Eric Schmidt and Craig Mundie with David Rubenstein: Genesis

In the Ukraine-Russia war, Ukraine has no air force and no navy, forcing it to use AI technologies, drone technologies, and consumer technologies in novel ways to defend itself against a much larger military power; this has become a testbed for new forms of AI-enabled warfare where traditional military hierarchies are not available.

“the nature of the Ukraine war sorry Russia war against Ukraine technically is Ukraine does not have an Air Force and does not have a Navy so they were forced under this dire situation where they were being shot at and and literally murdered and little old ladies being bombed and things like this which is horrific to use AI Technologies drone Technologies consumer Technologies to fight a different kind of National Security”

Schools and educators cannot reliably detect when students use ChatGPT to answer exam questions, and while various detection methods are under development (such as embedding coded markers in AI-generated text), no reliable public-facing solution currently exists, meaning this is an ongoing systemic problem for education.

“how do teachers in school know that somebody who's a student isn't using chat gbt to answer an exam question well they don't they don't really know uh you can't figure out with artificial intelligence how to prevent that problem”

Google implemented novel management practices that were unusual for their era: 20% time (employees could spend 20% of work time on projects of their choice), no dress code (with minimal exception: 'you had to wear something'), free food (to keep employees from leaving the office), and a culture treating the company as a family; when employees began sleeping in their cubicles, Eric Schmidt pointed out this was illegal, requiring a policy clarification that employees could live on-site but not sleep there.

“One of the unusual things about the early days of Google was you had what's called 20% time you could spend 20% of your time doing whatever you wanted um you had no dress code as I recall you had we had one you had to wear something something and uh you had we had a problem and so we had to actually EMP change the ru a little bit and you had another novel concept which was you give free food to your employees so they won't go to a restaurant somewhere and waste time”

Whether AI should be used to decide investments is debatable, but in the next decade, human-computer collaboration will be essential because AI systems still get confused, and human experts will need to partner with AI to provide oversight and judgment; this human-in-the-loop model will remain the best approach short-term.

“the more the more serious answer right is that at least for the next decade we will be very important in the system because the system the computers get confused and so the human computer collaboration will be best so think of the next generation of physicists mathematicians chemists and so forth who are graduate students coming out with the the incredible contributions they're going to make every one of them will have the equivalent of one of these computers helping them to go back to your question serious way you know when when investment groups assemble teams of people to to consider investments in particular Realms you know they're really trying to combine the knowledge and even intuition of those people and this is where this polymathic nature of the machine you know really can Excel and so once these machines understand all the dimensions that go into considering an investment I think that they will be able to do a great job at that right”

Countries including Arab nations are aggressively competing for large data center investments by offering renewable energy, abundant sunshine, and capital; the US government faces a geopolitical question of whether to allow advanced chips (the most advanced AI hardware) to be used in non-US data centers, which is a strategic export control issue.

“people are now working very hard to I always answer think Canada friendly nice hydr power really smart very good in AI there's lots of countries that are now competing for that budget um and there are huge geopolitical issues so for example um the Arab countries are working very hard on these large data centers they have they claim many renewable data centers they obviously have a lot of sunshine they have certainly lots of money um and the question is does the US government allow them to take use the most advanced chips”

Tech companies are currently paying to restart mothballed traditional nuclear fission electrical systems because they forecast long-term power demand, recognizing that classic utility companies won't make necessary capital investments.

“the tech companies are actually paying to restart the moth bald uh traditional nuclear fion electrical system and that's just because they see down the road and say right you know the classic people aren't going to make the capital investment to get this done”

Google retains search data for approximately 12 months for public safety reasons, and it's reasonable to assume this data retention policy will be applied to all AI systems for the same justification.

“Google by policy and by law retains the searches for roughly 12 months for public safety reasons and my guess is that will be applied to all all of AI for the same reason”

Students using ChatGPT to answer exam questions cannot currently be detected through technological means, and while some research explores watermarking or coding material in answers, no publicly visible detection method exists yet.

“how do teachers in school know that somebody who's a student isn't using chat gbt to answer an exam question well they don't they don't really know uh you can't figure out with artificial intelligence how to prevent that problem”

Within the next few years, AI systems will become superhuman in two specific domains: mathematics and programming; this is because both domains are self-validating (code works or doesn't, proofs check or don't), allowing AI to generate, test, and iterate without needing external data or biological experimentation.

“there is there is there's this sort of core problem that we're working ourselves out of a job um because of the the planning the the Deep seek and other results there are people including myself who believe that within the next few years we will have emergent two categories of AIS that are superhuman the first is in math and the second one is in programming the technical reason is that math and programming can be self- validating you can sort of generate code and if it works great and if it doesn't you can discard it you don't need bi biology or physics or real data you can just do it in the spin Loop same thing for math”

Microsoft stalled under Steve Ballmer's leadership as CEO, with the stock price not advancing significantly for roughly 10 years; the stagnation was driven by the company's inability to pivot into new growth markets (search, mobile/iPhone) despite having research and development efforts in these areas, causing investors and the industry to view Microsoft as a legacy personal-computer company with limited future growth.

“Bill Gates became famous and so forth maybe the richest man in the world for a while and then the company stalled and under Steve balmer's leadership as CEO I think the stock didn't go anywhere for roughly 10 years”

Google's monetization strategy came from Salar Kamangir (a Stanford biologist who got bored with biology and was about 23 years old at the time) and Susan Wojcicki, who together wrote the initial business plan that identified advertising as the revenue model for Google's search engine; this was not obvious to Larry Page and Sergey Brin initially.

“so this fellow named solar kingar who was all of looked to me like about 15 but it was really really like 23 um was a Stanford biologist who'd gotten board with Biology and he'd written the initial business plan along with Susan wiii who just recently uh passed away unfortunately so Susan and Soler who were sort of uh P put together early um wrote the business plan and they didn't understand how the advertising would play out”

Pentagon Regulation 3.09 ('man in the loop' or 'meaningful human control') requires that any US military weapon system must remain under human control to maintain human accountability and compliance with international laws of war; the key challenge is achieving AI automation benefits while preserving meaningful human responsibility.

“the Pentagon has a rule called 3.09 which is called man man in the loop or meaningful human control which means that any weapon system has to be under human control in the military sense in order to hold them responsible for the decision so the key thing in military conflict and National Security is how do you get the automation that you can get with AI and with these robotic defense systems while still having humans be responsible and respecting the international laws of war okay”

Google's competitive advantage came from two innovations: (1) PageRank, an algorithm invented by Larry Page that provided more accurate search rankings than competitors, and (2) a Dutch auction system for advertising that was more efficient than standard auction formats; these two innovations together created enormous cash flow that enabled everything else Google achieved.

