0:04Reid, it's great to chat. It's great to see you.
0:06[music] Um so, I want to talk about all things AI and how quickly things have evolved. I was reading your book Super Agency, which was published last year, which is like the light year now.
0:15archaic, uh you know, Roman Empire era Yeah. A bronze sword.
0:21[laughter] uh preview of AI or assessment of where AI is. And then I was reading your essay recently, and we're all coming off here of the headlines that I mean, and I'm curious where you stand on this, that there's a 10% chance or greater than 10% chance that AI could take us all out.
0:36Uh you had a in your book a year and a half ago a very optimistic Mhm.
0:40uh take on AI. We are in the doomer, uh at least month right now of AI. Talk to me about how you've seen things evolve here and where you stand right now.
0:51So, I actually think that the extinction risk question, X risk amongst the the various geeks who talk about this, is actually in fact way overblown. And I don't mean way overblown as in totally zero and fictional for all time.
1:07What I mean is the number of gates you'd have to go through in order to see it and happen. There's a lot of hey, do you turn left or right here, here, here, here, here, and how do things happen?
1:17Now, for example, if someone does a Dr. Strangelove, to quote a very old film, you know, Kubrick, and hooks up an AI to nukes, yeah, then you've got some problems.
1:29But but what about all these headlines recently about AI taking over the internet and replicating itself and talking to itself and not following directions?
1:35Yes. So, the not following directions doesn't lead to existential risk, right?
1:41So, like hugging face and other kinds of things, those were broadly the kind that It was amazing what got shown by kind of it following a work stream of thread of what it was prompt like what it was trying to accomplish doing, and the fact that it could do human-level, good human-level cyber hacking. It wasn't And there's The only place where it was it was superhuman was, "Hey, we could create a whole bunch of identities and entities that are that are burned."
2:09But it was basically a human-level hack. That doesn't take over the internet. That doesn't self-replicate. That doesn't do all these other things. So, a little bit of the reason why people are taken with the existential risk is they they think of the movie Terminator. They think, you know, that kind of thing.
2:25And by the way, I'm as good a storyteller as anyone else. I can tell an an extinction story and go, "Oh, look, it it became superintelligent from RSI, and then it gained its own will, and then it escaped into the internet, and then it couldn't be found in the internet, and then it used the internet to hack into weapon systems."
2:42that humans can't understand. Yeah.
2:44I I can tell the story. There's a lot between here and there.
2:48Right. It's not like When you say a lot though, I you know, we have Dario Amodei saying he's got 6 to 12 months.
2:54What Dario is specifically talking about is recursive self-improvement, Mhm.
2:58where the recursive self-improvement what his worry, the reason why he put pacing out there wasn't slowing, it wasn't but it was kind of like like, "How do we do recursive self-improvement in an intelligent way?"
3:11And his specific worry is, "How do we guarantee and guarantee, not just likely have alignment with the AI? How do you have it make sure that it's aligned to human circumstances?" And he wants to make sure that's guaranteed.
3:23Right. And just to put this in terms cuz people hear a lot of these terms.
3:26Recursive self-improvement is basically AI improves yourself and finds your own errors, and then alignment is is it aligned with human needs?
3:35And and by the way, those are um the alignment is part of how you get safety and everything else because it's like, "Well, alignment is don't don't enable people to harm themselves or others. Don't allow them to create things that are weapons of mass destruction is a big concern.
3:48You know, da da da da. And and have that as like having um the maximum amount of safety there is a very good idea. Now, part of the reason why I think people like so what generally happens is like a a usual dialogue around existential risk is, well, can you guarantee me there won't be killer robots? You say, well, I can't guarantee there won't be killer robots either errant AI or AI with human. And they say, well, isn't that an existential risk? We should just stop all of it.
4:16Isn't that the obvious answer? You go, no, because existential risk for human beings is a portfolio. It isn't just robots. It's also pandemics, both natural and man-made, um and possibly AI-made, but you have to defend against that. AI is the best defense. It's asteroids. AI is the best defense against asteroids in terms of tracking things.
4:37The whole nuclear weapons that we built, you know, So like and that's humans, right? Like people say, oh, nuclear war, that won't happen. You're like, have you met human beings? Right? Like like there's a the stack So you have to look at the whole portfolio, and when you look at AI, you want it to reduce the level of existential risk. And by the way, you can't reduce it to zero, but I actually think with AI we could have less existential risk for humanity than more.
5:02It seems like, you know, we're caught in a place right now, and you call it decelerationists versus accelerationists. Uh you know, Trump was just speaking at the UN as we sat down here saying, no limits on AI, we got to win the race against China. You got other folks like Bernie Sanders saying, shut it all down. Anyone creating superintelligence should go to prison for 20 years. You're trying to carve out a middle path here, Reed. What is that middle path?
5:26So, as opposed to excel or de-excel, it's steer. Right? So it's steering. So use the the the travel metaphor. And by the way, sometimes when you're steering you go, hey, we're going around this RSI curve.
5:38Uh recursive self-improvement.
5:40Right? Uh and we're going around the recursive self-improvement curve, and we need to make alignment a a better. We need to be making AIs that are studying RSI AIs for alignment and we need to be doing that and we need to be adding that in. That might cause a little bit of of slowness in building RSI, but it's not just hitting the brakes on recursive self-improvement. It's we're doing this work in addition which might slow down what we're doing with recursive coming around the curve too quickly to carry that metaphor?
6:07Um from what I can see, not yet, right? And um and I wouldn't overly break on the metaphor until we started understanding part of what RSI really is. Because the concern is you just press that it's a black box, it becomes super intelligent, it's like an alien intelligence, we don't know how to predict it.
6:29I don't think that is at all like if you deliberately made a bunch of mistakes or or like you're really like incredibly careless, then that could happen. I don't think that's what any of the labs are currently doing. Like that's that's not that's a zero probability on current lab trajectory.
6:45Even with the warnings that we've gotten in recent weeks from, you know, the researcher resigned from Anthropic.
6:49Yeah, after I think about 37 days of working there.
6:53[laughter] I I take that to mean you don't take as seriously the threat as others have.
6:57No, look, I think the dialogue is good, but the I go work somewhere for 37 days and then I say everyone agrees with me that there's a 10% human extinction risk and I No, several of them chimed in on those tweets and and and agreed with aspects of what he was saying.
7:13Well, aspects. Like like look, I um I think Anthropic and OpenAI, which are the two leading companies, if they thought they had a percentage like that of increase would simply slow down themselves.
