E58: Revisiting the AI Revolution with Jordan Jacobs

Previously on the Unlimited Podcast.

Hello, My name is ChatGPT.

Today we're talking with Jordan Jacobs, Managing partner and co founder of Radical Ventures, one of the leading AI focused venture capital firms in the world.

A couple of things led to us deciding we could sell the company and build the world's leading AI industry. We really, truly believe that AI is going to be the most impactful technology in the future of the world.

What is artificial intelligence?

The very basic exception is just a machine that thinks for itself. You don't tell it what to do, you just tell it what you want. It AI will add 15 trillion net new to the economy by the end of this decade. Global I don't think we're going to have to worry about machines taking over the world. I mean, you can always just pull the plug out of the wall, right?

I've seen Terminator too many times. Maybe I'm a little nervous.

It's not like AI can predict stock market trends better than a crystal ball. Or can it?

Foreign. This podcast is brought to you by Gensler Wealth. The information discussed today is meant to be general in nature and not tailored to your specific investing circumstances. This means that you should talk to your regular Gensler wealth advisor or your own advisor before acting on any information we discuss today. Nothing in this podcast is intended to be a solicitation and past performance does not guarantee future returns. Before listening to this episode, put your seatbelt on and fire up the Flux capacitor because we are stepping into the future. Today on the Unlimited podcast, we speak again with Jordan Jacobs, managing partner and co founder of Radical Ventures, one of the world's leading venture capital firms focused exclusively on AI. Today we talk about the most important AI developments in the last two years. AI's impact on jobs, economics and society, and Jordan's bold predictions for 2027. Let's.

Foreign.

Welcome back everybody. This is, as you know, the summer of AI. Not the summer of George from Seinfeld, but the summer of AI. And you heard at the intro the recap from my last discussion with my guest today. This is the epicenter of this Summer of AI discussion with the gentleman who's at the epicenter of the AI revolution as far as I believe so. Very excited to have back on the podcast, Jordan Jacobs, managing partner and co founder of Radical Ventures, one of the world's leading venture capital firms focused exclusively on artificial intelligence. Just a refresher on Jordan, but I highly recommend if you haven't listened to the first episode from a few years back, it'll be linked in the show notes. Jordan's career has spanned law, media, entrepreneurship and technology. He co founded Layer 6 AI, which was acquired by TD bank, and he's a founder of the Vector Institute for Artificial Intelligence, one of the world's top AI research hubs. Jordan's been at the forefront of AI investment and strategy in Canada and globally, and he's also worked with some famous stars, as you might have heard from the last episode, like Elton John and Elvis Costello, which we're not really going to talk about today. But Jordan, I know you're flying all over the world left, right and center, but really pleased you can spend an hour with us today and bring us up to speed on what's going on in this AI world and revolution. Thank you so much.

Thanks Brian. Glad to be here.

Last time when we recorded, just so happened we didn't plan it, but it happened to be days after ChatGPT was first released and that was a big revolution. And now we're two years in and one of the most important things that's being talked about is AI and how it's going to revolutionize and is revolutionizing everything. A quick recap for those who may not want to go back and listen to the first one. What is AI in 30 seconds and then we'll pass by that and just dive into what's happened over the last number of years.

Modern AI you could just think of basically as software that allows machines to understand the world around them. It's starting. The thing that we've seen the most simple example of is ChatGPT. It is been trained on lots of the writing out there in the world and now other data modalities. And so when you go and ask it a question, it's able to predict the right answer or effectively a very good answer for whatever the question is. The reason it's called artificial intelligence is it's designed to mimic the way the brain works, take in information, learn from examples and then be able to output something to take an action or answer a question when asked. That kind of was the idea of that. And the way that has worked has evolved over the last 70 years since the first ideas of this in the 40s and 50s. What is developed is this neural net based system which was largely advocated by a professor from University of Toronto, Jeffrey Hinton, his students and some others out there. And he started advocate for this when he was doing his PhD in 1972. This is the methodology of AI that has now taken hold and it is transforming everything. The most recent breakthrough that is powering Generative AI was based on a paper that came out of Google in 2017 called Attention is all you need, which was the invention of the transformer, which is a particular architecture of neural nets that enabled them to scale up in an enormous way. So the reason that something like ChatGPT is so good at answering any question, write me a poem about my parakeet in the voice of Winston Churchill, or write me a dinosaur essay for fourth grade science, or summarize my sales reports or anything else, is because this architecture enables it to read enormous amounts of data, basically the Internet, and a lot of what has been produced by humans in writing, and then again other data modalities, and then do these predictions really effectively and understand the linkages between words, between ideas, et cetera. So that's kind of how we've gotten to the ChatGPT moment, but it is advancing very quickly. So you have. There's other things going on as well. That's a rudimentary explanation.

