Are We Ready for the AI Revolution That's About to Reshape Our World?
In this episode of Leveraging AI, Isar Meitis talked about the convergence of groundbreaking technologies poised to revolutionize how we live, work, and conduct business. From advanced AI models to quantum computing leaps, we explore what's on the horizon for business leaders, C-suite executives, and entrepreneurs.
Topics include:
This episode isn't just about understanding the future; it's about preparing for it. Whether you're at the helm of a startup or leading a Fortune 500 company, the insights shared today could be the key to unlocking unimaginable opportunities and navigating the challenges of tomorrow's business landscape.
Tune in, get inspired, and let's navigate the future of business together. Don't forget to subscribe for more episodes that equip you with the knowledge to lead in the AI era
About Leveraging AI
If you’ve enjoyed or benefited from some of the insights of this episode, leave us a five-star review on your favorite podcast platform, and let us know what you learned, found helpful, or liked most about this show!
Are We Ready for the AI Revolution That's About to Reshape Our World?
In this episode of Leveraging AI, Isar Meitis talked about the convergence of groundbreaking technologies poised to revolutionize how we live, work, and conduct business. From advanced AI models to quantum computing leaps, we explore what's on the horizon for business leaders, C-suite executives, and entrepreneurs.
Topics include:
This episode isn't just about understanding the future; it's about preparing for it. Whether you're at the helm of a startup or leading a Fortune 500 company, the insights shared today could be the key to unlocking unimaginable opportunities and navigating the challenges of tomorrow's business landscape.
Tune in, get inspired, and let's navigate the future of business together. Don't forget to subscribe for more episodes that equip you with the knowledge to lead in the AI era
About Leveraging AI
If you’ve enjoyed or benefited from some of the insights of this episode, leave us a five-star review on your favorite podcast platform, and let us know what you learned, found helpful, or liked most about this show!
Hello and welcome to Leveraging ai, the podcast that shares practical, ethical ways to leverage AI to improve efficiency, grow your business, and advance your career. This is Isar Meitis, your host, and this is a special episode in which instead of diving into a specific use case, we are going to talk about multiple technologies that has either reached or reaching maturity in the near future, and that the convergence of them is going to completely change everything we know about the world we live in. It's important for each and every one of you to understand where these technologies are, where they could be in the near future, and why the combination of them will be so impactful to everything we know, including how we run our businesses, because we'll help you better understand what the future may bring and hopefully help you plan for your business and personal future. This episode is brought to you by the AI Business Transformation Course. We have been running these courses open to the public as well as to close the organization and companies. We are fully booked until April, but we're opening another cohorts that starts on April 1st. If you want to learn how to use AI within a business context, tools, use cases, processes, strategy, how to take your business into the AI era. It's a fantastic course. It's four sessions of two hours each, once a week with homework and exercise and in hands-on. And it's always with a fascinating group of people like you, managers and leaders in business organizations who want to understand how to use AI within their businesses. So if that's something that's interesting to you or to somebody that you know, please go and check our website at Multiplai.ai and we'll put a link to the course page in the show notes. Let's start by talking about AI models. The AI large language models have been growing dramatically in this past year, right? We started with GPT-3 0.5 when it was released back in November of 2022, but since then there's been major upgrades and improvements across the board. GPT-4 Turbo, which is the one we're on right now, still probably the reigning king. There are several different aspects that Gemini already does better, and this is Gemini Ultra One-point-zero. We don't know what Google actually has as far as more advanced models and whether they chose to release something that is just a GPT level because that's the only thing they could develop, or because they're keeping some of their cards close to their chest to see what OpenAI is going to do next. It is very clear at this point that now the Gemini is out, that OpenAI is going to release something. Soon, it may not be GPT-V. That is supposed to be a Game-changer. It might be GPT-4. Something that we're going to see coming out. A lot of people call it 4.5, but either way, we're going to see more advanced models. There have been huge improvements on the open source side initially with Llama, but since then, many open source models came out including very powerful, and very capable open-source models like Mixtral and Mistral from the Mistral company. Llama 2, and its different variants from Meta and the latest one that we talked about last week, Smog that achieves very high scores across multiple benchmarks. Now in addition to that, across universities and different research groups, they are able to run more and more advanced models with less computing powers and on multiple devices, including natively on mobile devices. So what we're seeing is that we're running better and more capable models with the opportunity to potentially run it with less required computing power. Now, the next thing that we've seen a lot of upgrades in is infrastructure. In order to run these models, especially the more advanced models, they currently require huge amount of data for training. They require the computing power and the farms in order to run these models. They require very specific chips. They require a lot of energy and water for cooling. But what we're seeing is that on one hand, huge investments that are going in that direction. And some of you may have heard that Sam Altman wants to raise trillions of dollars in order to build AI infrastructure for the future of our planet. Whether that's feasible or not, I don't know, but this is the direction, or at least the way these companies and these leaders think. The flip side, as I