Ex-Google Exec (WARNING): The Next 15 Years Will Be Hell Before We Get To Heaven! - Mo Gawdat

0:00 The only way for us to get to a better place and succeed as a species is for the evil people at the top to be replaced with AI. I mean, think about it. AI will not want to destroy ecosystems. It will not want to kill a million people. They'll not make us hate each other like the current leaders because that's a waste of energy, explosives, money, and people. But the problem is super intelligent AI is reporting to stupid leaders. And that's why in the next 15 years, we are going to hit a short-term dystopia. There's no escaping that. They having AI leaders. Is that even fundamentally possible? Let's put it this way. Mold Ga is back. >> And the former chief business officer at Google X is now one of the most urgent voices in AI with a very clear message.

0:46 >> AI isn't your enemy, but it could be your savior. >> I love you so much, man. You're such a good friend. But you don't have many years to live. Not in this world. Everything's going to change. Economics are going to change. Human connection is going to change. and lots of jobs will be lost including podcasting. >> No, no. Thank you for coming on today, Mo. >> But but the truth is it could be the best world ever. The society completely full of laughter and joy. Free healthcare, no jobs, spending more time with their loved ones. A world where all of us are equal. >> Is that possible? >> 100%. And I have enough evidence to know that we can use AI to build the utopia. But it's a dystopia if humanity manages it badly. a world where there's going to be a lot of control, a lot of surveillance, a lot of forced compliance

1:31 and a hunger for power, greed, ego, and it is happening already. But the truth is the only barrier between a utopia for humanity and AI and the dystopia we're going through is a mindset. >> What does society have to do? >> First of all, I see messages all the time in the comments section that some of you didn't realize you didn't subscribe. So, if you could do me a favor and double check if you're a subscriber to this channel, that would be tremendously appreciated. It's the simple, it's the free thing that anybody that watches this show frequently can do to help us here to keep everything going in this show in the trajectory it's on. So, please do double check if you've subscribed and uh thank you so much because a strange way you are you're part of our history and you're on this journey with us and I appreciate you for that. So, yeah, thank you.

2:16 Mo, two years ago today, we sat here and discussed AI. We discussed your book, Scary, Smart, and everything that was happening in the world. Since then, AI has continued to develop at a tremendous, alarming, mind-boggling rate, and the technologies that existed 2 years ago when we had that conversation have grown up and matured and are taking on a life of their own, no pun intended. What are what are you thinking about AI now, two years on? I know that you've started writing a new book called Alive, which is I guess a bit of a follow on or an evolution of your thoughts as it relates to scary smart. >> What what is front of your mind when it comes to AI? >> It's a scary smart was shockingly accurate. It's quite a I mean I don't

3:02 even know how I ended up writing predicting those things. I remember it was written in 2020, published in 2021 and then most people were like um who wants to talk about AI you know I know everybody in the media and I would go and do you want to talk and then 2023 CH GPT comes out and everything flips everyone realizes you know this is real this is not science fiction this is here and uh and things move very very fast much faster than I think we've ever seen anything ever move ever And and I think my position has changed on two very important fronts. One is remember when we spoke about scary smart I was still saying that there are things we can do

3:48 to change the course. Uh and we could at the time I believe uh now I've changed my mind. Now I believe that we are going to hit a a short-term dystopia. There's no escaping that. >> What is dystopia? >> I call it face rips. We can talk about it in details but the the the way we define very important parameters in life are going to be uh completely changed. So so face rips are you know the way we define freedom uh accountability human connection and equality economics uh reality innovation and business and power. That's the first change. So the first change in my mind is that uh is that we uh will have to prepare for a

4:35 world that is very unfamiliar. Okay. And that's the next 12 to 15 years. It has already started. We've seen examples of it in the world already even though people don't talk about it. I try to tell people you know there are things we absolutely have to do. But on the other hand I started to take an active role in building amazing AIs. So AIS that will uh not only make our world better uh but that will understand us understand what humanity is through that process. >> What is the definition of the word dystopia? >> So in my in my mind these are adverse circumstances that unfortunately might escalate beyond our control. The problem is the uh there is a lot wrong with the

5:22 um value set with the ethics of humanity at the age of the rise of the machines and when you take a technology every technology we've ever created just magnified human abilities. So you know you can walk at 5 km an hour you get in a car and you can now go you know 250 280 m an hour. Okay. uh basically magnifying your mobility if you want you know you can use a computer to magnify your u calculation abilities or whatever okay and and what AI is going to magnify unfortunately at this time is it's going to magnify the evil that man can do and and it is within our hands completely completely within our hands to change that but I have to say I don't think humanity has the awareness uh at this

6:09 time to focus on so that we actually use AI to build the utopia. >> So what you're essentially saying is that you now believe there'll be a period of dystopia and to define the word dystopia, I've used AI, it says a terrible society where people live under fear, control or suffering and then you think we'll come out of that dystopia into a utopia which is defined as a perfect or ideal place where everything works well, a good society where people live in peace, health and happiness. >> Correct. And the difference between them, interestingly, is what I normally refer to as the second dilemma, which is the po point where we hand over completely to AI. So, a lot of people think that when AI is in full control, it's going to be an existential risk for humanity. You know, I have enough uh evidence to

6:58 to argue that when we fully hand over to AI, that's going to be our salvation. that the problem with us today is not, you know, that intelligence is going to work against us. It's that our stupidity as humans is working against us. And I think the challenges that will come from humans being in control uh are going to outweigh the the challenges that could come from AI being in control. >> So, as we're in this dystopia period, did you do you forecast the length of that dystopia? Yeah, I count I count it exactly as 12 to 15 years. I I believe the beginning of the slope will happen in 2027. I mean it we will see signs in 26. We've seen signs in 24 but we will

7:44 see escalating signs next year and then a a clear uh slip in 27. >> Why? >> The geopolitical environment of our world is not very positive. I mean you really have to think deeply about not the not the symptoms but the the reasons why we are living the world that we live in here in today is money right and uh and money for anyone who knows who really knows money's you and I are peasants you know we build businesses we contribute to the world we make things we sell things and so on real money is not made there at all real money is made in lending in fractional reserve, right?