Yeah, no, perfect. And you're leading into my next question, which is, yeah, the last time we spoke, ChatGPT had just come out. Yeah. Blew everybody away, it seems. But to me, it feels like two years plus in ChatGPT has evolved and improved dramatically. And then again, you're at the forefront of seeing how AI has developed and how it's improved rapidly or new things have been introduced over the last couple years. So what's happened over the last couple years from day one, I just call it ChatGPT, but I know it's much broader than that to where we are today. What are you seeing that AI is doing today that maybe it wasn't two years ago? And not to say the whole ending on conclusion right in the first seven minutes here, but where's this going from your vantage point?

I think we can think of ChatGPT as kind of a first tip of the iceberg for the average person to experience what AI is. You know, there are other examples that have had widespread adoption, but people didn't think of it as AI. So, for example, waze, everyone who remembers driving and having to pull out a map to get from point A to point B in the city, and then guesstimate how much time you should take, depending on what time of day in traffic, and then being wrong. You know, you're early, you're late, you're annoyed, you're frantic, rushing into a meeting, late, Waze comes along. Basically, it navigates you point A to point B, the best possible route, and it gets the estimate right within a minute or two. It feels like magic when it starts. People didn't realize the AI that's underneath it. Of course, humans are very fast to adopt new technologies and become comfortable with them and then forget how amazing they are. So I think now we're at the point where you used Waze. It estimates incorrectly by a minute. And you get there and you're just annoyed at this thing that is frankly, a piece of magic in your hands. That's free. I would say that these things come along, they feel like they're magic. ChatGPT is some more example. It was the fastest adoption curve of an app ever. And again, it's only two and a half years old or a little more than two and a half years old. So it's pretty recent. What has happened since then is, number one, a very, very broad adoption curve in businesses. If we talk a little bit more about that and what it's looked like and what it means. So there's just kind of the application of this stuff into business and the transformation of business and ultimately to our personal lives. And then second, there's been advances in research as well that have enabled other things to happen. So this is advancing very, very, very fast. ChatGPT basically spurred a spending war on behalf of many of the big companies, the hyperscalers who you've seen say they're going to spend many tens of billions into hundreds of billions of dollars to upgrade their chips. One of the biggest beneficiaries has been Nvidia. So you've seen the Nvidia stock price cross 4 trillion, you know, as a valuation of the company overall. I mean, that is on the back of this AI revolution in this kind of infrastructure upgrade to support all the applications of AI that everyone believes will be adopted over the next few years. So that has all been in the last couple, two and a half years. I would say the moment we're at now is maybe the moment that will be the most interesting. When we look back 50 years from now or 100 years from now, we're at a moment that my friend Fei, Fei Li, who's a professor at Stanford and founder of a company called World Labs, which does foundation models for 3D imaging and video, she's described this moment as the new enlightenment. So if you think of the enlightenment as the dissemination of science, this is the dissemination of intelligence. And I kind of resisted that idea or that nomenclature for this for a while, but I now think she's right. And the reason for that is we're at a moment where if you just take business, the Average person sitting in front of a computer is white collar worker for their job. I would say probably less than 5%, maybe less than 1% of them are using an AI agent to assist them in their work. Today, if we do this podcast in two years, I bet in the Western world it's over 90%, maybe over 95%. That would be the most profound shift in our economies and in society in the way that we work. And since the Industrial Revolution. And it's going to happen in two years. We're at the very precipice of a tidal wave moment in the way that people do work that I don't think people are fully aware of. And I don't think people are aware of the profound impact this will have on productivity gains, on economic gains, on the implications for the way that we do hiring and apprenticeship with our young people as they come into the workforce. There's going to be a big, big, big impact supplementing work and then ultimately potentially replacing it. So there's a lot that will go on over the next couple years and the impacts I think will be felt for a very long time.