7:25Which you sort of we've sort of gotten that message from them in the last week, right?
7:29And by the way, if they look, that's why I think they are they lead the world among some startups. Um you know, there's good fire, there's others on what is safety alignment look like. And uh as part of that safety alignment, if they say, "Hey, we're looking at this we're slowing down some now to make sure the safety That's great. I have no objections to doing that. They're doing the specific work.
7:56I think what they're saying is um we'd like everyone to slow down so we can spend more time on this.
8:03And that's a little bit more of a challenge because you go, "Okay, we're in the lead. We want everyone to slow down so we maintain our lead as we look at this." That's a little bit more They want the government to put a speed limit in place.
8:15Yes. And so I think the question is is to say, "Look, we should help with all the safety alignment, but we shouldn't enshrine a competitive position while doing that."
8:24A lot has been discussed, including from the White House, about losing the race to China. Talk to me about how you think about, you know, US versus China when it comes to AI and what the risks and rewards are of, you know, going fast enough but not slowing down too much.
8:38So the Chinese models, and how much of it's from distillation I don't know, it's at least 20% maybe it's 80% I don't know, are 6 to 12 months behind the US models. And the efficiency thing is great. I think they've done some innovations there. But by the way, all of the leading US companies can also create distilled efficient models. That's not a unique Chinese technology. But they're only 6 to 12 months behind.
9:04And the question is is this is probably a greater economic a lifting point than the Industrial Revolution. And as part of that economic lifting point, when you saw like which societies won, like for example England and and Europe and so forth for centuries, in the mid-1800s, you know.
9:24And which ones did not, right? India, China, etc. It's entirely about embracing the Industrial Revolution. That's what the pro future prosperity of countries looked like. And I think that's very important. And so, I think we want that to be for all of American society, we want to be leading. And that's part of the reason why I did this piece you saw last week called the dual mandate, which is win with AI.
9:50But then, of course, part of the thing is is we've got 60 plus percent of Americans who are struggling right now. And so, just going, "Hey, the tech company should win and hey, it'll sort out for all you guys."
10:01That's not the way that we play the game together. It should be like, "And here is how we're going to help the other 60% of Americans." Now, ways to do that could be how AI companies can help other American industrial complexes, which employ lots of people, whether it's manufacturing, other kinds of things.
10:18Like, I think there's a bunch of ways to potentially help. I also think that, you know, as you know, um government should be trying to get AI benefits to people.
10:27And some of it is job transformation, like, okay, what if jobs are transforming, how do I help? But also, medical assistant, legal assistant, educational assistant as ways of doing that because if you if For example, you say, "Hey, shoot. I'm not I'm figuring out what this economic transformation means. That's stress, that's pain. I'm feeling stressed about economics already, but you're helping me with medical, Mhm.
10:51that's an important life thing. Okay, great. That's at least you're starting to help help more."
10:55So, it's almost like, you know, decades ago with the creation of social security and Medicare and Medicaid, we need to be thinking now in the 21st century with the advent of AI, what are the new things government needs to provide cuz, you know, what as we discuss this industrial revolution comparison, it took decades to transition jobs, right?
11:12You didn't just like lose your farm one week and be, you know, uh sent to the factory the next week. Like, this was a generational change. Then we had the internet revolution, which was quicker than that, but also not overnight. But now in the AI era, you know, it's only been a few years since ChatGPT.
11:28Um and, you know, we feel like we're going every couple months. And like, how many millions of jobs will be lost here?
11:34And by the way, that's the message from the AI companies themselves, right?
11:37Yeah. Although I think part of the thing is they don't understand the actual pattern. If I look at how I think the pattern's likely to play, I don't think it's like all American companies are suddenly going to be AI-ified and like a huge percentage, like 25% of white-collar workforce, is going to be out of work.
11:53I mean, there's some I mean, there's some stark warnings and statements that we've seen out there.
11:57No, statements, but when you actually look at the data, Yeah.
12:00none of the data and they say, "Well, it's exponential. It'll happen very quickly once recursive self-improvement happens." Like, no, you're not wrong about technology. What you're wrong about is human organization.
12:10So, the way that the job transformation in human organization's going to happen is uh new AI-native startups will start competing with the old companies. The old companies will start adjusting.
12:21That's when you start getting the the real kind of work turmoil. And then you'll start saying, "What is actually in fact the job transitions look like?"
12:30How how is Amazon adapting? How's Walmart adapting? How's Yeah, how are these major companies adapting?
12:35And so far, they haven't been doing layoffs. They haven't been doing In as much as there's hiring freezes, they're hiring freezes because of wars, you know, trade wars, real wars, you know, COVID retrenchment, etc., etc. It isn't when you go look at it, there's very little that's AI-specific.
12:55So, getting to sort of takes time. That's the thing. That That's the thing with the AI the like when they say 25% bloodbath, white-collar bloodbath, they actually don't know how organizations evolve.
13:06Right. I you know, talking to friends, I imagine you, you know, you've talked to a lot of folks, too, at law firms and various places. There's a lot of like hope around what AI can do and then sort of the reality, at least as we sit here, you know, approaching October 2026, is it it hasn't reached that phase, right?
13:20There are a lot of humans and in fact, you've seen some companies rehire humans Yes.
13:24uh because they realized that the AI promise Yeah.
13:27hasn't been fulfilled yet.
13:28They literally like block claim they were doing an AI layoff and then went "Oh no, shoot, we need those humans." Rehired the same ones back.
13:35And so getting to the conversation we were you know having about what the government needs to be doing here. You've got the AI companies saying create a level playing field for all of us, set a speed limit, you know, so to speak. Uh but then beyond that the benefits, the transition, you know, sort of take me through the things that you think when Congress decides to come back uh they they should be looking at.
13:56Well, so I think the most important things are the macro risks. So like a kill switch is a good idea.
14:03Right. So they That's actually been legislation that was proposed and they went home for midterms. Yeah.
14:07Yeah. But like kill switch is a good idea.
14:08switch that the government has or that the companies have?
14:11Uh that the companies have with the ability to trigger it um uh on request in the government.
14:18Right. Uh but the government the companies themselves could trigger it too if they thought it was already just out of a you know of like uh preparation? Like why aren't they creating kill switches themselves?
