How AI Will Change Business Forever w/ Salim Ismail | EP #46

0:00We're living in a world that has been pummeled by exponential technologies. The world has now turned into information. The first time we started hearing about artificial intelligence was at a conference in Dartmouth in 1956. The first article that came out saying that robots are going to take all the jobs in five years appeared in 1964. AI is not going to take your job. It's someone using AI that's going to take your job. The AI plus a human being always turns out to be the best competitor. It's a tsunami of change and you're either an EXO surfing on top of the tsunami or you're being crushed by it.

0:36Get started as fast as you can because this is one of those where you cannot afford to be left behind on this paradigm. Welcome to Moon Shots and Mindsets here with my dear friend, Selene Missmeowl, the CEO of OpenEXO, and my co -author on a new book called Exponential Organizations 2 .0. So, Celine, the world is so different than it's ever been. And for an entrepreneur, especially an exponential entrepreneur, you're either building an exponential organization or you're dead on arrival. I mean, I think it's that black and white. It's gonna be an EXO or nothing in this decade ahead. We used to say the world changed radically 10 years ago, but the world changed radically like four months ago.

1:22Sorry, everything is different from that. And it's accelerating. You know, on stage and people say, when is it going to slow down? When is the world going to like, you know, where's the on -off switch? There is no on -off switch. There's no velocity meter. It's accelerating. And we can talk about why it's accelerating. But in this session, we're going to talk about one of the most important external attributes of being an exponential organization, which is AI and algorithms. But before we get there, for those just hearing this for the first time, Selim, what is an exponential organization? An exponential organization is a 21st century organization focused and delivering minimum 10x better faster and cheaper than your more linear peers.

2:04The 20th century is all about building scalable organizations focused on efficiency and predictability. Today, you need to be agile, flexible, adaptable and fast. And that's what an EXO is with a set of characteristics and A .I. on algorithms is one of those. I like to say that we're living in a world that has been pummeled by exponential technologies. I liken it to the asteroid that impacted the planet 65 million years ago, and it changed the world so rapidly, the environment so rapidly that the slow lumbering dinosaurs died and went extinct. The furry little mammals are very proud ancestors evolved to dominate.

2:42And the asteroid that struck the world now over the last 20, 30 years and accelerating is all of these exponential technologies. Computation, sensors, networks, AI, robotics, three printing synthetic biology, ARV, ARB, blockchain, all these technologies are just like transforming, you know, it's a tsunami of change and you're either an EXO surfing on top of the tsunami or you're being crushed by it. And I think anybody who's not been asleep for the last six months knows that one of those tsunami causing attributes is AI and algorithms. So let's jump into that. It's a super exciting opportunity.

3:25And I've mentioned this in one of our previous podcasts, but I saw a tweet from a friend of ours that said, you know, we're not far from the first three person billion dollar company starting, you know, literally overnight. And I think that's true. Yeah, I think that's exactly right. And it goes back to your six D's premise, right? Where you're digitizing the world. The world is now turned into information. Our relationships are now digital. Our memories aren't in our heads. And now we, with all of that tsunami of data, we actually need AI and algorithms to help manage and massage it all in the business small with an emerge for that, it can be really profound.

4:02So if we look at a quick definition, AI and algorithms are the use of artificial intelligence and algorithms to basically make sense of data sets and create value from those data sets. And then eventually in an EXO create a business model around Google is a quintessential example of this with the creation of ad sense and and PageRank. And PageRank. And the massive amount of value is created from that is unbelievable. And now we're seeing that pervasively being applied across the board. The world runs on AI today. Yeah, it does. And just to put AI in a frame, I think that people can appreciate is why now what has changed?

4:45Because there's something that's actually changed in the last five years. We've met certain thresholds. That's transformed the world. You know, if you think about it, AI is over as a concept, is 60 plus years old. The first time we started hearing about artificial intelligence was at a conference in Dartmouth in 1956. And then people say, well, no, it's the deep learning models and deep learning concepts are in you. But you know, that actually emerged in like 1960s, mid -60s. So why now, why we seeing this huge tsunami of change where every trillion dollar company is betting their future on these generative AI models.