8:31 And and you know the biggest lender uh in the world would want reasons to lend and those reasons are never as big as war. I mean think about it, huh? Uh the world spent $2.71 trillion on war in 2024, right? A trillion dollars a year in the US. And when you really think deeply, I don't mean to be scary here. You know, weapons have depreciation. They depreciate over 10 to 30 years. Most weapons, >> they lose their value. >> They lose their value and they depreciate in accounting terms on the books of an army. The current arsenal of the US, that's a result of a deep search with my AI Trixie. You know the current

9:18 arsenal I think we we think cost the US 24 to 26 trillion dollars to build. My conclusion is that a lot of the wars that are happening around the world today are a means to get rid of those weapons so that you can have replace them. And uh you know when when your morality as an industry is we're building weapons to kill then you know you might as well use the weapons to kill. >> Who benefits? the lenders and the industry, >> but but they can't make the decision to go to war. They they have to rely on >> remember I I said that to you when we I think on on our third podcast. War is decided first then the story is manufactured. You you remember 1984 and the Orwellian approach

10:04 of like you know uh freedom is slavery and uh war is peace and they call it uh something speak uh basically to to to to convince people that going to war in another country to to kill 4.7 million people is freedom. You know we're going there to free the Iraqi people. Is war ever freedom? you know, to to tell someone that you're going to kill 300,000 women and children is for liberty and for the the the you know, for human values. Seriously, how do we ever get to believe that the story is manufactured and then we follow and humans because we're gullible uh we cheer up and we say,

10:51 "Yeah, yeah, yeah. We are we're on the right side. They are the bad guys." >> Okay. So, let me let me have a let me have a go at this idea. So, the idea is that really money is driving a lot of the conflict we're seeing and it's really going to be driving the dystopia. So, here's an idea. So, I um I was reading something the other day and it talked about how billionaires are never satisfied because actually what a billionaire wants isn't actually more money. It is more status. Correct. And I was looking at the sort of evolutionary case for this argument. And if you go back a couple of thousand years, money didn't exist. You were as wealthy as what you could carry. So even I think to the human mind, the idea of wealth and money

11:37 isn't a thing. what we've but what has always mattered from a survival of the fittest from a reproductive standpoint what's always had reproductive value if you go back thousands of years the person who was able to mate the most was the person with the most status so it makes the case the reason why billionaires get all of this money but then they go on podcasts and they want to start their own podcast and they want to buy newspapers is actually because at the very core of human beings is a desire to increase their status. >> Yeah. And so if we think of when we going back to the example of why wars are breaking out, maybe it's not money. Maybe actually it's status and and it's this prime minister or this leader or this, you know, individual wanting to create more power and more status because really at the heart of what

12:24 matters to a human being is having more power and more status. And money is actually money as a thing is actually just a proxy of my status. >> And and what kind of world is that? >> I mean, it's a [ __ ] up one. all these all these powerful men have uh >> correct >> are really messing the world up. But >> so so can can I can I can I >> actually AI is the same >> because we're in this AI race now where a lot of billionaires are like if I get AGI artificial general intelligence first then I basically rule the world >> 100%. That's exactly the the concept what I what I used to call the the the first inevitable now I call the first dilemma and scary smart is that it's it's a race that constantly accelerates. You think the next 12 years are going to be AI dystopia where things aren't >> I think the next 12 years are going to

13:10 be human dystopia using AI >> humaninduced dystopia using AI >> and you define that by a rise in warfare around the world as >> the last the last one the RIP the last one is basically you're going to have a massive concentration of power and a massive distribution of power okay and that basically will mean that those with the maximum concentration of power are going to try to oppress those with with democracy of power. Okay, so think about it this way in today's world um unlike the past uh you know the Houthis with a drone the Houthis are the Yemeni uh tribes basically resisting US power and Israeli power in the Red Sea. Okay. They use a

13:59 drone that is $3,000 worth to attack a uh a warship from from the US or an airplane from the US and so on that's worth hundreds of millions. Okay, that kind of democracy of power makes those in power worry a lot about where the next threat is coming from. Okay, and this happens not only in war but also in economics. Okay, also in innovation, also in technology and so on and so forth, right? And so basically what that means is that like you rightly said as the the the tech oligarchs are attempting to get to AGI. They want to make sure that as soon as they get to AGI that nobody else has AGI and and basically they want to make sure that nobody else has the ability to

14:46 shake their position of privilege if you want. Okay. And so you're going to see a world where unfortunately there's going to be a lot of control, a lot of surveillance, a lot of um of forced compliance if you want or you lose your privilege to be in the world and and it is happening already. >> With this acronym, I want to make sure we get through the whole acronym. So >> you like dystopians, don't you? >> I want to do the dystopian thing, then I want to do the utopia. >> Okay. >> And ideally how we move from dystopia to utopia. >> Mhm. So the the the the F in face R >> is the loss of freedom as a result of that power dichotomy. Right? So you have you have a massive amount of power as you can see today in uh one specific

15:32 army being powered by the US uh funds and a lot of money righting against peasants really that have no weapons almost at all. >> Okay. Some of them uh are militarized but the majority of the mill two million people are not. Okay. And so there is massive massive power that basically says, you know what, I'm going to oppress as far as I go. Okay. And I'm going to do whatever I want because the cheerleaders are going to be quiet, right? Or they're going to cheer or even worse. Huh? And so basically in in that what happens is max maximum power threatened by a democracy of power leads to a loss of freedom. A loss of freedom for everyone. >> Because how does that impact my freedom? >> Your freedom. Yeah,

16:18 >> very soon uh you will if you publish this episode you're going to start to get questions around should you be talking about this those topics in your podcast. Okay. Uh you know uh if I uh have been on this episode then probably next time I land in the US someone will question me say why do you say those things? Which side are you on? Right? and and and and you know you can easily see that everything I mean I I told you that before doesn't matter what I try to contribute to the world my bank will cancel my bank account every 6 weeks simply because of my ethnicity and my origin right every now and then they'll just stop my my bank account and say we need a document my other colleagues of a different color