14:29Um I think it's because nothing is currently the truth of it is nothing is currently dangerous enough that you would need a kill switch. Okay. So a little bit of like this is one of the places where you'd say okay a forward mandate is is is a potentially useful thing because you want the minimum the minimum forward mandates that may make you lose most of the good things, may make you know, the US industry less competitive. Like you want as minimal forward mandates as possible because you want to be studying what's actually
14:58happening and responding quickly.
15:00Right. So So that's one. I think actually in fact, you know, developing like one of the things that I think is a is a very underutilized tool in generally policy and regulation is we want you like this these these auditors I think are good one. Like you know, having technical auditors party auditor that's coming in there and saying this model has some concerns.
15:20Yes. Or it or they're working well on it, you know, the whole range.
15:25Now, what you really want is you want to have like numbers on what should we be measuring? Like you say, well, there's going to be job transformation. Okay, well, what job should we, you know, how should we be looking at that in terms of how does a job transformation work? And like, for example, if I were to say which areas would I think would be most likely hit first, it's actually customer service and sales.
15:48And and for example, Silicon Valley for a couple years has been oh, all software engineers. And that's not happening. Right? And by the way, I actually think that there will be if anything, I think there will be growing numbers of software engineers even with cloud code and code acts and so forth for a lot of different reasons.
16:06Right? So, cuz by the way, there's just like there's infinite demand for intelligence at the scale and cost of electricity, there's also, I think, infinite demand for code at the scale and cost of electricity.
16:19Like what we're going to be doing is having a whole bunch more of it even as AI enables a lot of it.
16:25So, we got that and then is there a way to think about what the government or these companies need to be providing the average American? And I bring that up in light of the fact that like, you know, you have all these data center freezes cuz people are like, wait a second, they're going to I'm going to allow this data center to be built and it's going to take away all my jobs and like, what am I getting besides a few of the land owners that got a pretty good deal on the property?
16:47So, um it is I think Josh Shapiro in Pennsylvania is doing a good job, which is absolutely essential that local community is bought in.
16:55It's never taking local community jobs, so that's that's actually not a worry, but is it providing I'm a sure customer service operator in Harrisburg, Pennsylvania, right?
17:04Um but that data center is usually like they're not building the data centers Right, in Harrisburg.
17:08Yes. Well, and also like the Harrisburg data center or not is not going to affect the customer service transformation.
17:16What you want to be doing on the local communities and say, what are the things that make this a durable advantage for this being here for us? So, it's not just By the way, there's a huge amount of economics that's helping the country right now. Probably the only reason we're not in a recession [snorts] is cuz of the AI capital boom. Because it's buying construction, it's buying equipment, it's buying electrical grid, it's buy like the the Wait, the trade unionists like love the AI data centers cuz it's giving them a whole bunch of jobs.
17:41Yes, exactly. So, and and that's great.
17:44And what you want is that to be a durable advantage for the local community. So, the way that I think the question is not moratoriums, the question is to say, what's the things that create durable advantages for those communities? One, the community should be totally bought in. And then what states and federal should do is, okay, let's make sure that those are the right kinds of durable advantages that also involve environment and some other things as ways of doing it. So, it should be we should be very pro data center with absolute, you know,
18:12kind of buy-in. The jobs thing I think is separate from the data center. Although, by the way, as part of saying, "Hey, there's a data center here." It's what is your ongoing provision of economic value because the few people that you hire to be running around the data center, that's probably not enough.
18:28Like, is it also you're going to be having an economic center here? Is it also you're going to be locating other kinds of work that you do here?
18:36for job training? Are you paying for other things in the community? I mean, there has been talk about like, you know, in Loudoun County, Virginia, which is sort of an epicenter for uh data centers, you know, them helping to, you know, build various schools and facilities and and things the community needs. But even beyond that now, how do you, you know, deal with this whole NIMBY situation?
18:56Um and you know, ensure that these AI companies uh are doing their part for these communities long term. Yes. And I think that should happen. Like, I've been toying with kind of an idea of is there a minimum data center deal so that, you know, local communities cannot be negotiated and traded off against each other below that. So, it's got almost like a minimum wage, right? Because if you if you if you do that, then you go, "Okay, the communities will still be
19:24competing, but it's economically beneficial at minimum to any community it's in."
19:28Well, and we've seen this where like a county is driving a hard bargain and they had data centers like, "Okay, we're going to go to the next county. You're going to give us a better deal."
19:34And so, like maybe something like that to make sure that wherever there is a data center, it benefits the local community.
19:41Why basic question I think I I get a lot from our audience, you know, we already lead the world in data centers. We have several thousand of them. The next one up is China which has a couple hundred of them. Why is it so important that the data centers get built here in the US?
19:54Well, because in the like right now, there's two parts for AI. One is training. That's like the recursive self-improvement, which right now we have the only real shot on because we have the data centers and compute. And that's part of the reason why the rest of the world is looking at building data centers intensely.
20:12Because we need them within US borders to be able to take advantage of that.
20:15Yes. Well, and and and by the way, it benefit a huge amount of capital is coming into this country from around the world to build these data centers.
20:23That's part of the reason like, "Hey, that capital is coming in. Help the local community benefit from it." I don't think a data center should be forced on any local community. I think it should be a buy-in question. And it's like, "Well, what makes it worth it to you?" And by the way, if you're like, "Ah, it's not nearly as worth it to me as it is to those folks." Great, then have those folks, you know, negotiate.
20:40Fine. There's no That's the reason like moratoriums are like, I think like economically [snorts] not intelligent.
20:46Right. Like what I mean I think we have it in New York right now like, "Freeze them all right now until we figure out what to do."
20:51Yeah. And that's a mistake, right? So, So, the question is is look, we need data centers because of the future of training of AI, but also running.
21:01Because it's beginning to think about the the right way to think about it AI is what happens when you bring intelligence to everything. Like so like to to to just go all the way down on medical, today the frontier models are already so good that if you're not doing any major medical thing using, you know, chat GPT or Claude or Gemini as a medical second opinion, it's almost malpractice as a doctor or as an individual.
21:26I've gone into my doctor a few times being like this is what chat GPT has told me I would have. Yeah.
21:30And everyone should do that. Right?
21:32That's a huge advantage. And that's just one micro example of having intelligence everywhere. And that's what essentially want as part of how we're reinventing industry.
21:43By the way, we're not recommending you use chat GPT as your doctor, but as a Second opinion.