mentioned, is that we're getting better and better capabilities with less and less required computing power. And now we have AI tools running directly on mobile devices, and we've seen the latest releases from Google with her Pixel phones as well as from Samsung. And there's no doubt in my mind that Apple is coming next and everybody else that is making mobile devices will have the ability to run some AI models on board the chipset that's on these devices. So from a model technology perspective, from the architecture of the models and from the infrastructure, it will allow to run more and more advanced models. Another very interesting development that, again, is not new but is getting to a peak scenario right now is AI agents. AI agents are AI tools that are able to take actions on our behalf. So not just review data, analyze data, and provide answers, but also to actually take action. There's a paper from 2017 called World of Bits. That paper was written by a group of people that were trying to build agents back then. One of the leading people that was a member of writing that paper is Andres Karpathy, who was back then working at OpenAI on exactly these kind of tools on agents. In the conclusion of this research, the researchers wrote the following,"We show that while standard supervised and reinforced learning technique can be applied to achieve adequate results across these environments, the gap between the agent and humans remains large and welcome additional modeling advances." In other words, the AI agent were not smart enough to understand things as humans, and it was very hard and probably did not yield positive ROI to train them on basically everything they may encounter when operating within web environments, forms, websites, tools etc. That has changed. So in an interview with Lex Friedman in 2022, Carpathi mentioned that some of the barriers they had when they diagnosed research do not longer exist. And the reason for that the large language models today are so advanced that they understand what they're seeing on the screen, right? The stuff that they couldn't do back in 2017, they get out of the box right now and AI agent that's looking on a website understands the website, understands the buttons, understands the forms, understand that different components drop down menus and so on. And if it's provided with the right tools, it will be able to take actions on any user interface. In an article on the information last week, it was clearly stated that internal sources at OpenAI confirmed that they're working on agents that will be able to take control over our devices and take actions on our behalf. Now will I allow a random tool to take control over my computer or my mobile phone? Right now? Absolutely not. But if that's somebody is Google and they already have access to all my data because I'm using an Android phone and I'm using Chrome as a browser and I'm using the G Suite workspace for everything that I'm doing, they already have access to everything that I know. So to give them access to stuff like that makes perfect sense. If the reward is high enough because the risk is relatively low above the current risks that I'm already giving them all my data. Based on these rumors, this is already at work at OpenAI, which means they will release it sometime in the new future. This may be six months, 12 months, 18 months, but sometime in the next two years, there will be working active agents that you can take advantage of if you want to, which literally changes everything we know. Once there's an AI agent that can literally use a website or any digital tool that you provide to it after you train it. And some of the companies that are working on these kind of tools allow you to train the models literally by showing them what you're doing on the screen. So they become really smart copycats because again, they understand the task, they understand the goal, they understand the required components, they understand the context, and they understand how to use the tool. Another AI field that is making a huge improvement in the past few months is video generation. Whether it's in a simple environment like creating avatars on people and having them speak in any language, in any format about any content with tools like HeyGen and Synthesia and DID or creating any video you want simply by using text or starting with an image in tools like Runway with its Advanced Motion Brush, or PicaLabs, or those of you who have seen the Google Lumiere demo, then that's where the world is going, right? So by the end of 2024, we will most likely be able to create incredible, realistic videos that will look just like real life, similar to what Mid Journey can do with images right now, and we'll be able to create those on the fly without a camera, without lighting, without microphones and without editing because we'll be able to basically tell it what we want the video to be about and it will generate the video for us. In addition to rendering these videos, as I mentioned in the news of this past week, there's just been a fraud for$25 million that was done by replicating in real time CFO and other senior leadership in a video conference call. So the ability to create deep fake videos in real time is currently available to anybody who wants to use it for good purposes. For bad, so huge improvements in video generation using ai. Which takes us to the next frontier. After all these things mature, which is AGI and superintelligence. AGI is artificial general intelligence, which basically means an AI system that will be able to do everything a human can do from a cognitive perspective at or above a human level. So many AI systems can do multiple things better than humans, but they cannot do everything that we can do better than us. They can do some very specific things better than us. And some say we're very far from achieving AGI. I think the consensus across multiple experts right now is two to five years we'll be able to achieve that. Part of the reason in the difference in timelines is that the definition of what is AGI is relatively vague and how exactly we evaluate that. But sometime in the next few years, we will get to AI systems that are smarter than us across everything that we do. And then the next level is super intelligence, which is basically AI systems, which exceed us in everything that we can do from a cognitive perspective. I think very few experts right now question whether that's possible. And the only question is when and how many resources will be required to do that? What's the implications of systems like