17:04 or a different ethnicity don't get asked for another document right but but but that's because I come from an ethnicity that is positioned in the world for the last 30 40 years as the uh enemy. Okay? And and so when you really really think about it, in a world where everything is becoming digital, in a world where everything is monitored, in a world where everything is seen, okay, we don't have much freedom anymore. And I'm not actually debating that or or I don't see a way to fix that >> because the AI is going to have more information on us, be better at tracking who we are, and therefore that will result in certain freedoms being restricted. Is that what you're saying? >> This is one element of it. Okay. If you push that element further

17:49 in in in in a very short time if you've seen agent for example recently manos or chat GPT there will be a time where you know you'll simply not do things yourself anymore. Okay. You'll simply go to your AI and say hey by the way I'm going to meet Stephen. Can you please you know book that for me? >> Great. >> And and and yeah and it will do absolutely everything. That's great until the moment where it decides to do things that are not motivated only by your well-being. Right. Why would he do that? >> Simply because, you know, maybe if I buy a BA ticket instead of an Emirates ticket, some agent is going to make more money than other agents and so on, right? Uh and I wouldn't be able to even catch it up if I hand over completely to an AI. Uh go go a step further. Huh? Think about a world where everyone

18:36 almost everyone is on UBI. Okay. >> What's UBI? >> Universal basic income. I mean, think about the economics, the E and face rips. Think about the economics of a world where we're going to start to see a trillionaire before 2030. I can guarantee you that someone will be a trillionaire. I'm I'm you know I think there are many trillionaires in the world today or there we just don't know who they are. But there will be a new Elon Musk or Larry Allison that will become a trillionaire because of AI investments, right? And and that trillionaire will have so much money to buy everything. There will be robots and AIs doing everything and humans will have no jobs. Mean >> do you think that's a there's a real

19:21 possibility of job displacement over the next 10 years? And the the rebuttal to that would be that there's going to be new jobs created in technology. >> Absolute crap. >> Really? >> Of course. >> How how can you be so sure? >> Okay. So again, I am not sure about anything. So So let's just be very very clear. It would be very arrogant. Okay. To assume that I know >> you just said it was crap. >> My my belief is it is 100% crap. >> Take a job like software developer. >> Yeah. >> Okay. Uh Emma would love my my new startup is me, Senad, another technical engineer and a lot of AIS. Okay. That startup would have been 350 developers in the past. >> I get that. Um but are you now hiring in

20:08 other roles because of that or or you know as is the case with the steam engine? I can't remember the effect but there's you probably know that when steam when coal became cheaper people were worried that the coal industry would go out of business but actually what happened is people used more trains so trains now were used for transport and other things and leisure whereas before they were just used for commu for um cargo. Yeah. So there became more use cases and the coal industry exploded. So I'm wondering with technology, yeah, software developers are going to maybe not have as many jobs, but there everything's going to be software. >> Name me one. >> Name you one. What >> job? >> Name you that's going to be created. >> Yeah. One job that cannot be done by an AI. >> Yeah. >> Or a robot. >> My girlfriend's breath work retreat

20:55 business where she takes groups of women around the world. Her company is called Barley Breathwork. And there's going to be a greater demand for connection, human connection. >> Correct. Keep going. >> So there's going to be more people doing community events in real life festivals. I think we're going to see a huge surge in things like >> everything that has to do with human connection. >> Yeah, >> correct. I'm totally in with that. Okay. What's the percentage of that versus accountant? >> It's a much smaller percentage for sure in terms of white collar jobs. >> Now, who does she sell to? >> People with probably what? probably accountants or you know >> correct she she sells to people who earn money from their jobs. >> Yeah. >> Okay. So you have two forces happening. One force is there are clear jobs that will be replaced. Video editor is going

21:40 to be replaced. Uh >> excuse me. >> I love as as a matter of fact podcaster is going to be replaced. >> Thank you for coming on today Mo. It was seeing you again. But but but the truth is a lot so so you see the best at any job will remain the best software developer the one that really knows architecture knows technology and so on will stay for a while right and you know one of the funniest things I I interviewed Max Tedmar and Max was laughing out loud saying CEOs are celebrating that they can now get rid of people and have productivity gains and cost reductions because AI can do that job. The one thing they don't think of is AI will

22:26 replace them too. AGI is going to be better than at everything than humans at everything including being a CEO. Right? And you really have to imagine that there will be a time where most incompetent CEOs will be replaced. Most incompetent even breath work. Okay. Eventually there might actually one of two things be two things be happening. on one is either uh you know part part of that job other than the top breath work instructors, okay, are going you know who are going to gather all of the people that can still afford to pay for a breath work you know class they're going to be concentrated at the top and a lot of the bottom is not going

23:13 to be working for one of two reasons. One is either there is not enough demand because so many people lost their jobs. So when you're on UBI, you cannot tell the government, hey by the way, pay me a bit more for a breath work class. >> UBI being universal basic income just gives you money every month. >> Correct. And if you really think of freedom and economics, UBI is a very interesting place to be because unfortunately I as I said there's absolutely nothing wrong with AI. There's a lot wrong with the value set of humanity at the age of the rise of the machines, right? And the biggest value set of humanity is capitalism today. And capitalism is all about what? Labor arbitrage. >> What's that mean? >> I I I hire you to do something. I pay you a dollar. I pay it I sell it for two.