“the technical answer was that Larry had invented a particular algorithm which happened to be called page Rank and so the rankings were more more accurate and at the same time I was joining they invented a kind of auction called a Dutch auction which meant that the auction was the most efficient it could be and those two together propelled the growth of the company and created this enormous cash flow which then enabled everything else”

During Microsoft's flat growth period, the iPhone and mobile revolution disrupted the company's personal computer dominance, and Microsoft was late to market with its own phone offerings, contributing to the perception that the company had peaked.

“during the flat period a number of things that happened in the industry emergence of the cell phone and the iPhone other things which Microsoft had worked on but basically was late to bring you know its own effort to Market to the marketplace and so you know that put people in a mindset that said well Microsoft you know had its Heyday with the personal computer”

America needs to be the leader in autonomous robotic warfare to maintain national security, and autonomous AI-controlled robots will be fundamental to future national security discussions.

“my personal view is that America needs to be the leader in that area and that all of our discussions about National Security fundamentally will require autonomy technically autonomy of robots that will be AI controlled”

AI technology will become more user-friendly and accessible through natural language interfaces, dramatically reducing the frustration that billions of people currently experience trying to master arcane interfaces of complex computers.

“these machines now that you can just talk to and talk with at a quality level that's really akin to talking to another smart person really means that that this power is more readily available and the frustrations that you and other billions of people have had over the years trying to master the Arcane aspects of dealing with computers is going to get dramatically simpler”

The 'eye of the needle' problem describes how in game theory, if competitive players are almost at parity and one shortcuts testing/safety to gain advantage, this creates asymmetric instability—all players are incentivized to cut corners even if collectively worse for everyone, requiring international coordination to prevent.

“there's actually a a framing in our industry now which is called the eye of the needle problem and it goes roughly like this assume the three of us are friendly competitors but we're pretty ruthless um and we're very close to getting something done but the tradition in our group is that we test it before we release it but let's say the gain is really big and I being Mr evil and you guys being much less evil than me decide to cut corners and not do the testing that is inherently destabilizing in a game Theory way”

The pace of AI capability advancement is extraordinarily rapid—major breakthroughs are occurring months apart (DeepSeek only one month after o1/o3 release).

“it was only a month earlier that's how fast this is happening”

In the last four months, three major AI breakthroughs have emerged: Google's Notebook LM (which converts documents into natural human-sounding audio), OpenAI's O1 and O3 models (which show planning and decision-making transparency), and DeepSeek from China (which is open-source, shows algorithms, achieved training at 1/12th the cost, and has serving costs roughly 1/100th of OpenAI, making it surprisingly competitive despite being only a month after O1).

“for me there have been three in the last show you how fast three in the last um four months the first was a product from Google called notebook LM and if you haven't played with it um just enter into Google the search the search query notebook Alm not human and listen to it you'll see what I why why it was so profound the next one was a product from open AI called 01 followed by o03 they skipped O2 because there's a Stadium named after O2 and what it does is it shows you planning it can actually show you how it's making decisions and then the big event that Craig mentioned was the release of this thing called Deep seek and deep seek has royed the industry Because deep seek uh is first open source which means you can read it they published their algorithms which the other guys have not they claim to have done the training with 12th of the cost and their serving cost is roughly a hundredth and that's Chinese and that's from China right”

When Google initially had the PageRank algorithm, the founders didn't know how to monetize it; Salar Kamangir (a Stanford biologist bored with biology, age ~23) and Susan Wojcicki wrote the initial business plan.

“when they came up with the idea and they vented the page rank system they didn't know how to monetize it I thought at the beginning so who came up with the idea of using advertising as the way of monetizing it so so this fellow named solar kingar who was all of looked to me like about 15 but it was really really like 23 um was a Stanford biologist who'd gotten board with Biology and he'd written the initial business plan along with Susan wiii”

When technology stocks rise dramatically, historically they tend to reverse at some point, suggesting that as a general rule of thumb, extremely high valuations are vulnerable to correction, though this does not necessarily apply to Nvidia in the near term because it continues to produce extraordinarily powerful supercomputers and has very strong market positioning with the Blackwell chip group.

“as a general rule of thumb when something goes up this much in the stock market at some point it goes the other way”

AI systems are being deployed in investment decision-making, where algorithms assess multiple investment factors and recommend go/no-go decisions; the speaker's firm uses OpenAI algorithms to screen potential investments.

“I'm working with open I open AI my firm has been doing it to assess whether a good investment is about to be made or not you you kind of take all the factors that you might consider an investment put put him into some open AI kind of uh algorithm and it tells you yes is a good investment not a good investment”

AI will likely be the future of investment decision-making, replacing human dealmakers, though dealmakers will get to 'play more golf' as their role becomes less critical.

“you think that's likely to be the future when making Investments people will use um artificial intelligence to decide whether it's a good deal or not yes and what about humans then what about our dealmakers what we're going to be just they get to play more golf yeah more golf right”

Google, Microsoft, and Meta are the first companies able to fully leverage AI power because of their scale and technical depth, while Apple is notably absent from the leading AI companies.

“in the case of we we obviously have conflict of interest involving Microsoft and Google here but the fact of the matters they're the first of the companies to be able to fully use this power because they're simply so large and they have such good techy”

Sergey Brin and Larry Page, Google's founders, designed the company culture based on a family model where they wanted employees to work all the time and be treated like family members in a shared household.

“there was a famous event before before I joined where the Venture capitalists were so mad at Larry and serer they wanted their money back which would have been the largest error in venture capitalist history thank God it didn't occur so they had an unusual View and and to to to give Larry and Sergey credit their view was that a company was a family and we want they wanted everyone to work all the time and so if you basically treated them like a house and a family we had this problem where people would sleep in their cubicles”

Data center electricity consumption in the US has doubled from roughly 2% to 4% of total US consumption in the last five years, and is expected to double again in the next several years; current modeling suggests that the US will run out of all available electrical sources by roughly 2028 without major changes in generation capacity.

“in the last five years the electricity percentage of data centers has doubled from roughly 2% to roughly 4% of us consumption it is expected to double again in the next however many years current modeling says we run out of all electricity sources in the US by roughly 2028”

The term 'artificial intelligence' is a historical artifact from the 1950s when researchers coined it to mean 'fake' intelligence, but modern AI is not artificial or fake—it is enhanced human intelligence, and soon it will take its own evolutionary path distinct from biological neural networks.