that? What we're gonna talk about this later on in this episode. Now from ai, we're gonna move to other aspects of technology that is maturing and converging together with AI technology. One of those is Apple Vision Pro. Apple released their vision pro. Now, don't think about this as a pair of goggles, but think about it as a wearable computer that runs in 360 degrees around you what Apple calls Spatial, Computing. This is a whole new generation of computing capability and concept that did not exist before. So if you want, this is another iPhone moment. Now, yes, they're still really expensive and really bulky, and only geeks are going to use them in the immediate future. But think about where this technology is going. What does that mean to AI and the way we interact with the world once the user interface stops being a mouse and a keyboard and a screen, but it's something you can carry with you wear on your head and interact with the real world all the time, while AI agents that now can analyze video, generate video, generate sounds, speak on your behalf, understand everything that's going on. Can view and experience the world together with you in real time. This is an incredible change of how we basically do anything, including not just communicating with computers, but communicating with other people from different places around the world, different languages. All these barriers basically disappear. You'll be able to talk to a person, they will listen in their language and they'll be able to talk back and you will hear them in their language and including analyzing everything around you, using very powerful AI capabilities in real time using glasses and maybe even contact lenses. Now that may sound science fiction, but there's a company called Mojo Vision that back in 2020 had a first prototype of a quote-unquote smart contact lenses that they demo that CES. Since then, they've developed several different prototypes that were even better, and then they run into some financial difficulties. But it means that the technology is going in that direction where we'll be able to put contact lenses and get data projected straight into our eyes and into our brains connected to AI capabilities of analyzing everything that we see. And you can start to understand where this is going. Another technology that has been evolving dramatically and is reaching maturity point is humanoid robots. So many of you probably seen different videos online of humanoid robots, whether it's from Honda or Toyota that presented their robots, or all the dancing robots from Boston Dynamics. Not all of them are humanoids, but they have incredible capabilities to do things like creatures and animals that we know and far beyond. And there's also companies like Agility Robotics, SoftBank Robotics from Japan, as well as obviously recently Optimus from Tesla. All these companies are already generating humanoid robots that can do anything from basic tasks at home as far as assisting the elderly or just anybody who has the money to do all the way to sadly doing military operations and so on. So combined the physical abilities of these robots with advanced AI systems that we're seeing today in ChachiPT and beyond or in AGI and super intelligence. And you start to see the potential for having AI go beyond the cognitive capabilities of humans, but combined with physical superpowers that will allow it to fly, run, climb, move, dance, and do anything that it will want to do better, than we can do on the physical level. This has obviously amazing positive implications, as I mentioned, as far as helping people, rescuing people while taking significantly less risks, and in the long run, probably costing less money. Even though right now these are extremely expensive, but on the flip side can be used for a lot of negative sides because it will have superhuman mental capacity combined with superhuman physical capabilities you can imagine what kind of applications militaries and just bad people can find for systems like that. Now if you think that having Apple Vision Pro, or even contact lenses as a way to communicate with AI systems is the most advanced this can get, I wanna remind you that there's companies out there like Neuralink and Synchron that are creating computer chips that talk directly to our brain. Neuralink is a company from Elon Musk. Synchron is a company that saw investment by Bill Gates and Jeff Bezos. But there's other companies and competitors like Precision Neuroscience. All three companies has already started doing human trials, meaning they're actually installing chips in people's brains to allow them to do different things. Now, while Synchron and Precision Neuroscience are focusing on curing specific illnesses, Neuralink is doing the same in the short run, but in the long run, Elon Musk is saying very clearly that our survival as the human race against AI depends on us becoming super humans. In other words, making us cyborgs that can keep our human aspects, but enhance it with actual computer chips that will be AI driven to provide us similar superpower capabilities as the AI will have. Now, I know what you're thinking. You're thinking there is no way I'm ever allowing anybody to put a chip in my brain, not to mention put a chip in my kids. The reality is if everybody will start doing this and getting quote-unquote superpowers, let's take a theoretical scenario, which I don't think is theoretical. It's just a question of when, not a question of if, let's say all the kids in your son's class now has a chip that makes them super smart that they can know basically everything instantly because they'll be connected to the internet in a seamless way through a brain-computer interface. Will you not put the same chip in your kid? You will have that option, obviously, but then you'll become the new Amish, right? You'll be the one that technologically stays behind while everybody else is making a quantum leap forward which is a gap that you cannot close without having those capabilities. So as much as this is a scary thought, this is where the technology is going, and once technology gets mature, it gets adopted across the board and nobody wants to be left behind. Now to run such things, we will need a huge amount of computing power and we will need high speed internet and connectivity basically everywhere. So the computing power there's been a quantum leap in quantum computers, pun intended, but quantum computers are now a fact. They are multiple quantum computers operating around the world today, they have significantly higher capacity than