24:00 Okay. And and most people confuse that because they say, "Oh, but the cost of a product also includes raw materials and factories and so on and so forth." All of that is built is built by labor, right? So, so basically labor goes and mines for the material and then the material is sold for a little bit of margin then that material is turned into a machine. It's sold for a little bit of margin then that machine and so on. Okay, there's always labor arbitrage in a world where humanity's minds are being replaced by uh by AIs, virtual AIs, okay, and humanity's power strength within 3 to 5 years time can be replaced by a robot, you really have to question how this world looks like. It could be the best world ever. And that's what I believe

24:46 the utopia will look like because we were never made to wake up every morning and just, you know, occupy 20 hours of our day with work, right? We're not made for that. But we've fit into that uh uh, you know, system so well so far that we started to believe it's our life's purpose. >> But we choose it. We willingly choose it. And if you give someone unlimited money, they still tend to go back to work or find something to occupy their time with. >> They find something to occupy their time with, >> which is usually for so many people is building something. Philanthropy, a business%. So you build something. So between Senad and I, Emma. Love is not about making money. It's about finding true love relationships. >> What is that? Sorry, just for context. >> So So you know,

25:31 >> it's a business you're building just for the audience context. So, so, so the idea here is I can, it might become a unicorn and be worth a billion dollars, but neither I nor Senate are interested, okay? We're doing it because we can, okay? And we're doing it because it can make a massive difference to the world. >> And you have money, though. >> It doesn't take that much money anymore to build anything in the world. This is labor arbitrage. >> But to build something exceptional, it's still going to take a little bit more money than building something bad >> for the next few years. So whoever has the capital to build something exceptional will end up winning. >> So so this is a very interesting understanding of freedom. Okay. This is the reason why we have the AI arms race. Okay. Is that the one that owns the

26:17 platform is going to be making all the money and and keeping all the power. Think think of it this way. When humanity started the best hunter in the tribe could maybe feed the tribe for three to four more years more days. H and as a as a reward, he gained the favor of multiple mates in the tribe. That's it. The top farmer in the tribe could feed the tribe for a season more. Okay? And as a result, they got estates and you know uh and mansions and so on. The best industrialist in the in a in a city could actually employ the whole city, could grow the GDP of their entire country. And as a result, they became millionaires. the 1920s. H the best technologists

27:04 now are billionaires. Now what's the difference between them? The tool the the hunter only rem depended on their skills and the automation the entire automation he had was a spear. The farmer had way more automation. And the biggest automation was what? The soil. The soil did most of the work. The factory did most of the work. the the network did most of the work. And so that inc incredible expansion of wealth and power and as well the the incredible impact that something brings is entirely around the tool that automates. So who's going to own the tool? Who's going to own the the the digital soil, the AI soil? It's the platform owners.

27:50 >> And the platforms you're describing are things like OpenAI, Gemini, Grock. These these are interfaces to the platforms. The platforms are all of the uh of the uh um tokens, all of the compute that is in the background, all of the uh all of the uh uh methodology, the systems, the algorithms, that's the platform, the AI itself. You know, Grock is the interface to it. I think this is probably worth explaining in layman's terms to people that haven't built AI tools yet because I think I think to the listener they probably think that every AI company they're hearing of right now is building their own AI whereas actually what's happening is there is really five, six, seven AI companies in the

28:38 world and when I built my AI application I basically >> pay them for every time I use their AI. So if Steven Bartlett builds an AI at stephvenai.com, it's not that I've built my own underlying I've trained my own model. Really what I'm doing is I'm paying Sam Alman's chat GPT. Um every single time I do a a call, I basically um I do a search or you know I use a token. And I think that's really important because most people don't understand that unless you've built AI, you think, "Oh, look, you know, there's all these AI companies popping up. I've got this one for my email. I've got this one for my dating. I've got No, no, no, no, no. They're pretty much I would be I would hazard a guess that they're probably all OpenAI at this point. >> No, there are quite a few quite

29:24 different characters and quite differently, >> but there's like five or six. >> There are five or six when it comes to language models. Yeah. Right. Uh but interestingly, so yes, I I should say yes to start and then I should say but there was an interesting twist with Deepseek at the beginning of the year. So what Deepseek did is is they basically uh nullified the business model if you want in two ways. one is it was around a week or two after uh you know Trump stood you know with pride saying Stargate is the biggest investment project in the history and it's $500 billion to build AI infrastructure and soft bank and Larry Allison and and uh Sam Alman were sitting and so you know beautiful

30:10 picture and then DeepSeek R3 comes out it does the job for a one over 30 of the cost okay and interestingly is entire open source and available as an edge AI. So, so that's really really interesting because there could be now in the future as the technology improves the learning models will be massive but then you can compress them into something you can have on your phone and you can download deepseek literally offline on a um um you know an off the network computer and build an AI on it. There's a website that basically tracks the um sort of cleanest apples to Apple's market share of all the website referrals sent by AI chat bots and chatbt is currently at 79% roughly about

30:56 80%. Perplexi is at 11, Microsoft copilot about five, Google Gemini is at about two, Claude's about one and Deepseek is about 1%. And really like the the point that I I want to land is just that when you hear of a new AI app or tool or this one can make videos, >> it's built on one of them. It's basically built on one of these really three or four AI platforms that's controlled really by three or four AI you know billionaire teams and actually the one of them that gets to what we call AGI first where the AI gets really really advanced one could say is potentially going to rule the world as it relates to technology. >> Yes. Uh if if they get enough uh head

31:44 start. So, so I actually think that uh what I what I'm more concerned about now is not AGI, believe it or not. So, AGI in my mind and I said that back in 2023, right? Uh that we will get to AGI. At the time I said 2027, now I believe 2026 latest. Okay. The most interesting development that nobody's talking about is self-evolving AIS. self evolving AIS is think of it this way if you and I are hiring the top engineer in the world to develop our AI models and with AGI that top engineer in the world becomes an AI who would you hire to develop your next generation AI that AI

32:30 >> the one that can teach itself >> correct so one of my favorite examples is called Alpha Evolve so this is Google's attempt to basically have four agents working together four AIs working together to look at the at the code of the AI and say where is the where are the performance issues then you know an agent would say what's the problem statement what can I uh you know what do I need to fix uh one that actually develops the solution one that assesses the solution and then they continue to do this and you know I don't remember the exact figure but I think Google improved like 8% uh on their AI infrastructure because of alpha evol Right? And when you really really think, don't quote me on the number 8 to 10, 6 to 10, whatever in Google terms, by the