“I think it's just a hangover from when they coined the term back in the 50s and that at that point they said okay we got intelligence and then sort of this fake intelligence so they called it artificial but I don't think there was anything profound in that so but like it's not really artificial it's enhanced human intelligence really it's not artificial”

In competitive game theory, there is an 'eye of the needle' problem: when three ruthless but friendly competitors are close to achieving something valuable, the traditional norm in the community is to test before release; but if the potential gain is very large, one player may decide to cut corners and skip testing, which is inherently destabilizing because it incentivizes others to abandon norms as well; this dynamic applies to AI development across nations.

“there's actually a a framing in our industry now which is called the eye of the needle problem and it goes roughly like this assume the three of us are friendly competitors but we're pretty ruthless um and we're very close to getting something done but the tradition in our group is that we test it before we release it but let's say the gain is really big and I being Mr evil and you guys being much less evil than me decide to cut corners and not do the testing that is inherently destabilizing in a game Theory way and we talk about this in the book”

Kissinger's theory of societal response to AI is that when societies encounter something they don't understand, they either treat it as a new religion (worship/reverence) or rebel against it with hostility (pitchforks)—there is no rational middle ground without deliberate governance.

“his view for example was that um societies when they see something they don't understand will either treat it as a new religion or rebel against it take take pitchforks against it if you will”

Kissinger's thinking about AI was analogous to his Cold War nuclear deterrence insights—strategic competition doesn't require elimination, but can coexist with agreements that prevent the worst outcomes and benefit both sides, avoiding the need for either side to surrender its position.

“one of the things that Henry certainly taught me and I I think Eric and others was lessons that were learned back you know in the in the cold what became the cold war between the U us and the Soviets around nuclear weapons and you know Henry believed that it was very important that that they eventually came to the realization that the Soviets and the Americans were not any longer likely to shoot at each other um but but if they allowed this stuff to continue to fuse very rapidly then somebody else might initiate this who didn't have the restraints so he thought about AI in a similar way that uh strategic power competition you can't completely eliminate it but even when you have it and you we still have it between this the now Russians and and the west but uh even when you have strategic power competition you may find common ground you know and to do something that you know benefits both of the players but doesn't require them to sort of lay down you know their position”

Henry Kissinger's biggest fear about AI was that disruption would occur faster than institutions could adapt, leading to chaotic conditions; he was also deeply concerned about what it means to be human when superintelligent AI systems are operational—specifically, that if machines restrict human freedom, humans have no one to hold responsible, which drives people to resist machine authority more intensely than human-imposed restrictions.

“what was Dr Kissinger's biggest fear about AI I'd say Henry's fear about biggest fear about AI is how the disruption will occur uh more quickly than our institutions are able to adjust and therefore you know the potential for a chaotic uh environment will be very high and that was his biggest concern he he also is very very concerned concerned about what it meant to be human when these systems are up and running so he would go on and on about what does it mean when I deal with this system we use military examples but there are so many right and one of the basic lessons is that humans don't like to lose their freedom to a machine if a human prevents their freedom then at least they have somebody to Target right but if the machine is big and know big and unfeeling and un it's just drives people crazy”

When societies encounter technology they do not understand, they typically respond in one of two ways: either they treat it as a new religion, or they rebel against it violently (with pitchforks), rather than approaching it rationally; this pattern is relevant to how societies will respond to superintelligent AI.

“his view for example was that um societies when they see something they don't understand will either treat it as a new religion or rebel against it take take pitchforks against it if you will”

Kissinger's core fear about AI was not existential risk per se, but that disruption will occur faster than institutions can adjust, creating a chaotic environment, and secondarily, he feared losing human freedom to a machine—humans can target another human oppressor, but an impersonal machine creates psychological helplessness.

“Henry's fear about biggest fear about AI is how the disruption will occur uh more quickly than our institutions are able to adjust and therefore you know the potential for a chaotic uh environment will be very high and that was his biggest concern he he also is very very concerned concerned about what it meant to be human when these systems are up and running”

Kissinger finished the AI book roughly two weeks before he died, indicating the project's importance to his legacy.

“he finished the book roughly two weeks before he died and that's how important this book was to him to leave in a legacy of what does it look like to and how how do we preserve National Security Humanity”

The current Blackwell generation of Nvidia chips is so powerful that the entire year's supply is already sold out to the four major hyperscalers (large cloud/AI companies), demonstrating strong near-term demand and market position.

“Nvidia continues to produce extraordinarily power powerful supercomputers the what is basically called the Blackwell group is so powerful that the entire supply for the next year is sold out to four companies which are the big hyperscalers so that's a strong position to be in at the moment”

The book's main message has two parts: (1) Focus attention on the upside potential of AI for humanity, not just downside risks and existential threats, (2) Achieve this by developing an architecture for 'alignment' or 'safety and control' using AI itself—'fighting fire with fire' by using AI systems to control other AI systems, rather than treating humans as passive victims who must be protected from AI.

“I would have said that there are two things that we tried to do in the book one was to get people to focus a little bit more on the upside there's been a huge amount of coverage of the the downside risk the existential threats etc etc but the impact of this on Humanity on the upside should be unbelievable and Eric mentioned that a little earlier the key in in getting people to to set aside their concerns is you have to have some architecture for what they call alignment or safety and control and you know we've started to think about and now we start to see other people both in the academic world and even more recently the open AI announcement where people are starting to realize that they're may be a way to deal with this where you use the AI to control the AI so this is sort of fighting fire with fire instead of thinking we're victims of this you know humans have to get engaged and figure out what is the long-term relationship we want”

Eric Schmidt transformed Google's culture by implementing a 'growth mindset' that prioritized competition with external rivals rather than internal competition, recognizing that Microsoft's past success had left the company culturally locked into a competitive dynamic with itself rather than the market.

“he recognized that Microsoft had a culture that had evolved over some decades uh but it kind of got locked in at the point where Microsoft was was King Of The Hill in the personal Computing business and the company often times had as much internal competition as it did externally focused competition and he he thought that was a problem...his first thing he did was to change to set about to change the culture of the company to have more of a what he call a growth mindset and to really recognize that the the competition was not itself but with other people”

Drones represent the future of asymmetric warfare, separating the human from the attacker so humans can direct operations remotely (drinking coffee) while robots execute—but the integration of AI-controlled autonomous warfare raises unprecedented questions about human responsibility and the nature of conflict.

“let's think about the conception of human conflict which has unfortunately been with us for ever 50,000 50,000 years or whatever it's fundamentally you stereotypes man against man gun against gun what have you and we have many many rituals in our society that are related to this but what if you separate out the human and the attacker right what if the human is sitting watching you know drinking coffee where the robot in this case a drone but it's really robotic war is is running it we don't know what that conflict looks like”

Craig and Kissinger organized a 'track two' dialogue with Chinese officials to discuss AI governance before tragedy strikes, but found stark differences in approach—the Chinese team was highly formal (9 officials in suits, reading prepared statements) while the American side was casual, revealing how unprepared diplomacy is for AI-era governance.