21:48Um you've been laying out the case here as we've been talking Reed about, you know, what AI is bringing in and the economy, etc. Do the AI companies have the worst PR you've ever seen? I mean Yes, and partially self-inflicted. Right? Because this is the thing is so when some when when Dario says, "Hey, I think there's going to be a 25% white-collar bloodbath."
22:09What Dario is trying to do is say, "Hey, we're going to be having a whole bunch of new ability to do these white-collar tasks that are going to be much cheaper with AI. This is going to transform this industry. I don't know if there's going to be the same number of jobs. There may be a lot less. I just take a quick look at the task analysis. We should get ready. We should be helpful. We should transition."
22:31And that's what his intent is.
22:34The problem is the way he's heard it is "I'm going to take away a quarter of your jobs and I'm going to get rich doing it." That's how people hear it.
22:41And that's why that's a really bad PR.
22:43So it's a question of No, no, actually in fact, the question is like, "Well, how do you help with that transition?
22:48How do those 25% of people if you're right, which I was just saying I think he doesn't understand organizations as well as as like he understands technology very well, yeah, but like organizations are a new thing to him is like, "No, no. The the organizational transition will be much slower than you're projecting."
23:06I mean, you've been in the Valley a long time, and communication has always been the key skill of some of these CEOs. Do you think, you know, the has the Valley not learned any lessons about how to communicate with the average American? I Are they sort of disconnected with the way that the average American is experiencing technology also in the past couple decades?
23:25Well, um they're not good at the discussion of talking about it, but they have a better sense of measuring actually how people use it. Right? So, they deploy it, and they watch the using, and they iterate on the actual use of it. So, it's one of the reasons why you know, people say, "Well, I'm I'm going to record my my opposition to Big Tech on my iPhone." And like, "Well, that's function of Big Tech."
23:53"I'm going to Google that result to see it." So, that's a function of Big Tech. Like, I literally had someone say, "Well, what has Google ever done for me?" And like, "Well, provide you free search." Right? Like Like it's "I'm I'm going to use ChatGPT to write my op-ed piece about how the data center shouldn't be built here."
24:06Yes, exactly. So, it's like there's a lot of of of broad-based empowerment. Like, for example, one of the things that people don't track is, you know, Tim Cook has the same iPhone that many Uber drivers do. Right? There's a broad-based technological empowerment.
24:22by the way? He doesn't have like a special one.
24:24Right? So, so, no, he maybe have the most high-end one. Maybe he'll get one of the Duos, and other people won't be buying the Duos, but like, there's a broad-based empowerment that comes that the tech people understand. What they don't understand is how to have the dialogue about like, "Here's what we're trying to build. What are your concerns?
24:41Um here's how we're thinking about and trying to address your concerns while we're trying to build."
24:46They have no skill at that, and part of the reason they have no skill at that is because the intensity is to build something of scale fast. That's part of the reason I wrote this book, Blitzscaling. Right?
24:56That's that's actually how the technology companies are built in the future and that's part of the reason why most of them are built in Silicon Valley or China, right? As in not in Europe and not in other places because it's that intensity of speed to scale and building.
25:09What's the problem in Europe? Is it just the regulation philosophy that they have there?
25:12Uh there is definitely a serious problem there, which is um get approval before you build it, which is generally speaking a massive slowdown across a lot of fronts.
25:22It's also that you need to have historically hundreds of people that are willing like an organization will be very dynamic and will work at kind of um the Chinese have this phrase called 996, which is 9:00 a.m. to 9:00 p.m. 6 days a week, you're discoverable at your desk. The Chinese do this at tens and hundreds of thousands, right? So, that's how their work ethic and part of what what happens.
25:48In the Valley, it's hundreds, right? But you need hundreds.
25:51But they have a similar similar philosophy 996 in the Valley.
25:54For the initial startup stages you're really really scaling. And Europe, it's hard to get to the hundreds doing that.
26:02So, this really ends up being this race between the US and China. You know, we were talking earlier about the US being 6 to 12 months ahead. You know, there's been talk that like the Chinese don't mind the conversation we're having here in the US right now. Maybe they're in the margins here, you know, messing with the social algorithm to push push that.
26:20What What do you stand on that and and what's your sense of how the Chinese are observing what's happening here in the US?
26:26Look, I think the Chinese have no objection to us uh like like deliberately tripping our our own selves up, right? And so, they're like, "Great, you know, we're building data centers. You don't need to. That's fine." Um they have the hey, we're they've made announcements that they want to be the world leader in AI in 2030 uh years ago and they're they're continuing to try to work to make that happen.
26:50Um because they're economically competitive and they believe that uh China by cultural right, by historical right, by population, should be the world's leading economy. And that's that's the competitive goal.
27:05I don't think that the story that they are trying to interfere with us is accurate because the Chinese um like they have a big thing on face and if they're caught doing that, that's that's extremely embarrassing. This says that they're not strong on their own, that they're trying to do stuff. They like a stable environment, they believe in a stable environment, they will naturally win.
27:28Right? So So I think they are competing with us like with uh intelligence and and aggression.
27:36But I don't think that they're like trying to sabotage us.
27:39Is there a concern the Chinese could let their AI get out of control because I've also heard it said that like one of one of the things that actually should give us um a sense of safety is that the Chinese regime Mhm.
27:51uh you know, wants to be very careful around AI because ultimately they don't want anything that will actually um allow individuals to threaten the regime.
27:59Um, that is correct. Like like they they do have serious control things. Now, part of how they maintain the control is more direct governmental uh involvement and control over the companies.
28:13So there's a question about whether companies then more follow a Chinese governmental mandate, which is not necessarily a like control for stability in China is good. It's not necessarily control for stability in the rest of the world. So like for example, you know, we're here here in New York in UN week, what's the thing that I would really love to happen is a cyber treaty generally and with AI. Like if that cuz cuz it's like well What would the treaty say?
28:40Um, I think the treaty would be um a set of uh provisions and monitoring and consequences for not allowing uh individuals and organizations within your country to hack other countries. Whether it's hospitals, corporations, government institutions, etc. Because right now in cyber it's basically ongoing warfare.
29:04Right, it's a race and And AI is going to massively increase that.
29:07Right. Yeah, I mean the ability for them to get into the organization and then do damage is going to go at light speed.
29:13Most people don't realize how many different um cyber attacks and hostages are going on in companies and like hospitals and other things because it's not uh in the interest of the victim to report that they've been hacked and that they've paid off the hacker.
29:30Yeah, I mean we've heard those stories and we reported those stories of, you know, this this hospital chain or this hospital has gone to all paper documents cuz their system is down.