33:16 way, that is massive. That's billions and billions of dollars. Now, the the the the trick here is this. The trick is again, you have to think in game theory format. Is there any scenario we can think of where if one player uses AI to develop the next generation AI that the other players will say no no no no no that's too much you know takes us out of control every other player will copy that model and have their next AI model developed by an AI. >> Is this what Sam Alman talks about who's the founder of um chatbt/openai when he talks about a fast takeoff? I don't know exactly what which what what which you're referring to but we're all talking about a point now that we call

34:01 the intelligence explosion. So, so there is a moment in time where you have to imagine that if AI now is better than 97% of all code developers in the world and soon we'll be able to look at its own code own algorithms by the way they're becoming incredible mathematicians which wasn't the case when we last met if they can develop improve their own code improve their own algorithms improve their own uh uh you know uh network architecture or whatever you can imagine that very quickly the force applied to developing the next AI is not going to be a human brain anymore. It's going to be a much smarter brain and very quickly as humans like basically when when we ran the Google infrastructure when the machine said we need another server or a proxy server in

34:47 that place we followed. we we never really you know wanted to to object or verify because you know the code would probably know better because there are billions of transactions an hour or a day and so very quickly those self-evolving AIs will simply say I need 14 more servers here and we'll just you know the team will just go ahead and do it. I watched a video a couple of days ago where he Sam Alman effectively had changed his mind because in 2023 which is when we last met he said the aim was for um a slow takeoff which is sort of gradual deployment and open AI's 203 2023 note says a slower takeoff is easier to make safe and they prefer

35:34 iterative rollouts society can adapt in 2025 >> they changed their mind and Sam Alman said He now thinks a fast takeoff is more possible than he did a couple of years ago on the order of a small number of years rather than a decade. Um, and it to define what we mean by a fast takeoff, it's defined as when AI goes from roughly human level to far beyond human very quickly, think months to a few years, faster than governments, companies, or society can adapt with little warning, big power shifts, and hard to control. A slow takeoff, by contrast, is where capabilities climb gradually over many years with lots of warning shots. Um, and the red flags for a fast takeoff is when AI can

36:22 self-improve, run autonomous research and development and scale with massive compute compounding gains which will snowball fast. So, and I think from the video that I watched of Sam Orman recently, who again is the founder of Open Air and HBT, he basically says, and again I'm paraphrasing here. I will put it on the screen. We have this community knows things so I'll write it on the screen. But he effectively said that whoever gets to AGI first will have the technology >> to develop super intelligence >> where the AI can can rapidly increase its own intelligence and it will basically leave everyone else behind. >> Yes. Uh so that last bit is debatable but but let's just agree that uh so so in in a live uh you know one of the posts I I shared and got a lot of

37:09 interest is I refer to the the altman as a brand not as a human. Okay. So the altman is that uh persona of a California disruptive technologist that disrespects everyone. Okay. and believes that disruption is good for humanity and believes that this is good for safety and like everything else like we say war is for democracy and freedom they say uh developing you know putting AI on the open internet is good for everyone right it allows us to learn from our mistakes that was Sam Alman's 2023 spiel and if you recall at the time I was like this is the most dangerous you know one of the clips that really went viral you so you're you're so clever at finding the right clips is when I said >> I didn't I didn't do the clipping mate

37:56 >> they're team teams remember the clip where I said we [ __ ] up we always said don't put them on the open internet until we know what we're putting out in the world I'm going to be saying that >> yeah we we we [ __ ] up on putting it on the open internet teaching it to to code and putting you know agents AI agents prompting other AIs now AI agents prompting other AIs are leading to self-developing AIS and and The problem is, of course, we, you know, anyone who has been on the inside of this knew that this was just a clever spiel made by a PR manager for Sam Alman to sit with his dreamy eyes in front of Congress and say, "We want you to regulate us." Now, they're saying, "We're unregulable."

38:43 Okay? And and when you really understand what's happening here, what's happening is it's so fast that none of them has the choice to slow down. It's impossible. Neither China versus America or OpenAI versus Google. the that the the only thing that I may have may see happening that you you know that that may differ a little bit from your statement is if one of them gets there first uh then they dominate for the rest of humanity that is probably true if they get there first uh with within enough buffer. Okay. But the way you look at Grock coming a week after open AI, a week after uh you know Gemini, a week after Claude and then

39:30 Claude comes again and then China releases something and then Korea releases some something. It is so fast that we may get a few of them at the same time or a few months apart. Okay, before one of them has enough power to become dominant. And that is a very interesting scenario. multiple AIs, all super intelligent. >> It's funny, you know, I I I got asked yesterday, I was in I was in Belgium on stage. There was, I don't know, maybe 4,000 people in the audience and a kid stood up and he was like, um, you've had a lot of conversations in the last year about AI. Like, why do you care? And I don't think people realize how, even though I've had so many conversations on this podcast about AI, you >> haven't made up your mind. >> I I I have more questions than ever.

40:16 >> I know. And it's and it doesn't seem that anyone can satiate. >> Anyone that tells you they can predict the future is arrogant. >> Yeah. >> It is. It's never moved so fast. >> It's nothing like nothing I've ever seen. And you know, by the time that we leave this conversation and I go to my computer, there's going to be some incredible new technology or application of AI that didn't exist when I woke up this morning. That creates probably another paradigm shift in my brain. Also, you know, I people have different opinions of Elon Musk and they're they're entitled to their own opinion, but the other day, only a couple of days ago, he did a tweet where he said, "At times, AI existential dread is overwhelming." And on the same day, he tweeted, "I resisted AI for too long,

41:03 living in denial. Now it is game on." And he tagged his AI companies. I don't know what to make of I don't know what to make of those tweets. I don't know. And you know, I I try really hard to figure out if someone like Sam Wman has the best interests of society at heart. >> No. >> Or if these people are just like >> I'm saying that publicly. No. As a matter of fact, so I know Sundur Pachai. I work CEO of Alphabet, Google's parent company. an amazing human being on in all honesty. I know Dennis Hassab is amazing human being. Okay. Uh you know these are are ethical incredible uh

41:50 humans at heart. They have no choice. Uh Sund by law is uh demanded to take care of his his shareholder value. That's that is his job. >> But Sund you said you know him. You used to work at Google. >> Yeah. He's not going to do anything that he thinks is going to harm humanity. >> But if if he does not continue to advance AI, that by definition uh uh uh contradicts his responsibility as the CEO of a publicly traded company, he is liable by law to continue to advance the agenda. There's absolutely no doubt about it. Now, so but but he's a good person at heart. Deis is a good person at heart. So they're trying so hard to make it safe. Okay? As much as they can.

