the or indoctrinate rather than seek truth, then now the only thing I can guarantee is, okay, the. The AI is responding to me from within a frame. They are using that to nudge my thinking in a direction, and it becomes a form of mind control. And, and if you've ever seen, Dear Listener, if you've ever seen an incredible debater, I promise you what you love about them is they can reject the frame and then put their own frame on it. And now they're arguing from a position of power. Most people can't do it. Most people don't even realize somebody just put them in a frame and they don't realize how constraining that frame is. And that's what really freaks me out, is everything else felt more like it was out in the open. Like, even when it was still Twitter and Twitter was being censored like crazy, everybody was like, bro, this is so obvious. Like, look, you post about this, poof, gone. I post about this, it's gonna explode. So when the Twitter files came out, I don't think anybody was like, wait, what? Everyone was like, yeah, that's exactly how it felt. This will be a game of frame and really does come down to it's hard for humans to determine what is true. We were talking earlier about why is technology stalled out. The reason technology stalled out, in my humble opinion, is physics broke somewhere around, call it 50, 60 years ago. It just got hung up. And we haven't been decoding the real world. That's truth. Now, once you're able to make contact with that ground level truth, truth, new things are open to you. And so that's my big concern with AI, is that we will not be getting informed by what is making contact with ground truth. We're going to be having the frame set and we're going to be taught as kids, as adults, as everybody, based on the frame that matches somebody's ideology and that scares the life out of me.
- [62:40]
Marc Andreessen
Yeah, it's. Should I agree with that, of all the radical things that Elon is doing, maybe the most radical is that he's declared that his goal and I've said we're investors in it with him. But his goal for Xai is what he calls maximally truth seeking. And if you've listened to him on this, what you know is he actually means two different things by that. I mean, they're the same thing ultimately, but two different angles. One is maximum truth seeking in terms of actually understanding the universe. And so to your point, actually like learning more about physics, but he also means maximally truth seeking in terms of social and political affairs. And so being able to actually speak openly about having the AI actually be fair and truth seeking when it comes to politics. And of course that's, you know, that is possibly the most radical thing anybody could do is build a legitimately truth seeking AI. And at least he's declared the determination to do that. So, yeah, there's a version of the world where he succeeds and that becomes the new benchmark. And, and by the way, open source AI plays a big role here because people can field open source AI to do this without, you know, permission. And so there's a version of the world where AI becomes an ally in trying to understand ground truth and trying to enable the, all the actual discussions and debates that need to happen. And then there's a version of the world in which, yeah, it's a, it's a, it's a, it's a Orwellian thought control. You know, my, my line on it is 1984. The novel 1984 was not written to be an instruction manual. Right. Like that was not the goal. Right. You know, it was supposed to be a dystopian, you know, future that we were trying to avoid. And so the idea that the machines are telling us what to think and that they're, that they're slanted and biased by the people who build them, I find to be completely unacceptable. But there is a, I mean, look, we have that today. Most of Most of the AIs in the world today are like that. And there is a very big danger. And by the way, those, and again, those companies, the people who are the most upset about freedom of speech, they I think justifiably aim Internet freedom of speech. They justifiably aim a lot of criticism at the companies. And I think that is valid in many cases. But I would just also tell you these companies are under intense pressure and there's tons of activists that are very powerful that are basically bearing down to these companies all the time but then also the government directly. And one of the things that has really kicked in in the last 10 years is governments both here and in Europe and other places basically seeking to censor and control even in ways that I think are just obviously illegal by their own laws. And that pressure remains very strong, strong and I think if anything that that pressure probably is going to intensify. And so I this for me is in the category of yes, these are the right concerns and then ultimately this is a democratic lowercase D democratic question which is, you know, do people care and are people going to be willing to stand up for this? And I think that's what's required.
- [65:21]
Host
That's it for part one, everybody. But the conversation with Marc Andreessen is definitely not over. Part two is packed with even more mind blowing insights. So be sure to come back tomorrow. I will see you then.