traditional computers and significantly faster communication capabilities. Now, are they available to everyone everywhere right now? Absolutely not. But can these machines over time drive more and more AI capabilities in order to make it available to the masses through high-speed internet, potentially even satellite-based, high-speed internet like Starlink from Elon Musk, or any other system, there's a Parallels system that Jeff Bezos is now developing, but the benefit of having a system that is satellite based is that you can get high speed internet everywhere. Meaning you don't have to be in a physical location that's connected to cables or that has cellular towers in the area, but it can literally be anywhere on the planet driving AI access to really advanced capabilities that potentially can run on quantum computers that will connect straight to a chip in your brain. This is where the technology is converging too. On the scary side, that's obviously really scary and doesn't sound like it makes any sense to us because we grew up in an analog world. The digital world was just in addition to the analog world. But the reality is the digital world and the analog world are now slowly infusing one into the other. You're gonna have robots that are gonna operate in the physical world, and you're gonna have digital aspects to our being most likely, not too far in the future the other issue that we talked about is obviously power consumption and the negative environmental impact that running these systems have on our planet. AI itself, I already enables the development of new materials. So imagine a world with no plastic because AI will allow us to develop materials that actually decay and do not leave any bad footprint on the planet. But that can be used to replace plastics that AI is already curing multiple diseases, and we will see more of that as AI becomes more and more advanced, and AI can help us develop clean energy. Now, another piece of technology that recently has reached the initial phase of maturity is nuclear fusion, is the ability to create the same process we see on the sun that generates energy right here on earth, and that particular process generates a huge amount of energy without generating any pollution. Now, the idea behind nuclear fusion exists since the fifties, so it's not a new idea, and there's been experiments with doing it since successful experience. The problem with all these experiments was that the energy that was required in order to run the experiment was significantly higher than the energy produced by the nuclear fusion. That was true until 2023. In 2023 was the first experiment that actually created net positive energy in a nuclear fusion experiment. And this was confirmed by multiple scientists in a review that was published a couple of weeks ago. This is another huge step into potentially having unlimited energy without polluting the planet, which means we might have the energy to run all these computer farms and potentially quantum computers that are required in order to run all these advanced AI capabilities that can drive all these things that we're talking about. So what's the impact of all of this on our society? I cannot obviously tell you because I don't know. I don't know if anybody knows, but it's very obvious that it's gonna have significant impact across everything that we do. And I'm not talking in a hundred years, and I'm not talking about 50 years. A lot of these developments, we will start seeing as available for us to use within the next three to 10 years. So within the lifetime of probably most people who are listening to this podcast. Now if you look on the negative side, it has the opportunity to increase loneliness and reduce happiness, which research is showing, has clear correlation to the more use we have with technology, specifically cell phones. and software, like social media that is running on our phones. That being said, there could be a lot of positive impacts of that because it may require us to work less. It may give us more free time to do stuff that we actually want to do, hopefully with people that we wanna do it with. It is very likely to expand our lifespan, whether our actual physical life or maybe beyond that by allowing us to quote unquote download or upload our actual consciousness into a cloud. And I know that sounds like science fiction and those of you who watch the series Upload may think that I'm crazy. But there are thoughts that something like this might be possible sometime in the future. Now the anticipation is that with super intelligence, there'll be huge advance quantum leaps forwards in science, which will allow us to benefit a lot of positive things to the planet, human race, and probably our world as a whole. So if your head is not spinning yet, you may be asking yourself whether what will be more significant? The negative implications or the positive implications? I'm an optimist person and I would really like to think that the positive implications on the maturity and the convergence of all these technologies will significantly outweigh the negative aspects. And hopefully as a society, we will find ways to jointly fight and suppress the negative implications of these things so we can all jointly enjoy the benefits of these technologies and really provide better life and a better outcome for everybody on this planet and potentially beyond. I hope you found this episode interesting. I know it's not my standard episode of how to do very tactical things with use cases. I promise you that's where we're going back to next Tuesday, but the amount of data that I've consumed around all these topics recently and the amount of understanding that I have that this is going to change everything that we know pushed me to want to share this with you. So I would love your feedback on this different episode to see if you want more of those in the future, please connect with me on LinkedIn. Just find me Isar Metis, I-S-A-R-M-E-I-T-I-S. And don't be shy. Let me know what you think. In general, if you find this podcast interesting and if you think it can help people that you know, please share it with those people. This is a great way for you to help them better understand AI and where it's going, and it also helps us grow the podcast and hence be able to provide you even more guests and more resources in order to help you in your AI journey. I wanna remind you that this episode is brought to us by the AI Business Transformation Course, and if you want to learn more about it, go and check our website. And until next time, have an amazing week.