5:23And it's four things. The first, as you well know, is computation, right? We're in this massive growth of computation that's doubling every 18 to 24 months, more as law. What recurs well calls a law of accelerating returns, doubling power and power and power. And finally, we've got enough computational power on the cloud. We have a new set of massive and video computational chips that are being deployed. And that's allowing us to run these deep learning models that were never possible at scale before. Yeah, in the past, we were focused on narrow AI, right? Fuzzy logic in your digital camera or anti -lock braking systems or credit card fraud detection very niche applications.

6:08Then we went to deep learning where you basically gave it and it learned as it went along. And that got us to a certain point. I remember chatting with Jeffrey Hinton a few years ago I think it was like 2017 and he's speaking at a conference and I said, you know, where does deep learning go? And he said we've hit the end of the life cycle. We need the next breakthrough to happen so that AI can progress to the next level I don't know what that is and literally at the same time the Transformer paper was being published and that turns out to be the big thing that all these LLMs are based on. The second major thing that's occurred is that we have a massive amount of data.

6:45Yeah, so global data has been doubling roughly every two years. We're about to hit 175 zetabytes. I love that, which is 175 billion, million megabytes of data. And this data is not labeled. It was an article recently that said we're going to create a new term called a yota bite, which is coming online. I love that. You know, the stat that always blows my mind is the one that goes in the last, you know, six months we created more data than the entire history humanity did until that point. And then we just did in the last six months and then we'll do it again in the next three months and then you're like, okay, I just can't.

7:22I just I would suggest that we can't comprehend or process that. A quick break from our episode on June the 6th. Selim and I are going to be running a free three hour workshop on how to actually build and design an exponential organization. We'd love to have you join us. If you join us on June the 6th, first of all, you'll get free access to the book, Exponential Organizations 2 .0, access to an AI that we've built that allows you to query the book and helps you design your exponential organization. It's June the 6th, it's three hours, it's free, we've never done this before. Click on the link below, dmandis .com, backslash, EXO, and join us.

8:02All right, now back to the episode. So computation, mass amount of data, label data, which is really the food stock for the large language models, open AI, stability, Bard, all of these. The third thing that's happening that is really changing this universe is the cost -trained AI systems has gone down by 99 .5 % in the last five years. So the efficiency of the algorithms. And this is something that people don't realize. This is what's speeding up the world, ladies and gentlemen. and it's the fact that the tech is getting so much better that every dollar goes so much further. So in genome sequencing, it cost Craig Vent or $100 million to sequence one genome.

8:47Now, it's $100 bucks, so $100 million will sequence, I mean, $100 million will sequence a million genomes, not one genome. And the same way it used to be, call it a thousand bucks to sequence, you know, to actually train an AI model, and now it's like 50 cents. So, or five bucks, it's transformative, how fast it's going, and then the third, you know, my favorite phrase, and half people listening and not gonna know this, but you know, no bucks, no buck Rogers, right? Is the amount of massive capital, I have a reference number, that in 2021 during the pandemic, like $160 billion of global corporate investment went into AI alone.

9:36It's crazy. And then Facebook pretty much doubled it on its own just in the trying to go after their different angles. I think that you add this together and this becomes one of the most important characteristics of an exponential organization because as you build out, you're looking at what data sources, what can I turn into information and then you need algorithms to make sense of it all. And the comment that you made earlier about the startups being created out of this are going to be profound. The two domains that blow my mind just specifically to LLMs are doctors and lawyers. And I can literally see the point where you literally don't need a doctor and you don't need a lawyer in a human form of any kind because you can just tell Chatsy Pt to build a contract and it'll write a contract for you.

10:24Well, it can right now, really well, And in fact, a lot of people are using CHAT GPT as their physician. The interesting reference point, right? So CHAT GPT goes live in November of 2022. In January, two months later, it passes the US Medical Licensing Exam, US MLE, which is insane. You normally need four years of medical school, two years of residency or internship to pass that and it does that. And here's a good example of how it's going to massively disrupt. I think eventually two things are going to happen. Your AI will be your diagnostician. Zero question about that. And before then, it's going to become malpractice to diagnose a patient without AI as your copilot in the loop.