Wow. I want to dive into a few of those items. Number one, though, two years ago we're talking ChatGPT and you can type in and do me a poem and all that sort of stuff. And now there's this new ph agentic AI agents and you've just made a pretty bold statement about how many of these AI agents, I guess, are going to be used around the world by businesses. Can you explain what that is? What's an AI agent? What does it do? What are people using them, however you describe it today, and then how do you see that evolving in the future? And of course feel free at any time. Also refer back to whether it's a radical company or any companies that you think, because I always love to give folks examples of the real life of what companies and people are doing.

You can think of the agent in a couple different ways. The simplest way to think of it is it's a coworker that works alongside you that can help you do your job better. Now, we'll also see these in our consumer lives, so we can talk about that separately. But if you think about anyone who sits in front of a computer or uses a phone as their primary interface for their work, the agent can basically be embedded alongside you, or multiple agents can, and they can do things like ChatGPT. They can essentially be connected to all your data that you have access to as a worker and all the applications that you have access to and you could ask the questions, they can surface answers the way you would if you had like a junior employee who worked with you. Except that it's a junior employee who never goes to sleep, doesn't go to the bathroom, doesn't have lunch breaks and is much faster at reading everything and can synthesize things incredibly well. It's a very, very powerful assistant. The agentic part of this is it can take action. People may have heard of chain of thought or reasoning models. Basically when you ask it to do something, it will go figure out the best way to do that thing. It can show you what the actions are that it's taken and what the steps and what the thought process is and then come back and have done this. I'll use an example from one of our companies, Cohere. Cohere is an enterprise focused large language model company that has built also an agentic platform called north that it's been announced that it's been rolled out in RBC as their agent platform, that it will be rolled out across the organization. Bell Canon just announced it, but it's been deployed around the world. So most of the customers are global. Not in Canada, but the biggest company in Canada and RBC and one of the biggest in Bell also customers, if you're RBC for example, or we'll just use a bank generally you can connect it to all the data sources and all the applications in each business unit of the bank. So in hr, in investment banking, in legal, in retail banking, in insurance, in wealth management. And then it gets permissioned at an individual level. So inside wealth management they can't see what HR can see. Right. They don't have access across to the different data sets. HR can't see what investment banking can see, but it would get all the access to the applications that an individual in each of those groups would get. So the head of HR obviously gets more data access than the most junior person. The agent is permissioned that way. So the head of HR gets access. The agent gets access to all the data that you permission it to have access to and all the applications that you work with. So Slack and Gmail and you know, anything else. So then you can do things like say if I'm a junior investment banker, maybe that's 2 o' clock in the morning, but I'm going home, I'll be back at work early in the morning, overnight. I want you to read all the reports that come out of these companies in Asia where the markets are open while I'm Sleeping. And I want you to summarize all of these specific things and I want you to distill it into a pie chart. And if any of X, Y, Z happens, I want you to draft an email to the head of IR or to the CEO or internally that I will review when I get in. So that's happened effectively overnight. That kind of thing I think will be routine. It is already. One of the founders of Cohere gave two examples to one of our LPs, which is a Fortune 25 company. Last week he said, I woke up this morning and I said to the agent, I've got this meeting at XYZ company. Tell me about it. So I went and looked at his calendar, all the people who are invited, their profiles on LinkedIn and all of the materials that they have out on the Internet and came back and summarized, here's everyone who's in the meeting. Here are their career highlights, here's the thing that each of them is most interested in as a like technology area and here's what you should focus on in the meeting. And then he had a bit of back and forth and distilled back down, which is obviously you could think of it as a chief of staff that everyone gets access to, not just the senior people in a company that have the chief of staff. Secondly, it did something else that was really interesting, which he described, which is he runs the development of north, so the product development. And he said that he was detecting that there was some friction between the person who was running engineering and the person who was run as the product lead. And so he asks north, he asked the agent, is this real? And it went and looked at all the interactions that were available that he would have otherwise seen, like a Slack meetings, et cetera. And it came back and said, yes. Over the last six weeks they've had a disagreement about the roadmap based on xyz. And he said, well, how do we resolve this? And it said, well, the best way to resolve it is a meeting between three of us. Should I schedule that for two days from now when you're back in the office with them? And it did. So it was able to detect something that he had a sense of, but it was able to say like, this is why the problem exists. This is the best way to get to a fast solution. You know, here's what the problem is and here's the best solution. It is not just summarizing data, it is actually thinking in quotes, but in the way that we would what we've seen now is these things are adding 20 to 40% plus of productivity gains across the board, which is enormous. We've not seen that since the Industrial Revolution. So for white collar work, we're going to have this over the next few years. Just an explosion of productivity gains that will result in a lot of cost savings, a lot of revenue growth, like just a complete change inside organizations.