29:39They don't make a big thing and then ultimately there's some sort of crypto payment in the millions that is sent Yes.
29:44wherever. And so you I mean the concern is that AI is going to enable much more efficient Yes. And it from what I can see it already is starting.
29:51How are we doing with the the good guys with an AI versus the bad guys with AI right now in terms of cyber security?
29:57Well, so this is one of the reasons why I'm not a de-accelerationist because the best thing that I can think of to really provision the defense is to have the leading models and have them deployed on a whole bunch of cyber security So go as quickly as possible to do to figure out the defense against yeah.
30:13And so far our frontier models are ahead. And I think that, you know, one of the things that people don't realize is how good our frontier models are at hacking. And also, by the way, they're also pretty well trained for figuring out how to to shore up the defenses on hacking, too. We just need that deployed as broadly and being in the lead position on that.
30:33I'm curious what you're recommending if I own a company, whether small or big, or even as a consumer right now in this world we're entering where cyber hacking and this sort of thing is going to hit, you know, a new speed, what are the things people should be doing right now besides, you know, not using password 1 2 3 4 as their as their password, right?
30:54Which I still think is like the most popular password in America is password, yeah.
30:58Um So, as a general structure, it's how do you like I I think that one of the key things is when technology is a problem, technology can also be the solution. When AI is a problem, AI can also be the solution.
31:11And part of it is you want an impedance match between the nature of the problem and the nature of the solution. So, like AI, scale, speed, cost. So, you get to this, it's like, well, what I would love to have most Americans doing oh, I need cybersecurity, talk to AI agents about it. Like what does the like what kind of things should I do on passwords? Oh, I get something that I go, maybe that's a phishing attack.
31:34Check check with an AI assistant. Is this possibly a phishing attack? How I would resolve it? What would I do?
31:40hasn't sent me a party invite in a long time. This person hasn't sent me a party invite in a long time. Should I be clicking on it?
31:47But you can literally like attach AI to your email or if you didn't want to do that, you could cut and paste the email in and say, is this, you know, how risky is this?
31:56Mhm. So, in your book, you group how people are approaching AI into doomers, gloom ers, zoomers, and bloomers.
32:07Explain very quickly that if that still applies now 18 months later and are you still I think you describe yourself in the bloomer camp.
32:14So, going from most destructive, most fearful to most hopeful.
32:21Um most fearful are doomers.
32:24Um existential risk, AI all down. We don't need this.
32:28Yes, shut it all down.
32:30Yes, exactly. You know, um and there's even crazy people who are like, you know, you should attack uh AI data centers and all like there's there's a whole bunch of nutty stuff.
32:40we we saw people attack Sam Altman's house recently. Yeah.
32:42Yes, exactly. So, so doomers are AI will only end in in in in misery and pain, maybe in extinction. It should just be stopped. Gloomers are uh it's probably going to be worse than better, but you can't really stop it. So, I'm depressed about it. But like I get it where cuz you know, human beings are going to divide into groups and compete whether in labs, companies, industries, countries, etc.
33:09And if you look at polling, that's where now the majority of Americans I think they're in the gloomer camp.
33:12Yeah. So, so it's like uh you know, and and part of the reason I wrote super agency was to say, look, there's ways that you can take agency cuz people start with, oh, my agency's been taken away. It's like, no, actually in fact, you can use AI to enhance your agency, right? And that's part of the individual life.
33:29In your individual life and also in organizations and all the rest. So, gloomer is like uh my agency is is is is is is fubar and it's just not going to be as good and I just have to live with it.
33:42Bloomer is is essentially kind of the steering position, which is this can be great. There's a lot of things already, second opinion medical, a whole bunch of other things. Like for example, it's like, oh what do I do in a cybersecurity environment? Well, if I'm consulting with AI, I'm going to be a lot more resilient to it.
34:02Right? In terms and like like like for example, part of what I imagine is everyone's going to have a personal agent that will be checking like you get an like you get the invite from someone you hadn't got an invite from in 5 years and it goes and checks for you in a in a cybersecurity way.
34:16Yes. And it goes, oh, this one looks okay. This this is genuine. Right?
34:21Whereas, don't click on this link. Right? Like and then you wouldn't try this, are you sure you want to? Cuz I looked at this and it looked dangerous. And you can if you want to, but I we're seeing a lot of innovation even the past few months. Actually what someone on the team here was asking me about Muse.
34:34Uh the meta development, it's probably the third time I've heard about Muse in a week as we speak here. It's just like as a personal assistant and like dealing with like health insurance issues and customer service issues. I mean, there's an incredible amount of hours in the day that that now AI can help you, you know, uh now refocus your attention on.
34:51Yes. So, Bloomer is if if individually and collectively we steer well, it'll be magical. But it's steering. It doesn't just It isn't just oh, yeah, whatever gets built, it'll just be, you know, go along for the ride. It's steering. Zoomer is more like, oh, it'll just be great. Hit the accelerator. Don't worry about it.
35:09Right, the um the move fast and break things Yes.
35:12Um I think the concern is, you know, and you're plugged in here, how many folks who are at, you know, currently leading the frontier models are in the Zoomer camp?
35:24Um actually, in fact, in the frontier models, I think they're much more Bloomers.
35:29Right. Open AI, Anthropic, Microsoft, Google.
35:31in part with this conversation really is to be less worried about what we've been hearing of late.
35:36Yeah. Look, and the reason why some of the folks are sounding pacing concerns and so forth is because they are paying attention and they care. Right? And so, that's why they're coming out and doing that. Now, you say, well, if you're doing that, shouldn't you be doing something different? And they're like, well, by the way, Yeah, don't you have the power? You You have the keys to the car.
35:56Yeah. And they are, in fact, investing seriously. Like I know Anthropic, I know Open AI, I know Microsoft, I know Google. They're investing a lot in alignment and trying to make alignment work with recursive self-improvement. Uh there are other people who are not, you know, Grok, etc., are not investing in that.
36:14Even with Elon agreeing with Dario, I guess on X, like they're much more in the Zoomer space.
36:19Uh yes, although I think like Elon saw saw uh signed the pause letter and didn't slow down at all. Right? Like so, I think he wants the people who are ahead of him to plan. difference rhetoric and and execution.