11:10Here's a data point. Do you have any idea how many medical journal articles are published every day? I may have shared this with you right now. 7 ,000. 7 ,000 per day. So ask yourself how many of those journals has my doctor read this morning? Well, Daniel Kraft, the one he referenced to me was that there's 2 ,500 cancer research papers published today, right? So if you're an oncologist, there's no way you can keep track of all that. You need an AI to just surf that tsunami of data and make sense of it all. Yeah, we have at Fountain Life, where I'm proud to be executive chairman. You know, we, in an upload, we upload our members, they go through an upload every year, and it's 150 gigabytes of data.

11:55It's their full genome, full body MRI, brain, brain mask literature, coronary CT, 80 blood biomarkers, and they're changing all the time. And the database that we look into of knowledge is changing all the time. And it's kind of pathetic to think about a poor human trying to make heads or tails of that information. Cannot. So it's changing fast. You know, let's look at our brain, right? it's not had an upgrade in 50 ,000 years. It's stuck inside this, it's a leader and a half volume stuck inside here and it's trapped. It has evolutionary incremental gains and we are stuck with all these cognitive biases like sunk cost bias and framing bias and all these other biases that we can't get around very easily, these cognitive biases.

12:43Yeah, so I, that's it. Let's dive into that. that one second, and we'll come back to the more basic uses of AI and algorithms. I think one of the things we're all eventually going to have our own version of Jarvis. I don't think it's very far away, right? We have, I wouldn't say her name, it's spelled ALEX, A, it's sitting across the room here, I'll show it wake up. We have that and Siri and Google's variation, and I don't know if Cortana is still a thing or not. But they're early versions. And we're going to get to a point where we all have an AI that is on our bodies, you know, literally in our bodies.

13:24And we give permission to read everything we write, listen to our conversations, read our emails, our text messages, everything. Because then that AI is going to be able to be there to facilitate anything we want. And one of the things I wanted to facilitate is, hey, tell me when I have a cognitive bias. Because cognitive biases are, unfortunately, we cannot process all the information that comes into our minds. We have a very little flow. And so over evolutionary time period, our brains evolve these cognitive biases. They were neural hacks to help you reach conclusions quicker. like a similarity or a familiarity bias.

14:08Someone who looks like you, you trust more, or a recency bias, you give higher value to more recent information or a negativity bias, right? Are damned amygdala's that are, you know, destroying the planet and the crisis and use network is taking advantage of them, which you give far more value to negative news and positive news. What was the stat you and P .S. Stephen uncovered in abundance? Is it 10x? Yeah, so yeah, this is something I'd steep and Kotler and I found in looking at the news media and you can do the experiment yourself. Pick up a newspaper. I don't read them anymore, but if I'm going to conference someplace, I'll play the game and count the negative stories to positive stories.

14:47It's 10 to 1, right? And on cable news network, whatever the case might be, you're getting, you know, 95 % negative stories. Every murder brought to you over and over again in living color. Streamed as running devices. You know, this is such an important point. This is one of the most important pieces out of abundance that I can ever remember reading was this identification of this because we evolved that in MIGDLA. When you're running around on the plains of Africa, if you heard a noise in the bushes, you ran because bad news could kill you, right? If I missed some good news, I might miss some fruit that I could eat, but if I missed a piece of bad news, I died.

15:27So we're way more optimized for listening to bad news and good news because it's a survival factor And when I noticed one thing that I noticed when we were teaching courses at Singularity University When you showed somebody something new they immediately because of the unknown factor they related to it as danger Right, so the first when you saw something the autonomous car the first reaction is oh my god stop the car because that car might kill somebody And then we start with this negative perception of the technology and over a long period of time we kind of evolve a better version of it, et cetera, et cetera.

16:04We're not evidentiary based at all in terms of how we cognitive or analyze the world. It's a problem in screwing what the world, right? Our core software is fear and scarcity. Versus abundance and optimism, right? And fear and scarcity He puts you back on your heels in a protective world versus abundance and optimism against a mindset shift where you're taking advantage and you're jumping on. How can I use that? Where can I go with it? How can I make a better product or service for people? I mean, just to wrap up on cognitive biases, we have so many cognitive biases as humans. That's what I think.

16:43There's a whole list on Wikipedia. I wrote about it. We write about it in the book, EXO2. and at the end of the day, an AI will help you if you want to navigate those cognitive biases. Another going back to the fundamentals of why AI and algorithms are so critical. We're living in a world of an explosion of data. We're seeing the amount of data double every two years, and an explosion in sensors. We're hinting towards a trillion -sensor economy, I call it, right where autonomous cars with LiDAR and radar in cameras and drones flying and augmented reality glasses and everything is just generating massive amount of data.