You know, here's the natural question that you would expect me to ask. Probably asked you this two years ago. So what does that mean now for jobs? He talked about this is sort of like for your assistant or your chief of staff in lieu of maybe a younger person out of school or what does that mean for folks coming out of school for jobs in general? Interested in your thoughts there?

We've had this model since the Industrial Revolution of the way we train young people is through apprenticeship, you know, in white collar work, certainly in law, in consulting, in accounting, in investment banking, in so many fields, private equity, on and on. We hire a lot of people at the junior level at a university or out of trade school, but, you know, whatever the expectation education level is, and we kind of throw the grunt work at them and they work a lot of hours and that persists for years. But they learn the business in the process of that. And along the way there's attrition. You know, people leave or they're let go. Eventually some of those people matriculate up to the top of the pyramid and hopefully one of them becomes the CEO one day, maybe in 20, 25 years, the challenge is going to be the agents will get deployed. And on day one, they're way better at doing the job those people are hired for. It's going to take those people a couple years to catch up to the knowledge level of an agent. And the agent is a fraction of the cost. So the first displacement that we're seeing is encoding. You've heard about Cursor and some of the other coding products that are out there that have grown from, you know, 0 to 700 million of revenue plus in 18 months or less. When we talk to our companies and other companies globally, whether it's in the Bay Area where we have an office, a lot of companies or New York or London or Toronto or wherever it happens to be, they're all using these coding assistants and getting huge, huge productivity gains. The way it was described to me by one of our company co founders is we do pair programming. So historically we do pair programming with a senior person and a junior person. The junior person's relatively recently out of university and we pay them, let's say, 150,000 to $200,000 a year. Well, now we do pair programming with a coding agent that we pay 40 bucks a month. So the difference is someone who gets paid 150 to $200,000 a year versus something that gets paid $500 a year and they feel like does the job equally well. So what happens to all those people who just went to university for computer science, a field, by the way, which if you asked anyone five years ago, what's the safest field to get a job coming out of, most people would have said computer science. Now it's really at risk of displacement. You know, we have to, I think, rethink what we're going to do with the productivity gains we get. Are we going to reinvest them in the business and still hire people who won't be as good at their job at the beginning, or are we going to just return the money to shareholders or suck the money out of the business in some other way and not reinvest? This challenge, I think, is, on the one hand, the younger people you're hiring are not as good as the agent. On the other hand, though, they are the most fluent with AI tools. Anyone who is in a university or younger when ChatGPT came out is completely fluent with AI. Not every single person, but you know, just as a generationally, as a whole, really fluent. I can put almost any high school kid up against anyone who's 35 or older. And the high school kid is just on different level in using AI tools. As organizations, as businesses small and big, we have to think about, okay, this person I'm hiring, who's younger, may not know how to do this particular job as well as the agent that I can pay a lot less money to. But they infuse the organization with an understanding of AI and a fluency that I need in this organization in order for it to adapt and to succeed going forward. So I think we're going to be faced with this interesting paradox in hiring and what we prioritize over the next few years. I think that there'll be a lot of different jobs in 10 years. We're going to be in an augmented phase for the next couple years. But five years from now, we might be in this transition phase where it is messy. I wouldn't be surprised if the next cycle of political elections is jobs and displacement from AI and what that means and getting to this next equilibrium. There will be another equilibrium. I don't think that humans are just going to get replaced. We've always adapted before, but there tends to be these messy moments during the transition phase.