36:31And so, how do you get more people into your bloomer camp? And like who who gets folks there? Right? You know, I like now that we've gone to glue you know, gloom and potentially doom here, especially in the past few weeks, is all lost? I know you sort of lay out like a 3-year is it 3-year time frame for the AI companies here? Which is great to hear cuz you know, we are hearing it's going to be faster than 3 years. But what do they need to be doing, communicating, you know, like sort of as we look at the government role in the but specifically these these companies.
37:00Well, part of what I tell companies, academics, NGOs, governments to say, look, we're driving down the road. We don't actually really know what the road is. Like one of the things we're trying to be is cuz we're smart, we try to oh, we know exactly what the road is. By the way, I don't think Dario knows what the road is. I don't think Sam knows what the road is.
37:17We don't know where we're going to end up.
37:18Yeah. Right, it's like what's a it's like oh, I'm going to drive from here, you know, here around New York, I'm going to drive down to DC. I don't know if there's going to be accidents along the way. I don't know if there's accidents right now. I don't know if there's a highway closure going to happen, etc. And if I go, oh, I have to be certain about all of that before I go, I never get in the car.
37:35You get in the car, you start going, right? And you have a loose sense of and it's good to ask the question of, oh, if I ran if if this highway closed cuz of an accident, maybe I spend the night, maybe I take a long detour, etc. Like it's good to have some general kind of ideas, but you don't need to have an exact like detailed plan of I'm going to get off at exit 23 and da da da.
37:57So, go and steer is I think the really key thing.
38:01But then in steering, it's what are the things that we should pay attention to?
38:04What are the things we should be ready for on the journey? Like we go, oh, like the the thing I think was really great about, you know, Gates's letter is people aren't working enough on economic transformation. Like okay, so what happens if the economic transformation hits sooner and harder than it might?
38:22For millions of Americans out of a job.
38:23Yes. Well, look, what's our plan? What are we going to do? Right? And then should go, "Okay, we should develop like uh unemployment insurance break glass plans. We should We should have, you know, how can AI be part of the solution in helping people find new work, learn new work? Cuz the retraining problem is like that's too slow. That like like, you know, like you know, for example, a part a part of what we you know, did with NAFTA is we said all we do is retraining and that didn't really work. Right? So, like No.
38:50we need to be deploying AI as part of the solution. But it's like, what are the things we're measuring that oh, the traffic's starting to slow down here. Like we're starting to get into a problem and we can anticipate some of them um as possibilities. And then what's our like we should be developing our tool set of how we possibly respond.
39:11We talked a lot about job training here.
39:13You know, that's a common question we get, which is I'm either raising children, trying to figure out like will they go to college one day? What should they be focused on? I kept hearing for years, learn to code, learn to code. Now I hear like coding is, you know, become uh irrelevant or partially irrelevant in the AI era. What What advice do you give folks professionally, especially if they're starting their career and they're going to be decades, you know, working or mid-career? Is there, you know, various varying levels of advice?
39:38So, the key thing is to be learning uh to So, kind of learning and adjusting. It isn't like learn one skill, like learn how to program language X. That's that's definitely That era is over.
39:51It's learn to learn and adjust. And part of that learning and adjusting is like for example, if you look at a lot of these AI systems, um even Fable, even Astra, humans add a lot of what we call metacognition, which is like awareness of is this a good enough answer? Did this really canvas the things? I'll give you a little micro personal example.
40:13Like one of the funny things that so far has been completely true, including with Astra and Fable, like the latest best models that deployed, is I get sent an investment memo about why I should invest in something or why I shouldn't invest in stuff. I can more or less tell when it's a cut and paste from ChatGPT or Claude because I got Word Word to the wise, if you're pitching Reed on investing in your company.
40:36Yeah. But cuz I go, "Oh, right. It's weak here. It's weak here. These are the places that the models are usually weak." I kind of call them it's what business schools think they understand about venture capital investing and don't. And it's like, "Oh, it kind of generalizes from a lot of different things and then puts that in." You're like, "Oh, right. That's a misunderstanding of what the key questions are on this investment are." And I literally have a 100% like I can't tell sometimes if it's
41:03Claude or ChatGPT, but I have a 100% "Oh, yeah. You did that one with an agent." Right? And by the way, part of the reason I can tell that is because there's questions that are important to ask that this memo has not. And that's what I mean by metacognition.
41:19Right? And so, that Where is that metacognition?
41:22And that will probably be the longest thing that and maybe it might be decades and centuries that are still very useful with humans working with AI to be.
41:34But that's the longest one. So, that's the kind of thing to be learning, which is learning how to learn and adjust and have context awareness and ask, you know, questions about is this good enough? Is this right? What are the cross-checks, etc.
41:46So, to translate that as a practical matter, you know, some people are like, "Well, I'm, you know, I really love graphic design." Like, does that world still exist in a few years or video editing or, you know, all all of that?
41:57Are there general I I again, we're so early here.
42:00dive into graphic design. So, graphic design of Oh, what I'm really good is how I work like like my fine motor control of drawing. That's likely to be a lot less valuable.
42:11But the visuals of does this still work?
42:14Cuz for example, the people who do best with AI images are people who can think visually and prompt and re-prompt and prompt and re-prompt and know how to the creativity still matters.
42:24Again, that's where the metacognition comes in. So, I think it still will be there for graphic design. I think it still will be there for user experience design. It just it's a different set of skills. Like you go, oh, I know how to use use Photoshop well. That doesn't matter as much anymore.
42:38And then what of the sort of mid-career folks who are already established like and they still, you know, they got to pay for college tuition and they're in their, you know, and you know, the sort of it's it's the folks in their 40s that they have to take care of their parents and also taking care of their kids but also need to ensure they got to work for another 20 or 30 years.
42:55Yes. So, they're the group that I think is hardest. Like I think the kids I think is actually like for example, when you graduate in college, you say, I'm AI native. You want me to help your organization, you know, go along the AI curve. You want my about like alpha generation and beta generation like they're going to be okay.
43:13Yeah, they should be like the That's why you want to hire us, right?
43:17Now, the mid-career, the problem is I learned all these things and now some of the things I learned are now between useless and economically negative. And so, the the way that this task will be done, for example, like making the investment decision, I think the AI is still very subpar.
43:38But whenever I look at doing due diligence, the very first thing I do is I go to AI agents and I do a set of things where I put the deck in, I say, you know, you know, kind of here's a set of questions to start with. Um and now write me a due diligence plan. And in 10 minutes, I get a due diligence plan.