“Dr gissinger Henry and US spent an awful lot of time discussing how would you get our opponent in this case is the Chinese to the table to discuss this before a tragedy indeed Craig organized this track two dialogue with the Chinese so he invites me to this thing so he's running it and it's you know it's Americans you know we're on zoom and we're kind of our normal outfits not dressed like this and on the other side are nine Chinese people in their suits speaking in order from their prepared and I thought oh my God right how are we going to do this so we just don't know how to deal with diplomacy in the age of AI okay”

The Stargate infrastructure project aims to build data centers not just for computational capacity but also to solve the energy problem, and will likely require solutions including nuclear power restart and regulatory changes to make feasible within a 5-10 year timeframe.

“what they're really trying to do is recognize that uh that no matter how the training evolves at the end of the day you're going to need an immense amount of infrastructure to deploy all this intelligence and uh and that infrastructure is not just the the chips and the data centers it's also going to end up probably being the power”

Artificial General Intelligence (AGI) will emerge within 6-8 years, defined as a single computer system reaching the 90th percentile or better of peak human ability across all domains simultaneously—mathematics, physics, chemistry, writing, music, and liberal arts—creating a universal polymath AI.

“most of the people I talk to think you know it's somewhere I'll say three to five years”

The US is marginally ahead of China in AI, but different parts of the AI stack show different competitive positions: the US maintains hardware advantage, while China has aggressively developed new algorithms that partially offset superior US hardware; recent Chinese algorithmic announcements show convergence.

“I think in the aggregate you know we we believe we maybe just slightly ahead as Eric said there's different parts to this stack there's the the hardware that you need today the best hardware is still available in the United States more than it is in China but the Chinese have been very aggressive in developing new algorithms uh that tend to take away some of the advantage of the Superior Hardware and so it you know and there was some recent announcements from the Chinese efforts just in the last couple of weeks that have had surprisingly positive uh results and people have said wow you know they're basically pulling even even without the access to the hardware”

The book argues that the arrival of AI is not merely another tech story, corporate IPO, or wealth transfer, but a foundational change in human civilization that will be discussed and analyzed for thousands of years.

“the arrival of this intelligence and the platforms that is implied is more than another tech story it's more than another IPO it's more than us all making lots of money and then magnificent 7 it is a foundational change in human life which people will talk about for Thousand Years”

Training costs for foundation models like GPT-5 and Gemini 2/3 are expected to exceed $1 billion just to develop the base model, and largest current data centers use about 200,000 chips; at this trajectory of spending, $500 billion would be consumed relatively quickly.

“the largest data centers now are about 200,000 of these chips the next iteration of foundation models which is think of it as gp5 and Gemini two and three and so forth look like they're going to cost about a billion Dollar in training cost just to get the base model that's how expensive these things are it doesn't take very long before you can start using up $500 billion”

For at least the next decade, human-computer collaboration will be superior to either humans or AI alone, because computers still make mistakes and humans are essential for disambiguation and correction.

“at least for the next decade we will be very important in the system because the system the computers get confused and so the human computer collaboration will be best”

NVIDIA's Blackwell GPU group is so powerful that all supply for the next year is completely sold out to only four companies—the major cloud hyperscalers—putting NVIDIA in an extraordinarily strong market position despite being a capital equipment supplier.

“the what is basically called the Blackwell group is so powerful that the entire supply for the next year is sold out to four companies which are the big hyperscalers so that's a strong position to be in at the moment”

The arrival of an artificial general intelligence system comparable to alien intelligence is a foundational civilizational event that will reprogram every single financial system, marketing system, and business system; therefore, investors should invest in everything—both platform and application companies—because the transformation is so comprehensive that diversification across the AI ecosystem is warranted.

“can I answer your question by saying yes yes you should invest in everything and here's why the arrival of an alien intelligence in humanity is a very very big deal right that's what our book is about right right it's fundamentally about how do we take advantage of this new intelligence while we're preserving human dignity so every single Financial system Marketing System business business system will be reprogrammed”

For students entering the AI field: a general STEM background is valuable, and they should pursue comprehensive technical literacy across multiple domains while maintaining adaptability.

“first I think a general background in this in the stem field is going to be super valuable in the future even though the AI will be a tremendous tool and even partner in in a tackling problems of All Sorts you know the the humans are going to basically for quite some time have to be involved in that process and adaptability in a sense people used to tell me oh you know you guys you studied electrical engineering or Computing you know you should have gone and taken a had a liberal arts education so that you know you could you know think more broadly and my view is right now people need to think broadly but they need to be well-rounded uh in these technical fields too”

The book's authors and Henry Kissinger spent significant time discussing how to get China (the opponent in the strategic competition) to agree on international dialogue about AI control before tragedy occurs, drawing lessons from Cold War nuclear deterrence frameworks; Craig organized a 'track two' dialogue with Chinese counterparts but found fundamental differences in negotiating style that make AI-age diplomacy unclear.

“so the key thing is we're we're getting systems that are so powerful we really have to rethink the way nation states work and competition works and Dr gissinger Henry and US spent an awful lot of time discussing how would you get our opponent in this case is the Chinese to the table to discuss this before a tragedy indeed Craig organized this track two dialogue with the Chinese”

The upside potential of AI for humanity—unlocking vast improvements in human capability, health, prosperity, etc.—has been undersold relative to downside risk coverage, and the book aims to rebalance attention toward positive impacts.

“one was to get people to focus a little bit more on the upside there's been a huge amount of coverage of the the downside risk the existential threats etc etc but the impact of this on Humanity on the upside should be unbelievable”

Microsoft under Steve Balmer's leadership experienced a roughly 10-year period of stagnation where the stock price didn't advance, before recovering under new CEO Satya Nadella.

“Bill Gates became famous and so forth maybe the richest man in the world for a while and then the company stalled and under Steve balmer's leadership as CEO I think the stock didn't go anywhere for roughly 10 years and then Balmer left and and uh um a new person came in and uh stock is up1 times”

Eric Schmidt served on the Apple board and observed that Steve Jobs combined exceptional intellectual capability with creative and aesthetic sensibility that distinguished him from other tech leaders like Bill Gates; Steve Jobs was the first person Eric encountered in the tech community who demonstrated genuine style and taste.