42:36 Reality however is the the the disruptor the altman as a brand doesn't care that much. >> How do you know that? >> In reality the disruptor is someone that comes in with the objective of I don't like the status quo. I have a different approach. And that different approach if you just look at the story was we are a nonforprofit that is funded mostly by Elon Musk money. It's not entirely by Elon Musk money. So context for people that might not understand Open AI. The reason I always give context is funnily enough I I think I told you this last time. I went to a prison where they play the D of CEO. >> No way. >> So they play the D of CO and I think it's 50 prisons in the UK to young offenders >> and no violence there. >> Well, I don't know. I can't I can't I can't tell you whether violence has gone up or down. But I was in the cell with

43:21 one of the prisoners, a young a young black guy, and I was in his cell for for a little while. I was reading through his business plan, etc. And I said, "You know what? You need to listen to this conversation that I did with Mo Gordat." So I he has a little screen in his cell. So I pulled it up, you know, our first conversation. I said, "You should listen to that one." And he said to me, he said, "I can't listen to that one cuz you guys use big words." >> So ever since that day, which was about >> I noticed that about days four years ago, sorry. >> I've always whenever I hear a big word, I think about this kid. >> Yeah. >> And I say like give context. So even with the you're about to explain what Open AI is, I know he won't know what Open AI's origin story was. That's why I'm >> I think that's a wonderful practice in general. By the way, even, you know, being a non native English speaker, >> you'll be amazed how often a word is

44:06 said to me and I I'm like, yeah, don't know what that means. >> So, like I've actually never said this publicly before, but I now see it as my responsibility to be to to keep the draw the drawbridge to accessibility of these conversations down for him. So, whenever I whenever there's a word that at some point in my life I didn't know what it meant, >> I will go back. I was like, what does that mean? I think that I've noticed that in the you know more and more in your podcast and I really appreciate and we also show it on the screen sometimes. >> I I think that's wonderful. I mean the the the origin story of open AI is as the name suggests it's open source. It's for the public good. It was an in you know intended in Elon Musk's words to save the world from the dangers of AI

44:52 right so they were doing research on that and then you know there was the disagreement between Sam Alman and and Elon somehow Elon ends up being out of uh of uh of open AI. I think there was a moment in time where he tried to take it back and you know the board rejected it or some something like that. most of the uh top um safety engineers, the top technical teams in open AI left in 2023 2024 openly saying we're not concerned with safety anymore. It moves from being a nonforprofit to being one of the most valued companies in the world. There are billions of dollars at stake, right? And if you if you tell me that Sam Altman is

45:38 out there trying to help humanity, let's let's suggest to him and say, "Hey, do you want to do that for free? We'll pay you a very good salary, but you don't have stocks in this. Saving humanity doesn't come at the billion dollar valuation or of course now tens of billions or hundreds of billions." And and and see truly that is when you know that someone is doing it for the good of humanity. Now the the capitalist system we've built is not built for the good of humanity. It's built for the good of the capitalist. >> Well, he might say that releasing the model publicly, open sourcing it is too risky because then bad actors around the world would have access to that technology. So

46:23 he might say that closing open AI in terms of not making it publicly viewable is the right thing to do for safety. We go back to gullible cheer leaders, right? One of the interesting tricks is of lying in our world is everyone will say what helps their agenda. Follow the money. Okay, you follow the money and you find that you know at a point in time Samman himself was saying it's open AI. Okay, my benefit at the time is to give it to the world so that the world looks at it. They know the code if there is if there are any bugs and so on. True statement. Also a true statement is if I put it out there in the world, a criminal might take that model and build something that's against humanity as a

47:09 result. Also true statement. Capitalists will choose which one of the truths to say, right? Based on which part of the agenda, which part of their life today they want to serve, right? Someone will say, uh, you know, do I do you want me to be controversial? Let's not go there. But if we go back to war, I'll give you 400 slogans. 400 slogans that we all hear that change based on the day and the army and the location and the they're all slogans. None of them is true. You want to know the truth. You follow the money, not what the person is saying, but ask yourself why is the person saying that? What's in it for the person speaking? >> And what do you think's in it for Chachi

47:55 Samman? hundreds of billions of dollars of of of valuation. >> And do you think it's that power? >> The ego of being the person that invented AGI, the position of power that this gives you, the meetings with all of the heads of states, the admiration that gets it, it is intoxicating >> 100% 100%. Okay. And and the real question, this is a question I ask everyone. Did you see you didn't you're every time I ask you you say you didn't. Did you see the movie Elysium? >> No. You'd be surprised how little movie watching I do. You'd be shocked. >> There are some movies that are very interesting. I use them to to create an emotional attachment to a story that you haven't seen yet because you may have seen it in a movie. Okay. Elissium is a

48:41 is a society where the elites are living on the moon. Okay. They don't need peasants to do the work anymore and everyone else is living down here. Okay. You have to imagine that if again game theory you have to im you know picture something to infinity to its extreme and see where it goes and the extreme of a world where all manufactured is done manufacturing is done by machines where all decisions are made by machines and those machines are owned by a few is not an economy similar to the to today to the to today's economy that today's economy is an economy of consumerism and and product and production. You know, it's the it's the

49:27 in in alive I call it the invention of more. The invention of more is that post World War II as the factories were rolling out things and prosperity was happening everywhere in America. There was a time where every family had enough of everything. >> But for the capitalist to continue to be profitable, they needed to convince you that what you had was not enough. either by making it obsolete like fashion or like you know a new shape of a car or whatever or by convincing you that there are more things in life that you need so that you become complete without those things you don't and and that invention of more gets us to where we are today an economy that's based on production consumed and if you look at the US