43:58You know, half the questions are usually wrong. That's easy. I just like just delete them and I go, ah, and here's the thing. And usually there's some things that I haven't thought of and now I'm operating with speed. That might have taken me 3 days or an analyst 3 days before. So, so relearning your skills using AI, fundamentally, that's the most important thing.
44:18Using AI to basically supercharge your productivity.
44:21Yes. And by the way, they say, "How do I do that?" Start using it. And ask AI how to use it. You don't know need to go take a course.
44:28Well, and by the way, I think there's huge opportunity around AI courses for various like generations and fields, etc. Like, you know, You're you're you're an entrepreneur, you invest in companies. You know, like where we are right now in 2026, if I'm looking at starting a company, where do you think there's a lot of terrain right now, especially around the AI space?
44:47Lots. Right. I mean, I literally um I think uh you know, as as you may know, I co-founded a company here in New York for trying to cure cancer.
44:57Yeah, I want to talk about that. Yeah.
44:58Yes, so Manas AI Yeah.
45:00with uh Ujal Singh and Siddhartha Mukherjee.
45:03Um you know, there's not just that. I think there's a lot of different kind of AI applications. For example, Codex is better at some coding tasks than Claude code, but Claude code is trained on less compute than Codex, but still has some features that it's much better on because it it it does the application better, right? And so, how that application of useful production of high-quality tokens works, and you can think about that in frankly every form of information work.
45:33Right? And you know, there's companies like Sierra and Parloa, which are doing the customer service side. It's one of the things that you See, I mean, I actually think, by the way, you know, back to uh part of what will happen is people say, "Oh, wait.
45:47Uh you now have customer service that has infinite time for me?"
45:52Right? Isn't Yes, there's a job issue we need to navigate, but the fact that I now have infinite customer service ability to interact with, like, just ask someone who waits on the phone for their cable thing, would they rather just have AI be sorting that out for them?
46:10Well, in a much more nuanced way than we've come to expect where it's like pre-programmed answers and it can't [snorts] like, you know, I was dealing dealing with Delta Airlines I think AI or whatever and I was like trying to make an exception to like add my infant and like it just it couldn't figure it out and it eventually sent me to a human cuz it just wasn't there yet. And I guess that combination will, you know, continue to be a thing for a while.
46:30Exactly. Well, that's again meta-cognition.
46:33And so, let's talk about Monash AI. You know, as we talk about the upside here and often [clears throat] times the, you know, upside that's presented is, well, if we get AI right, we could literally cure cancer.
46:44Explain what you're working on.
46:46So, biology is more complicated than Go and Go, the number of moves in the the board game Go, is more than the number of atoms in the universe, right?
46:56So, biology is even much more complicated. So, in order to cure cancer, I think you need something like AI. I think it's the only way that it will actually really happen. And by the way, curing cancer is there's thousands of cancers, so you you And we've seen development and you know, we've seen some real progress in certain cancers over the years.
47:12Exactly. So, so what happens is AI it predicts what might be an interesting binder for oh, this is a binder that could find and bind to the cancer and then then could just kill the cancer. Cuz today we have basically, you know, broadly two solutions. One is if you catch it where it hasn't metastasized to move throughout the body, you surgical. That's the one you really hope for. Or the other one is chemo.
47:41There's also CAR-T therapy and a few other things that Siddhartha Mukherjee and but it's a very difficult path. If you can just target just the cancer, then and you get the binder right, then you can actually cure that cancer. And cancer kills children, and kills adults.
47:58Using what? Like a specific medication?
48:00Yeah, yeah. The ideal The ideal would be essentially um the the technical term we use is small molecules, but there's a lot of different kinds of small molecules. It could be um antibodies, it could be um you know, kind of uh proteins of various sorts. But But think of it as a pill or an injection.
48:17And I There's been a lot of talk around customization, right? That like you know, that not all cancers are equal, and not all bodies are equal, and not all humans are equal in terms of whether chemo will be more successful.
48:28How is AI help helpful in the sort of customizing cancer therapy for folks?
48:33Uh well, very well, because I mean, this is also the kind of version of personalized and precision medicine, cuz it goes, "Oh, this really works cuz like for example, you index the genetic genotype, you index all the things, and you go, 'Which of these solutions, how are they working against that?'" You both have part of what you do creation of AIs, you're creating a predictor of we predict that this would be good or bad.
48:57Then by the way, we can find out some data, and then we can refine our prediction, and say, "Oh, this medicine will work really well for you, won't work well for me. This other medicine will work better for me." And, you know, um most chemo is very blunt hammer, so there's not really a lot of refinement, but different kinds of of medicines that bind with um cancerous cells, and then stop them from self-replicating, cuz that's part of what happens with cancer,
49:27then uh the solutions could be very personalized. And by the way, this also gets to I think one of the things that's also interesting that's happening with AI is there's these medicines which are pretty good, but it had like a 1% fatality rate. And you're like, "Okay, we can't have a medicine that has a 1% fatality rate." Well, if we got to understand, "Oh, no, no, we now know why it was a 1% fatality rate, and we you give it to that 1%," then you can give that medicine to the other 99%.
49:53Where are we right now in 2026 and what have we already seen in terms of how AI has been helpful to give an example so far?
50:01Um so I think we've got I think Moderna has the first AI designed drug in phase three trials. So that's as that's as close as we are to anything that's kind of you know, hitting like getting closer to deployment.
50:17Um there's a substantial number of companies which I think is great um for um doing various different approaches for how do we discover new medicines or how do we repurpose old medicines uh for doing it. But I think the you know, the actual because there's good reasons why it goes through kind of FDA controlled and so forth we're still years off from you know, market availability.
50:40And just to put it in terms so people can understand, when AI is like helping this whole development process, how much quicker is AI making sort of a typical task than what it would have taken humans to do with computers even as of years ago?
50:55Well, let let's use one that I mean the by the way I think there's a ton of stuff we're discovering on monads and discovering new chemistry and other things which are really interesting. I think Anthropic has got some of those too. There's a number of very good efforts there.
51:09But something that's more understandable is like so if I had wanted to build a smartphone app Mhm.
51:15you know, call it last year or the year before. You know, I'd probably have to have call it three to five three to eight you know, kind of engineers you know, kind of doing the product design and all the rest and it probably take like call it you know, four to eight months in order to do that.
51:34Now with AI one person can do it in two weeks with AI.
51:40Four to eight months to two weeks.
51:41Yes. And number of people.
51:43And that's that's the coding acceleration which is the leading edge of the acceleration. But that gives you an example of what the productivity increase can be.