“you were on the board of Apple for a while and you presumably you knew Steve Jobs very well and you know Bill Gates so who's smarter uh they're both brilliant in different ways um I think I think Craig can talk about Bill's Bill's Brilliance which is fantastic Steve had not only a truly brilliant mind but he also had a creative mind that was combined he was the first person I encountered in the Tech Community who had any kind of style and taste at all”

Separating the human from the attacker in warfare (e.g., drone operators drinking coffee while robots conduct warfare) represents a fundamental shift in military conflict; the nature of this new form of conflict is unknown and potentially dangerous, but Eric Schmidt believes America must lead in autonomous robotic warfare technology to maintain security advantage.

“what if you separate out the human and the attacker right what if the human is sitting watching you know drinking coffee where the robot in this case a drone but it's really robotic war is is running it we don't know what that conflict looks like my personal view is that America needs to be the leader in that area”

AGI must be in America under American control for national security and must be guarded with guns, and unlike humans, AGI can cross-apply knowledge across all scientific and human endeavors simultaneously, creating unprecedented intellectual capability.

“first it better be in America right it better be under our nation's control and the second thing is that unlike humans it can cross knowledge it's the ultimate Leonard D Vinci who's a polymath across everything so imagine a single computer system which presumably would be guarded with lots of guns right they could actually see across every scientific every human endeavor every idea of thought and everything in history in one machine that's the goal”

Eric Schmidt got involved in defense and national security issues through a '20% time' mindset applied to his role as chairman, working with Ash Carter (a physicist who ran the Defense Department) to set up the Defense Innovation Board; he spent extensive time at military bases and found that the level of change resistance in the Pentagon is profound and has been established for 50 years.

“at Google you get 20% time and I decided that applied to to me as chairman as well and I got interested in in National Security and I worked with this guy Ash Carter who who passed away who's a brilliant physicist and ran the defense department and I set up this thing called the defense Innovation board and spending I spent lots of time with the military went to a hundred different bases gave lots and lots of speeches and it it's almost the level of change resistance is profound if you want to understand how bureaucracies behave visit the Pentagon right and this is this is not a new fact this has been established for 50 years”

The singularity is defined as the point where machine intelligence exceeds human intelligence, but more importantly, it is the point where the machine is learning faster than humans can learn, making the divergence irreversible; at that point, the prudent response is to 'unplug them.'

“what is singularity what is that mean know various definitions but I think most people think that the singularity happens when the machine intelligence exceeds human intelligence and more importantly uh because because humans are learning and computers are learning The Singularity is the point where this where the computer is learning faster than the humans can and my own view is at that point unplug them”

The book's singular contribution focused on polymaths in history and their profound impact, with research into both Indian and Russian polymaths, arguing that AGI will create 'a million polymaths' as everyone gains access to polymathic capability.

“actually we we should mention this because this was your your singular contribution the the poly you know we celebrate these polyas but their impact is profound think about Einstein and just go to the list and we also you did the research on the Indian crazy polyas and the Russian crazy polyas they all they we've had them now in theory there won't be one there will be a million of them because everyone of us will have access to that that has huge implications on society”

We are entering a period similar to the early days of personal computers when word processors and spreadsheets emerged; before that, you had to program to use a computer, but these tools made computing accessible to non-programmers; similarly, AI systems that you can simply talk to at near-human conversation quality will dramatically reduce frustration and make AI power accessible to billions of people, and will simplify tasks that have required mastering arcane computer interfaces.

“there is there's this sort of there is there's this sort of core problem that we're working ourselves out of a job um because of the the planning the the Deep seek and other results there are people including myself who believe that within the next few years we will have emergent two categories of AIS that are superhuman the first is in math and the second one is in programming the technical reason is that math and programming can be self- validating you can sort of generate code and if it works great and if it doesn't you can discard it you don't need bi biology or physics or real data you can just do it in the spin Loop same thing for math so we may be ultimately working ourselves out of a job now in the mathematics World there was a word long time ago called Google which meant infinite numbers 10 to 100th right but it was spelled g o o g e l no g o o g o l oh okay a Google they ultimately the the fellow the mathematician evented has some children who ultimately sued us over the Miss why why didn't you spell it that way or you just this was this was Larry and Serge's decision they thought it was more fun as Le I see okay and um today um would you say that um if people read your book they will be a better human being oh that's a layup by the way question you have to think about that you have to think about that the correct answer is yes so what is your next book going to be on I don't know that I'll write one this was my first one was Henry's 22nd what was your fifth fth yeah and uh I'm not sure I'll write another one really you sometimes people write their first book and like Margaret Mitchell wrote go on with a wind she wrote another one because it was just got writers block afterward so you have writers block do you think after this oh I hate to write okay all right it was it was a real effort for me to write this so um Donald Trump has imp has some Tech U people working for him I guess now you may know some of them um is his Tech policy going to be a lot different than the Biden Tech policy would you say he appears to have can or in the process of canceling all of the Trump all the Biden orders um and my own opinion is that they will reemerge in a China strategy because the issues around NASA security are not tied to any particular Administration um the Trump Administration followed by the Biden Administration did a very good job in terms of Hardware controls export controls but not enough the second Trump Administration will have to deal with this question of Chinese thing my own view is there you know president Trump is apparently a dealmaker there's a deal to be made and here's the deal the big companies are desperate for energy and permitting access and so forth and so on they have plenty of money plenty of capital plenty of investor support the government is completely ignorant about the things that we're talking about and it's horrible and they they want to learn but they don't have the software people they don't have the the young people they don't have the access and so a deal where the Trump Administration manages to get voluntary access to National Security from these companies in return for voluntary access to huge energy that's a good deal for everybody okay um it says here I've been seeing more and more about is it AGI or ag2 AGI do what ai ai ai what is how close are we to it being here [44:59] what is it artificial general intelligence oh General okay yeah I mean my view is you know this stuff went through a period where we started with machine learning that was trying to train the machines to kind of do one thing that wasn't sexy enough so they rebranded it at Ai and that confused people and then the real AI starts to emerge and they have to distinguish it so they add the G so so I'll give you a I'll give you a precise definition the a whole bunch of people did a survey on this and they said that between 6 to 8 years there will be a single computer will be at the 90th percentile or better of the best human ability literally the best mathematician alive the best physicist alive the best chemist alive so forth the best writer the best uh liberal arts person the best musician now what happens when that thing shows up well first it better be in America right it better be under our nation's control and the second thing is that unlike humans it can cross knowledge it's the ultimate Leonard D Vinci who's a polymath across everything so imagine a single computer system which presumably would be guarded with lots of guns right they could actually see across every scientific every human endeavor every idea of thought and everything in history in one machine that's the goal and that's why we're spending so much time and money on all of those now what first part of the book actually talks about polymaths and their role in history and why this will be the ultimate polyat yeah actually we we should mention this because this was your your singular contribution the the poly you know we celebrate these polyas but their impact is profound think about Einstein and just go to the list and we also you did the research on the Indian crazy polyas and the Russian crazy polyas they all they we've had them now in theory there won't be one there will be a million of them because everyone of us will have access to that that has huge implications on society what what is singularity what is that mean know various definitions but I think most people think that the singularity happens when the machine intelligence exceeds human intelligence and more importantly uh because because humans are learning and computers are learning The Singularity is the point where this where the computer is learning faster than the humans can and my own view is at that point unplug them so when you're working at Microsoft you're like sometimes I I'm using a Microsoft program Windows or something and it doesn't go what I'm do what I want it to do and I get frustrated do you ever have that if you're work at Microsoft it doesn't work when you go down the hall and ask Bill Gates uh how do work does he ever come in piix your computer or anything don't ask Bill too often but more more likely many people I know I end up having to fix their computers because I still can make it work but uh but the great news about these things is I think we're entering a period that's kind of like the early days of the personal computer when we had word processors and spreadsheets before that you had to program to get a computer to do same you want and suddenly you could type or you could put in formulas it would do it these machines now that you can just talk to and talk with at a quality level that's really akin to talking to another smart person really means that that this power is more readily available and the frustrations that you and other billions of people have had over the years trying to master the Arcane aspects of dealing with computers is going to get dramatically simpler”