50:13 economy today 62% of the US economy GDP is consumption it's not production okay Now, this requires that the consumers have enough purchasing power to to buy what is produced. And I believe that this will be an economy that will take us hopefully in the next 10, 15, 20 years and forever. But that's not guaranteed. Why? Because on one side if UBI replaces purchasing power. So if people have to get an income from the government which is basically taxes collected from those using AI and robots to to make things the then the the mindset of capitalism labor arbitrage means those people are not producing anything and they're

50:59 costing me money. Why don't we pay them less and less and maybe even not pay them at all? And that becomes illissium where you basically say, you know, we sit somewhere protected from everyone. We have the machines do all of our work and those need to worry about themselves. We're not going to pay them UBI anymore, right? And and you have to imagine this idea of UBI assumes this very democratic caring society. UBI in itself is communism. Think of the ideology between at least socialism. The ideology of giving everyone what they need. That's not the capitalist democratic society that the west

51:45 advocates. So those transitions are massive in magnitude. And for those transitions to happen, I believe the right thing to do when the cost of producing everything is almost zero because of AI and robots. because the cost of harvesting energy should actually tend to zero once we get more intelligent to harvest the energy out of thin air. Then a possible scenario and and I believe a scenario that AI will eventually do in the utopia is yeah anyone can get anything they want. Don't over consume. We're not going to abuse the the planet resources but it costs nothing. So like the old days where we were hunter gatherers, you would, you know, forge for some berries

52:30 and you'll find them ready in in nature. Okay, we can in 10 years time, 12 years time build a society where you can forge for an iPhone in nature. It will be made out of thin air. Nanopysics will allow you to do that. Okay? But the challenge, believe it or not, is not tech. The challenge is a mindset. Because the elite, why would they give you that for free? Okay. And the system would morph into, no, no, hold on. We will make more money. We will be bigger capitalists. We will feed our ego and hunger for power more and more. And for them, give them UBI and then 3 weeks later give them less UBI. >> Aren't there going to be lots of new jobs created though? Because when we think about the other revolutions over

53:16 time, whether it was the industrial revolution or other sort of big technological revolutions, >> in the moment we forecasted that everyone was going to lose their jobs, >> but we couldn't see all the new jobs that were being created >> because the the the machines replaced the human strength at a point in time. And very few places in the west today will have a worker carry things on their back and carry it upstairs. The machine does that work. Correct. >> Yeah. >> Uh similarly AI is going to replace the brain of a human. And when the west in its interesting uh virtual colonies that I call it uh basically outsourced all labor to the to the developing nations.

54:03 What the West publicly said at the time is we're we're going to be a services economy. We we're we're not interested in making things and stitching things and so on. Let the Indians and Chinese and you know Bengali and Vietnamese do that. We're going to do more refined jobs. Knowledge workers. We're going to call them. Knowledge workers are people who work with information and click on a keyboard and move a mouse and you know sit in meetings and all we produce in the western societies is what words right or designs maybe sometimes but everything we produce can be produced by an AI. So if I give you an AI tomorrow h where I give you a piece of land, I give the AI a piece of land and I say here are

54:50 the parameters of my land. Here is its location on Google maps. Design an architecturally sound villa for me. I care about a lot of light and I need three bedrooms. I want my bathrooms to be in white marble, whatever. And the AI produces it like that. How often will you go to an to an architect and say right so what will the architect do the best of the best of the architects will either use AI to produce that or you will consult with them and say hey you know I've seen this and they'll say it's really pretty but it wouldn't feel right for the person that you are yeah those jobs will remain but how many of them will remain how how often do you think uh how many more years. Do you think I will be able

55:36 to create a book that is smarter than AI? Not many. I will still be able to connect to a human. You're not going to hug an AI when you meet them like you hug me, right? But that's not enough of a job. So why do I say that? Remember I asked you at the beginning of the podcast to remind me of solutions. Why do I say that? Because there are ideological shifts and and concrete actions that need to be taken by governments today rather than waiting until COVID is already everywhere and then locking everyone down. Governments could have reacted before the first patient or at least at patient zero or at least at patient 50. They didn't. H what I'm

56:24 trying to say is there is no doubt that lots of jobs will be lost. There's no doubt that there will be sectors of society where 10 20 30 40 50% of all developers, all software uh you know all graphic designers, all um uh uh u online marketers, all all all assistances are going to be out of a job. So are we prepared as a society to do that? Can we tell our governments there is an ideological shift? This is very close to social socialism and and communism. Okay. And are we ready from a budget point of view instead of spending a a trillion dollars a year on on arms and and explosives and you know autonomous weapons that will oppress people because

57:11 we can't feed them? Can we please shift that? I did those numbers. Huh. Uh again I I go back to military spending because it's all around us. 2.71 trillion. 2.4 to2.7 is the estimate of 2024. how much money we're spending on military >> on Yeah. on military equipment on things that we're going to explode into smoke and death. Extreme poverty worldwide. Extreme poverty is people that are below the poverty line. Extreme poverty everywhere in the world could end for 10 to 12% of that budget. So if we replace our military spending 10% of that to go to people who are in extreme poverty, nobody will be poor in the world. Okay. You can end uh world hunger for less

57:56 than 4%. Nobody would be hungry in the world. You know, if you take uh again 10 to 12% universal healthcare, every human being on the planet would have free healthcare for 10 to 12% on what we're spending on war. Now, why why do I say this when we're talking about AI? Because that's a simple decision. If we stop fighting because money itself does not have the same meaning anymore because the economics of money is going to change because the entire meaning of capitalism is ending because there is no more need for labor arbitrage because AI is doing everything just with the $2.4 trillion we save in explosives every year in arms and

58:42 weapons just for that universal healthcare and extreme poverty. You could actually one of the calculations is you could end climate or combat climate climate change meaningfully for 100% of the military budget. >> But I I'm not even sure it's really about the money. I think money is a measurement stick of power. Right. >> Exactly. It's printed on demand. >> So even in a world where we have super intelligence and money is no longer a problem. >> Correct. >> I still think power is going to be insatiable for so many people. So there will still be war because you know >> there will be in my view >> the strongest I want the strongest AI. I don't want my >> and I don't and I don't want you know what Harry Henry Kissinger called them the eaters. >> The eaters.