17:25You're heading towards a world where you can know anything you want, anytime you want, anywhere you want. The data is there, but no way to access it without AI and algorithms. It becomes a critical filtering point. I mean, every EXO will have to be using AI. And what's magical I think about the LLMs and the generative AI that's emerged is we can use it everywhere. Right? One of our team members said, hey, optimize my email subject lines on our newsletter to Increase our open rates and our open rates on up 25 % just from that Right like him. So we'll start using it everywhere across every business function accounting fulfillment sales marketing the whole lot.

18:07So you know when you talk about the importance of of AI. I want you to read a couple of quotes here. Sundar Pichai, the CEO of Alphabet, has a quote that I love. He goes, artificial intelligence could have more profound implications for humanity than electricity or fire. I love that quote. And Elon has one which is a little more punchy. He goes, companies have to race to build AI or they'll be made uncompetitive. Essentially, If your competitor is racing to build AI, they will crush you. You know, I put it differently and I said, but still, you know, really black and white, which is there going to be two kinds of companies at the end of this decade, those that are fully utilizing AI and those that are out of business.

18:55And I think it's that black and white. Yeah, I really truly do. Hey, completely. It's kind of short. Can I show you an image? Yeah, please. Okay. Check this out. So we incorporated it, an AI, generative AI chatbot into the book where we you can ask it a question And it will literally look up the entire corpus of the book Pose it with your quething and then come back with an answer So I've asked it the question here. How do I turn a Brazilian shipping company into an EXO? And here's the answer, right? It literally tells you Identify transformative of purpose. There's your MTP, right? It has to align the purpose with the transformational shift in company leverage external resources like you can benefit from collaborating with other organizations in a bunch of ways.

19:39Consider an autonomous unit. You create a separate autonomous unit to develop new disruptive business models and then build the culture of experimentation. Right? And this is pretty damn good. That's like 50k worth of consulting right now. My entire community of 24 ,000 consultants, EXO people are going, hey wait a minute, what did you just do? I I think this is so fun. I have to digitize them. Well, as you often say, right, the crowd is a proxy for AI. Yeah, yeah. I tell you that. Just like a Uber driver is a proxy for an autonomous car. Yeah. So for sure. So for sure. We're going to automate more and more of these tasks.

20:18I think there's this unbelievable thing that's happened that nobody expected that we thought we'd automate manual tasks, watching dishes, and we just couldn't figure out to do that. And what we've automated is all the white color tests. So let me show one image here as well. And it's something that we've talked about a lot. Here we go. And when will AI reach and or exceed human level AI? So Ray Kurzweil has for the last 20 years consistently predicted 2029. I mean, the guy is committed to that date. And everybody in the AI industry was laughing at him. And but eventually and it's been pulled by all the AI experts it's gone like from 100 years ago to 50 years ago to you know 50 years They all come closer to right they all come closer and closer and then recently, you know Elon put this one out he said AI will be vastly smarter than any human and would overtake us by 2025 So I mean pretty Extraordinary now he since tweeted that you know, he said I think Ray is right about the timing So maybe it's somewhere between 25 and 29.

21:30But for those listening, I mean, here's the question. There's no question that we're heading towards and put aside whether this is AGI or conscious. It's a matter of will this be able to be as good or better than humans in your company? It's coming. And so how are you gonna start thinking about this if you're the owner of a business, If you're an investor, if you're an entrepreneur running a company, one of the things I talk about critical to this is every company needs to create or identify, hire, bring in what I call a chief AI officer. So that's one tip that I want to offer out. What is a chief AI officer?

22:15It is not someone coding your own large language model. It is not someone who is really coding. it's more someone who understands what's going on. They understand it at a core, and they know the players, and they are helping you as the entrepreneur, the CEO, the business owner, the investor, understand, you know, the platforms and what to utilize, because it's moving so fast that you need that person inside your organization. So one piece of advice and get a chief AI officer. In the book, we write, the by the time you read this chapter, it'll be able to date. It's such a pain they asked to write.