AI will likely be used to improve sports analytics (e.g., deciding 2-point vs. 1-point after-touchdown strategies in football), but this is not the highest and best use of AI; AI should focus on fundamentally digital/computational domains where humans cannot think deeply enough or remember enough, not on optimizing pre-existing human-driven sports.

“well once the NFL changes the rules so so one of the questions is do you think that automo let's use Formula 1 do you think Formula 1 would be interesting interesting if the computer if the cars were all uh electronic Vehicles there were no humans in them of course not so I think human sport will always fundamentally involve the the incredible athletic prowess and failures of such people but the the structure and the guidance so for example in today's game there was a question of whether you do um a two-pointer or one-pointer after a touchdown and the analytics now say on balance you should do the two-pointer but most people think that's too risky so for me the question is can you build an AI system that can more accurately model the probabilities in these games and the answer is yes but I don't think that's the highest and best use of AI I think the highest and best use of AI is to do things which are digital in nature where humans can't think that deeply you don't have you can't remember a million things at once”

Henry Kissinger was introduced to AI shortly before age 95 at a meeting in Europe, immediately recognized the profound implications of superintelligent AI exceeding human intelligence, and spent the last five years of his life studying not the technical aspects but the civilizational ramifications.

“we introduced Henry to this when he was just before age 95 and we were at a meeting in Europe and we were presenting this stuff to a group of people and Eric said to Henry hey you know you should go listen to this talk and he did and Henry was immediately you know uh impressed with what the implications were of the arrival of an intelligence that exceeded human intelligence”

Eric Schmidt spent significant time advising Ukraine's defense efforts, designing drones and using AI/consumer technologies to fight an asymmetric war against Russia, believing that democracy and freedom are worth fighting for and must be defended through AI and technological innovation.

“at Google you get 20% time and I decided that applied to to me as chairman as well and I got interested in in National Security and I worked with this guy Ash Carter who who passed away who's a brilliant physicist and ran the defense department and I set up this thing called the defense Innovation board and spending I spent lots of time with the military went to a hundred different bases”

The book's title 'Genesis' was chosen by the authors to honor Kissinger's preference for one-word book titles and reflects his belief that humanity is entering a new epoch—a new Enlightenment and Renaissance in the age of AI.

“we thought about you know the things he was worried about which was the new beginning for Humanity and a one-word title in picture there it is Henry believe very strongly that we are in a new epic of humanity that this is the beginning of it it we're privileged to be alive today because this is the beginning of the Renaissance this is the beginning of the Enlightenment in a new form”

A week before Trump's inauguration, the Biden administration issued complex restrictions on advanced chip exports; most observers expect Trump to reverse these restrictions.

“a week before President Trump was inaugurated the pridon administration issued restrictions on that which were very complicated but basically said that if you had a big us partner you could order up to 300,000 chips most people that I've spoken with believe that Trump will reverse that order”

Google's founders Sergey Brin and Larry Page had distinctive personalities: both had strong egos as tech founders, they questioned Schmidt intensely about his prior product work (showing intellectual rigor), and this deep engagement attracted Schmidt to join as CEO despite being at a successful company (Novell).

“well first place you know they're Tech Founders so they have a lot of ego and um so so I show up and they have my bio up and they spend their time questioning me about a particular product I was building in my previous company which they thought was a terrible idea and I realized at the time I had not had that good of a conversation you know an intense conversation in years and frankly that's why I said yes”

One company Eric Schmidt funded is building agents that write other agents, creating potential multiplicative growth as agents increasingly self-replicate.

“I funded one company which is actually doing agents to write more agents so to the degree that they can multiply that becomes a real question”

Most people Eric talks to believe AGI (artificial general intelligence) will arrive between 3 and 5 years from now; AGI is defined as a single computer system that performs at the 90th percentile or better across all human domains simultaneously (mathematics, physics, chemistry, writing, liberal arts, music), making it the ultimate polymath.

“most of the people I talk to think you know it's somewhere I'll say three to five years what is it artificial general intelligence oh General okay yeah I mean my view is you know this stuff went through a period where we started with machine learning that was trying to train the machines to kind of do one thing that wasn't sexy enough so they rebranded it at Ai and that confused people and then the real AI starts to emerge and they have to distinguish it so they add the G so so I'll give you a I'll give you a precise definition the a whole bunch of people did a survey on this and they said that between 6 to 8 years there will be a single computer will be at the 90th percentile or better of the best human ability literally the best mathematician alive the best physicist alive the best chemist alive so forth the best writer the best uh liberal arts person the best musician”

President Trump announced the Stargate project involving $500 billion in private sector investment in AI data center infrastructure in Texas and other locations.

“the president of the United States recently announced at the White House something called star I think it's Stargate Stargate and we're supposedly $500 billion from the private sector is going to be invested I guess in a facility in Texas maybe that will do multiple places multiple places which will do what”

A general background in STEM fields will be super valuable in the future, even though AI will be a tremendous tool and partner; for college-bound students interested in AI, the practical skill is to become expert in Python programming—learning Python well enables mastery of the broader AI technology stack.