59:28 >> Yeah. Brutal as that sounds. >> What is that? The people at the bottom of the socioeconomic >> that don't produce but consume. So if you had a Henry Kissinger at the at the helm and we have so many of them, what would they think like why why prominent military figure in the US history? Uh you know what why would we feed 350 million Americans America will think but more interestingly why do we even care about Bangladesh anymore if we can't make our textiles there or we don't want to make our textile there. Do you you know I imagine throughout human history if we had podcasts conversations would would have been warning of a

1:00:13 dystopia around the corner. You know when they heard of technology and the internet they would have said oh we're finished and when the the tractor came along they would have said oh god we're finished because we're not going to be able to farm anymore. So is this not just another one of those moments where we couldn't see around the corner so we we forecasted unfortunate things. You could be. I I am I'm begging that I'm wrong. Okay. I'm just asking if there are scenarios that you think that can provide that. You know, uh uh Mustafa Sulleman in in uh you hosted him here. I did. Yeah. He was in the coming wave. >> Yeah. >> And he speaks about uh about pessimism aversion. Okay. that all of us people who are supposed to be in technology and business and so on, we're always supposed to, you know, stand on stage

1:00:58 and say the future's going to be amazing. You know, this technology I'm building is going to make everything better. One of my posts in life was called the broken promises. How often did that happen? >> Okay. How often did social media connect us? And how many and how often did it make us more lonely? How how often did mobile phones make us work less? That was the promise. That was the promise. The promise. The early ads of Nokia were people at parties. Is that your experience of mobile phones? And and I think the whole idea is we should hope there will be other roles for humanity. By the way, those roles would resemble the times where we were hunter gatherers, just a lot more technology

1:01:44 and a lot more safety. >> Okay. So, this is this sounds good. >> Yeah, >> this is exciting. So, I'm gonna I'm gonna get to go outside more, be with my friends more, >> 100%. >> Fantastic. >> And do absolutely nothing. >> Well, that doesn't sound fantastic. >> No, it does. Do be forced to do absolutely nothing. For some people, it's amazing. For you and I, we're going to find the little carpentry project and just do something. >> Speak for yourself. I'm still People are still going to tune in. >> Okay. >> Correct. Yeah. But what? And people are going to to tune in. >> Do you think they will? I'm not I'm not I'm not convinced they will. And for for as long >> will you guys tune in? Are you guys still going to tune in? >> I can let them answer. I believe for as long as you make their life enriched, >> but can an AI do that better

1:02:29 >> without the human connection? >> Comment below. Are you going to listen to an AI or the Davosio? Let me know in the comment section below. >> Remember, as incredibly intelligent as you are, Steve, uh there will be a moment in time where you're going to sound really dumb compared to an AI. and and and I will sound completely dumb. >> Yeah. Yeah. >> The the depth the depth of analysis and and gold nuggets. I mean, can you imagine two super intelligences deciding to get together and explain um string theory to us? They'll do better than any physic physicist in the world because they possess the physics knowledge and they also pro possess social and language knowledge that most deep physicists

1:03:16 don't. I think B2B marketeteers keep making this mistake. They're chasing volume instead of quality. And when you try to be seen by more people instead of the right people, all you're doing is making noise. But that noise rarely shifts the needle and it's often quite expensive. And I know as there was the time in my career where I kept making this mistake that many of you will be making it too. Eventually I started posting ads on our show sponsors platform LinkedIn. And that's when things started to change. I put that change down to a few critical things. One of them being that LinkedIn was then and still is today the platform where decision makers go to not only to think and learn but also to buy. And when you market your business there, you're putting it right in front of people who actually have the power to say yes. and you can target them by job title, industry, and company size. It's simply

1:04:02 a sharper way to spend your marketing budget. And if you haven't tried it, how about this? Give LinkedIn ads a try, and I'm going to give you a $100 ad credit to get you started. If you visit linkedin.com/diary, you can claim that right now. That's linkedin.com/diary. I've I've really gone back and forward on this idea that even in podcasting that all the podcasts will be AI podcasts or I've gone back and forward on it and and where I landed at the end of the day was that there'll still be a category of media where you do want lived experience on something >> 100%. >> For example, like you want to know how the person that you follow and admire dealt with their divorce. >> Yeah. Or or how they're struggling with AI, >> for example. Yeah. Exactly. But I but I

1:04:47 think things like news, there are there are certain situations where just like straight news and straight facts and maybe a walk through history may be eroded away by AIS. But even in those scenarios, you there's something about personality. And again, I I hesitate here because I question myself. I'm not in the camp of people that are romantic, by the way. I'm like I'm trying to be as as orientated towards whatever is true, even if it's against my interests. And I hope people understand that about me. like um cuz even in my companies we experiment with like disrupting me with AI and some people will be aware of those experiments >> because there will be a mix of all there you can't imagine that the world will be completely just AI and completely just podcasters you know you'll see a mix of of both you'll see things that they do

1:05:33 better things that we do better >> the the the message I'm trying to say is we need to prep for that >> we need to be ready for that we need to be ready by you know talking to our governments and saying hey it looks like I'm a a parallegal and it looks like all parallegals are going to be, you know, financial researchers or analysts or graphic designers or, you know, call center agents. It looks like half of those jobs are being replaced already. >> You know who Jeffrey Hinton is? >> Oh, Jeffrey. I I had him on the documentary as well. I love Jeffrey. >> Jeffrey Hinton told me >> trained to be a plumber. >> Really? >> Yeah. 100% for a while. >> And I I thought he was joking. 100%. >> So I asked him again and he he looked me dead in the eye and told me that I I

…