22:54We had to rewrite big chunks of the book after the chat, GPT came out, it's like we're shaking or fist at the gods around this. I want to go on a little bit of a rant. Okay, let's do that. You know, the 2029 number, which is called the technological singularity offer, which is the word. Well, Ray talks us about the singularity as 2045, yeah, versus 2029. I wait. I hate the discussion. And I'll explain why. Right. One, we don't know what intelligence is. Right now we have about a dozen facets of intelligence and we measure two pieces of it, which is the speed of thought processing and the ability to match concepts across frameworks.

23:29That's how we currently think, but we have emotional intelligence, we have spatial intelligence, we have the Eastern concept of presence or awareness. We don't touch on any of that. So what do we mean by intelligence is one big hairy problem of a question. And the second, which touches on the Chachy is what would you constitute overtaking? The minute I can prescriptively describe a task, an AI or a robot is gonna do much better than me anyway. So I find it's a great kind of framing discussion, but I don't find it that useful. I think as you said before, it's how will you use AI? Just like in chess, we all use AI now in chess to help us man navigate and do the box standard opening is then it comes down to the chess, the AI plus a human being always turns out to be the best competitor.

24:16I think that's where the world will end up. I mean, there's a lot of dystopian conversation going on right now around AI. Let's dive into that because I really want to take it head on. So there are two conversations. One is, will AI take your job? The second conversation is, will AI destroy humanity? And the consciousness are just igniting our amygdala's. It's like boom, boom, boom, right? And I don't think the global news has really taken on yet, but they'll have a heyday. Why are they going to see a why didn't you tackle the consciousness one and I'll tackle the jobs one? All right, well listen first of all, you know, I was reading a book called A Thousand Brains by Jeff Hawkins and it's it talks about how the brain works and these idea of these cortical columns and three -dimensional spacing as that means by which and I think it's clear to a lot of people who've studied this that currently AI as it's being built is orders of magnitude.

25:14You know, it could be millions, not billions of times, less complicated, in which it has a long way to go to get to the level of complexity of the human brain. And so I'm not worried about AI becoming conscious anytime soon. The place I get concerned is nefarious actors, you know, terrorists, individuals who are unfortunately looking to do more harm than good, using AI to do that. And it's going to be, as always, a virus, antivirus competition in that world. And so that's my near term concern and how do we do that? So we've seen Intel, for example, come up with a set of algorithms to determine deep fakes at 99 .5 % accuracy recently.

26:05So we're going to have problems and we're going to solve them. And I could say the world's biggest problems, the world's biggest business opportunities. The presidential candidate in the Turkey presidential election just resigned because of a sex porn tape that surfaced that he claims is a deep fake. It's just crazy how pervasive and high -pitched. Oh, wait till the US elections in 2024. It's gonna be, you know, we have all of these technologies that are being able to clone your voice and you're gonna get a call from your mom asking you, you know, passionately, please, on go vote for, you know, for this candidate.

26:43It's crazy. So I have a different take on the consciousness thing. You know, we have the same problem. We have no idea what consciousness is, right? And, and we don't know how to define it. We don't know how to test it. And when I think about AI, and I've heard you make this, when you look at page rank, it's evolving its own intelligence, scanning billions of webpages at the same time. And it's a completely complementary type of intelligence to human intelligence. It's not replicative, right? The Hollywood side will always portray it as a dystopian matrix terminator, a sky net over thing. If we're lucky, the robot overlords come and treat us like pets and if we're unlucky, we're food, right?

27:25That's pretty much that range that occurs. But in reality, we're adding capability to human intelligence, not subtracting or replicating in any way that I've seen. I want to go on a rant about the jobs thing, right? Because we worried, so we did some research on this. The first article that came out saying that robots are going to take all the jobs in five years appeared in 1964. And we have the clipping the thing. And every five years, you see that, you see the same kind of basic article. So can we go back to that 1964? I don't think it was robots. I think it was, um, automation was going to steal our jobs.

28:01Yeah. But this is a key point. Now, Now, let's look at some data. If you look at, say, Germany or Sweden or Korea, the most automated robotics enabled manufacturing countries in the world, employment has not dropped at all in traffic that's increased. They have the lowest unemployment because when you automate the robot work, there's so much more work to be done in problem solving, increased efficiency, better design, et cetera. My favorite example at a macro level is in the 70s when we created ATM bank machines, There was all this hand wringing going, what will we do with millions of bank tellers that will be out of a job?