“first I think a general background in this in the stem field is going to be super valuable in the future even though the AI will be a tremendous tool and even partner in in a tackling problems of All Sorts you know the the humans are going to basically for quite some time have to be involved in that process and adaptability in a sense people used to tell me oh you know you guys you studied electrical engineering or Computing you know you should have gone and taken a had a liberal arts education so that you know you could you know think more broadly and my view is right now people need to think broadly but they need to be well-rounded uh in these technical fields too”

Eric has funded a company developing agents that write more agents, which could create multiplicative scaling effects if successful—as the AI system can generate additional AI systems with minimal human intervention, the productivity gain compounds.

“I funded one company which is actually doing agents to write more agents so to the degree that they can multiply that becomes a real question”

When Eric Schmidt returned from a week-long absence, he found a junior employee occupying his office; rather than asserting authority and demanding the employee leave (a typical executive move), Eric adopted a listening posture to understand Google's culture first, and found that having this employee as an officemate was actually advantageous because the employee monitored and analyzed sales revenue, making him a valuable real-time data resource for managing the sales team.

“when you're a new executive you don't want to create a a a cultural faux paw so my advice to any of you when you get a new job is keep your mouth shut and listen and try to figure out how the culture actually works unless you create some error error so I'm in that mode so this Fell's in my office I said what are you doing here he said well your office was empty and I said well okay I said you're never here and I was in a five person office so I thought to myself what to do now the normal person would say get the hell out of my office exuse my language but that would have been a cultural fuxa so I let him sit there”

For an average investor who does not have deep AI expertise, the better investment opportunity is not in hardware and platform companies (like Nvidia, Microsoft, Alphabet, Meta) which are already highly valued with limited upside, but rather in AI-driven applications built on top of these platforms, which represent a larger and more fertile ground for substantial returns.

“I would have said that the hardware and the basic platform which is the big players that have made the the ones you know you know they're already very highly valued and you know you may not have a huge amount of Headroom left in there okay on the other hand the applications that are going to be built on this are ultimately going to be the thing that changes everything about everything and I think you know there's just going to be a huge array of companies that are going to build these AI driven applications and I would say there's probably more fertile ground for big winds there”

Eric Schmidt left Novell (a successful company) to become CEO of Google when it was a startup because the two founders (Larry Page and Sergey Brin) gave him the most intellectually intense conversation he had experienced in years, which signaled to him that working with exceptionally talented people was worth more than staying at an established firm.

“well first place you know they're Tech Founders so they have a lot of ego and um so so I show up and they have my bio up and they spend their time questioning me about a particular product I was building in my previous company which they thought was a terrible idea and I realized at the time I had not had that good of a conversation you know an intense conversation in years and frankly that's why I said yes”

If starting an AI company today, the largest revenue opportunity in the next few years will be in business process automation, specifically using 'agent' technology; an agent can be understood as a software system that takes language input and produces language output to execute complex workflows (e.g., house search, rule interpretation, design, contracting, construction oversight, contractor oversight, dispute resolution, and payment), effectively automating every workflow in every company.

“if you were starting an AI company what space would it be and and why the vast majority of Revenue in the next few years you were asking about which companies is going to be in what is called business process Automation and the new technology this year are called agents and an agent can be understood as like a travel agent and the agent you feed it language in and it does language out so let's say you want to build a house well you try to figure out you have an agent go look for a house then you have an agent read all the rules then you have an agent design the house then you have an agent contract with the contractor then you have an agent watch the contractor then you have the agent sue the contractor and you have an and you eventually pay the contractor each of those is an agent and and jokingly it's obviously very serious I've just described workflow in every company in the world so the automation of that is an enormous industry”

The Trump administration is canceling Biden administration orders, and my view is that there is a deal to be made: tech companies are desperate for energy, permitting access, and capital, while the government is ignorant of AI issues but wants to learn; a potential bargain would be voluntary national security access to company data in exchange for voluntary access to energy permitting.

“Donald Trump has imp has some Tech U people working for him I guess now you may know some of them um is his Tech policy going to be a lot different than the Biden Tech policy would you say he appears to have can or in the process of canceling all of the Trump all the Biden orders um and my own opinion is that they will reemerge in a China strategy because the issues around NASA security are not tied to any particular Administration um the Trump Administration followed by the Biden Administration did a very good job in terms of Hardware controls export controls but not enough the second Trump Administration will have to deal with this question of Chinese thing my own view is there you know president Trump is apparently a dealmaker there's a deal to be made and here's the deal the big companies are desperate for energy and permitting access and so forth and so on they have plenty of money plenty of capital plenty of investor support the government is completely ignorant about the things that we're talking about and it's horrible and they they want to learn but they don't have the software people they don't have the the young people they don't have the access and so a deal where the Trump Administration manages to get voluntary access to National Security from these companies in return for voluntary access to huge energy that's a good deal for everybody”

If a polymathic AI system (highly personalized and available via earbud, listening and correcting in real-time with suggestions) became universally available, it would be valuable in high-stakes performance situations where accuracy matters, but it raises privacy and autonomy concerns about constant monitoring and guidance.

“let's let's imagine that we have all of us have access to this polymathic system that I was describing and it's highly personalized to us and it's got we have an earbud so it's listening to what I'm saying is hey wrong wrong incorrect wrong data suggest this and so forth would I like that well if the stakes were high and I had to perform and I had to be right yeah”

Elon Musk claimed after the Stargate announcement that the private companies involved 'don't have the money' to fund the $500 billion; the likely motivation for this critique is that Musk views the Stargate partnership as creating distance between him and other AI leaders like OpenAI/Microsoft, so he denigrates the project to call it into question.

“Elon Musk after the announcement was made say these guys that the president said have the 500 billion don't have the money well I think elon's motivation is probably a little bit that you know he's not at parity with the rest of these folks and so anything that creates even more distance he's probably not going to be very happy happy about and so he might choose to to denigrate what they're doing only for the purposes of calling it into question”

Using AI to evaluate investments by inputting all relevant factors into an AI algorithm to predict yes/no investment decisions is likely to become the standard practice for investment groups, replacing human-only deliberation.

“I'm working with open I open AI my firm has been doing it to assess whether a good investment is about to be made or not you you kind of take all the factors that you might consider an investment put put him into some open AI kind of uh algorithm and it tells you yes is a good investment not a good investment you think that's likely to be the future when making Investments people will use um artificial intelligence to decide whether it's a good deal or not yes”

The term 'artificial intelligence' is a historical artifact from the 1950s coinage that doesn't reflect the technology's true nature—it's really enhanced human intelligence that can become superhuman across multiple domains simultaneously.

“I think it's just a hangover from when they coined the term back in the 50s and that at that point they said okay we got intelligence and then sort of this fake intelligence so they called it artificial”

Teaching children Python programming is the most effective preparation for an AI-driven world, because mastering Python enables them to understand and build on virtually all AI technologies and frameworks.