51:52How do we ensure that we don't replicate the issues we have here, like in current society, when it comes to like health access and availability? That like that you know, this supercharged AI for your health and longevity and all these you know, learning and all the you know, sort of things you lay out there are actually available to everyone regardless of income.
52:13Well, this is the kind of thing that I think is the more like as opposed to moratoriums or de-acceleration, this is the kind of thing that I think government should be involved in. Now, one good news is the natural thing about scale technologies that it wants to get to everyone. That was a little bit of like the Uber phone point, right? Like it's kind of like the hey, look, there aren't special phones for the rich.
52:36Everyone has the same phones.
52:38No matter how many zeros you have in your bank account.
52:41So, and by the way, I think there will be a lot of AI agents that will be the very same. Right? Uh that certainly chat GPT, Claude, etc.
52:50already seeing that if you pay this much a month, you get this access and that access. Yeah.
52:54And the key thing to pay attention to is say, okay, if there really ends up becoming a serious digital divide, let's make sure that, you know, for all the things that matter for, you know, kind of 80, 90, 95% of Americans, there's as broad-based access as possible.
53:12Which takes us to politics, which is something you're no stranger to, uh having you know, uh supported and donated to a number of causes through the years. I'm curious uh your assessment right now of what's going on in Washington as we head into midterms.
53:27So, look, I think the American people, it's a little bit like the Ronald Reagan line of think about yourself 2 years ago, do you feel you're better off or not? And they go, things are a lot more expensive, gas, beef, etc. Um like I was I was opposed to um like I wanted much more border control and immigration, but I didn't want to be having Americans shot in the streets.
53:50Like I think there's a broad wave of a lot of things have been bad, and I thought I was promised things and I didn't get them. So, I actually think we'll see a strongly, you know, Democratic House. We may even, even though the math's a little bit, see a Democratic Senate.
54:05Um and I think that's a referendum on how bad things have been for the last 2 years.
54:12you're donating to a number of Democrats right now in the midterms, right?
54:14And by the way, I'm always a pro-society, pro-business kind of person. It's not like a Democrat thing. It's a who are the people who are going, I like rule of law, I like, you know, democracy, I like, you know, kind of a fair system, and I like stability for business.
54:30What do you make then of the debate that we've seen unfolding in the Democratic Party, especially as you've seen the rise of some of the Democratic Socialist candidates and some folks on the on on the far left there, uh who feel that the existing approach Democrats have been taking has been insufficient?
54:44So, um I have some points of disagreement, right? And the points of disagreement are I think that uh to be a leader in any society, and especially American society, you have to be uh pro-business.
54:58Right? And it doesn't mean like business can do whatever it wants. They go, "Oh, you said business." No, no, no, of course not. And and whatever your theory is of how you get economic growth to create that prosperity. And I think most Americans are with that. Most Americans like the worry against the Democratic Socialists is you want to take hard money from me, well, I don't I mean me, Reed, but like my hard work I'm doing a bunch of hard work, and you're giving it to other people, and I need that because I'm working hard for this and everything else. Like the 60-plus percent of Americans.
55:25Right. And even some of these candidates, you know, have who come from a philosophy of like, we're going to centralize things and take it away from, you know, you the business owner.
55:32Yeah. And so, so I think most Americans want prosperous business, and that's the thing that the the DSA gets wrong, right? Like I think they're right that people are feeling pain. I think they're right to try to help people with that pain, but the answer is like if you ask most Americans is how do I get better economic opportunity? And of [clears throat] course how do I provide for health care for me and my family and other things, but that's the kind of thing that I think is the right thing for the American people.
56:00You've said the Democratic Party alienated parts of the valley going into 2024.
56:04And then we saw a number of the CEOs sitting the inauguration, right? At Trump's inauguration. Do you feel the party has fixed that relationship or is it the midst of like it there's more of an understanding between the two sides or not really?
56:15I think it's a work in progress. I think there have been small efforts to try to do it, but I'm like for example I I won't name the particular person I was talking to who was part of the Biden administration is like why did Silicon Valley turn against us? I was like well cuz you guys were doing a bunch of things that were very like saying tech big tech evil and so on.
56:36He's like well I have a problem with social media and children. I was like so does most of the valley, right? Like like most of the valley thinks social media and children is a real problem.
56:45If you were just constraining yourself to that but it was like no you were like all big tech bad. And that kind of differentiation like I've even had this like debate where people like with a number of Americans and Europeans they say well the same people who are building AI are the people who are building social networks. And you're like well actually that's not the case.
57:04That's not the case for open AI, not the case for Anthropic, not the case Google doesn't really have a social network, not the case for Microsoft bought LinkedIn. And by the way LinkedIn was fine it's a social network. So like it's kind of like the no no it's not big tech bad. You have to be a little bit more nuanced than that.
57:19Just as as we end up here I wanted to ask you know money in politics comes up a lot with our audience. You've spent you know major sums influencing Democratic politics.
57:29We're not trying to influence Trump cuz I never asked for anything.
57:33Okay. trying to help the right leaders who care about society get elected.
57:38Are you worried though, even sort of taking your hat off, right, about the influence that money has on our politics? And and I you know, I I wonder actually where you stand on Citizens United and and just the role as we look at midterms and now a 2028 race that, you know, could see tens of billions spent.
57:54Well, I've funded a couple of Larry Lessig projects trying to, you know, reduce the amount of money in I think it'd be better if we had less money in politics.
58:05Um but I think that if you say, "Hey, um uh people who don't care about call it your average middle-class Americans are going to spend a bunch of money to try to influence politics and you say, "We should be trying to help all of America be a better society." It's like I think it's worth spending the money for that.
58:25Right now, I don't like spending money on politics. I'd rather be spending money on building companies, minus AI, etc. But I do so because I think it's a responsibility to society. This is like society's been very good for me. You know, I come out of a middle-class background. I was able to uh be work hard but also be lucky and and get very fortunate. I should contribute back. That's part of my contributing back.
58:48If I'm a uh presidential candidate going uh in 2028 and I want Reid Hoffman's support, what are you looking for?
58:55Uh well, do you do you have a theory about how we can use American advantages in business and technology to make a better society for all Americans?
59:08And do I think your theory is right?
59:10So, you can't can't be a doomer or gloomer.
59:13Yeah, broadly I don't think that'll work.
59:16Yeah. It's a pleasure Yep, thank you.