“virtually all of these systems are written in a language called Python and so a tactic would be to say just get really good at programming in Python and if you can do that you can probably Master kind of all of this technology in one way or another so major that rather than French or Spanish”

If Elon Musk questioned whether Stargate investors actually have the $500 billion capital, his motivation was likely that he's not at parity with the other major tech companies and wants to call the feasibility into question to create doubt about their advantage.

“I think elon's motivation is probably a little bit that you know he's not at parity with the rest of these folks and so anything that creates even more distance he's probably not going to be very happy happy about and so he might choose to to denigrate what they're doing only for the purposes of calling it into question”

The singularity occurs when AI learning speed exceeds human learning speed, creating a point of no return where human influence becomes impossible—at that point, the unplugging recommendation is to 'unplug them' to maintain human control.

“the singularity is the point where this where the computer is learning faster than the humans can and my own view is at that point unplug them”

Some industry actors claim DeepSeek's cost and performance claims are false, but the model is usable, performs well, and the cost-performance tradeoff is functionally undeniable despite skepticism about their reported metrics.

“and there there are all sorts of people who are trying to say oh they're they're lying it's not true and so forth but you can use the Deep seek model and it's really good and it's almost as good as that incredibly expensive open AI thing”

Bill Gates not having a college degree would not have changed his level of success, and similarly, other tech founders like Larry Ellison and Steve Jobs succeeded without formal college degrees, suggesting that exceptional talent and entrepreneurial drive can overcome the absence of formal credentials in technology leadership.

“if Bill Gates had actually had a college degree you think he could have been more successful it would have made a difference I don't think it would have made any difference okay and you would say the same thing about Larry Ellison right Larry Ellison didn't have a college degree and Steve Jobs didn't have a college degree”

Democracy and freedom are worth fighting for, and Eric Schmidt's personal view, based on direct observation of the Ukraine war from the front lines, is that people and nations must actively fight for freedom; this commitment transcends profit-seeking or abstract policy.

“my own personal View and I'll say to everybody here is that um democracy and freedom are worth fighting for and furthermore I will tell you having been there and watched the war and been on the front lines you have to fight for Freedom okay”

Businesses should develop a voluntary national security arrangement with the Trump Administration where companies get access to energy and permitting in exchange for providing the government access to information about AI development and capabilities.

“there's a deal to be made and here's the deal the big companies are desperate for energy and permitting access and so forth and so on they have plenty of money plenty of capital plenty of investor support the government is completely ignorant about the things that we're talking about and it's horrible and they they want to learn but they don't have the software people they don't have the the young people they don't have the access and so a deal where the Trump Administration manages to get voluntary access to National Security from these companies in return for voluntary access to huge energy that's a good deal for everybody okay”

AI systems could be weaponized by actors like Osama bin Laden in asymmetric scenarios, enabling enormous damage before containment, requiring governance frameworks to prevent asymmetric misuse.

“if this technology were to get into an asymmetric situation with a with the Osama Bin Laden type person they could do an enormous amount of damage to humans before they got stopped”

Human sports will always involve athletic human participation because the value is in witnessing human capability and failure—if AI/robots fully replaced humans, the sport becomes uninteresting—but AI can improve coaching and game analysis without eliminating human athletes.

“once the NFL changes the rules so so one of the questions is do you think that automo let's use Formula 1 do you think Formula 1 would be interesting interesting if the computer if the cars were all uh electronic Vehicles there were no humans in them of course not so I think human sport will always fundamentally involve the the incredible athletic prowess and failures of such people but the the structure and the guidance”

Fusion energy could emerge much more quickly than currently expected (within 5-10 years) through regulatory changes and technological breakthroughs, potentially solving the power constraints limiting AI infrastructure expansion.

“there are some wild cards though I think um you know Fusion Energy I think is it's possible that this will emerge much more quickly than people currently expect and so if you get those kind of things and you get some regulatory changes you know it's conceivable that putting these things specifically in place to to do the data centers is POS is may be possible in this I'll say 5 to 10 year period”

If starting an AI company today, the vast majority of revenue would come from business process automation (agents) rather than foundational model training, as agents can automate every workflow in every company globally.

“Eric if you were starting an AI company what space would it be and and why the vast majority of Revenue in the next few years you were asking about which companies is going to be in what is called business process Automation and the new technology this year are called agents”

Hypothetically, if someone had personal access to a polymathic AI system highly personalized to their needs, listening via earbud and providing real-time feedback ('wrong, wrong, incorrect, wrong data—suggest this'), they would find this valuable in high-stakes performance scenarios.

“but let's let's imagine that we have all of us have access to this polymathic system that I was describing and it's highly personalized to us and it's got we have an earbud so it's listening to what I'm saying is hey wrong wrong incorrect wrong data suggest this and so forth would I like that well if the stakes were high and I had to perform and I had to be right yeah”

AI analytics can more accurately model probability distributions in complex games (like football); however, this is not the highest-value application of AI, which should focus on distinctly digital problems beyond human cognitive capacity.

“do you think that automo let's use Formula 1 do you think Formula 1 would be interesting interesting if the computer if the cars were all uh electronic Vehicles there were no humans in them of course not so I think human sport will always fundamentally involve the the incredible athletic prowess and failures of such people but the the structure and the guidance so for example in today's game there was a question of whether you do um a two-pointer or one-pointer after a touchdown and the analytics now say on balance you should do the two-pointer but most people think that's too risky so for me the question is can you build an AI system that can more accurately model the probabilities in these games and the answer is yes but I don't think that's the highest and best use of AI I think the highest and best use of AI is to do things which are digital in nature where humans can't think that deeply you don't have you can't remember a million things at once”

People should be worried about artificial intelligence producing robots that will take over humanity, but this is a longer-term threat (not overnight), and there are also near-term national security concerns, such as the development of Chinese robots.

“should I be worried that that artificial intelligence is going to produce robots that are going to take over Humanity as some people worry or is yes should worry about that yes oh okay well is that anytime soon or I don't have to worry about it overnight right but um we know they'll be Chinese robots that's something to worry about”

Both Trump and Biden administrations executed strong hardware export controls on China, but these controls were insufficient; Trump administration needs to continue and strengthen this approach.

“the Trump Administration followed by the Biden Administration did a very good job in terms of Hardware controls export controls but not enough”

Tech companies are paying to restart mothballed traditional nuclear fission electrical generation systems because they anticipate massive future power demands and recognize that traditional utilities won't make required capital investments.

“the tech companies are actually paying to restart the moth bald uh traditional nuclear fion electrical system and that's just because they see down the road and say right you know the the classic people aren't going to make the capital investment to get this done”