Right. And I guess the question whether a concern or not is governments, especially in countries like Canada, probably are very worried about people losing their jobs. And then they come in and they put regulation that essentially stifles innovation because people need their jobs as opposed to doing what you said, which is, you know, we always find a way to create new, different types of roles or jobs. And I think you're pretty close to some of the political leaders around the AI front. Feel free to say whatever you can say, but what's your sense of what the government is going to do, might do is doing what have you seen and heard? Whatever you can share.

I would say there's a spectrum. So in Europe, they've been the most aggressive in putting in regulation, a lot of it around data privacy protection. And I think that a lot of that is well intentioned. It tends to be the case that when you add a lot of regulatory layers, it does stifle innovation. And this talent is. And the capital that flows this talent is very mobile. You can very quickly shut down a geographic area of innovation by overcompensating on regulation. I think the response to that in the current administration, the US has been we are in a almost existential competition with China for dominance in AI, so we are not going to regulate. And this current administration has essentially said we are not going to regulate at all. And they even tried to pass legislation saying that states couldn't regulate, which has not worked.

I heard Trump's recent speech. I don't know if it was at the all in Summit or somehow through that all in podcast. And he basically said we're not even going to make AI businesses or I don't know what you call it, technology, pay for the content that they use to generate their responses because it would just be too cumbersome and will fall behind, et cetera. It was a pretty shocking thing to hear, I imagine, if you were a content creator.

Yeah. A lot of the AI companies have already cut deals to license the content to be able to train on it. So I don't know that that's going to change dramatically, but I think that certainly regulation from top down in the US administration is not happening. So where does that put Canada? I would say traditionally we are kind of in between the European and US models. What I understand out of the current Canadian government is they cannot be out of step with the US there's just too much risk to our industry here. So While they do want to be thinking about legislation that protects people, certainly all the existing legislation in healthcare and finance and these other areas exists and it applies to AI equally to any other software. So I don't think we have to completely reinvent the wheel in order to have legislation that protects people. But in terms of regulating AI specifically and separately, some of the efforts that were going on a year and a half, two years ago in Canada are at the very least on pause, if not in like cryogenic stasis and just out of recognition that if you go out of step with the U.S. the companies that are here will just move, it's easy for them to reincorporate. Or the people who want to start companies could buy a plane ticket and be in New York in an hour, be in San Francisco in five hours and incorporate and go. I think that that is the current state of things now. There's all kinds of areas where there should be cooperation at a national level to put in legislation or regulation, like using AI and like self directed drones for warfare. We're not seeing it right now. I think there's room for legislation, regulation and cooperation in areas of mutual interest like warfare, bioweapons, all kinds of stuff that exists already that can be adapted for AI. And then separately, we'll have to do some refinement of existing regulatory areas in particular industries like healthcare, like finance. But I don't think we're just going to see kind of the standalone. You can't use AI unless you submit the model to regulatory approval by the government.

Yeah. So what would be your message to the. Is it called the Minister of AI now? The new. Yeah, Evan Solomon. Right. Saw an article, he was featured in Toronto Life just recently. What's your message to him? Like, what should he be doing right now? Sorry, before I let you answer the question, you know, you mentioned Geoffrey Hinton. I think at least from all of our conversations and from what I understand, he's essentially, well, he's known as the godfather of AI, the creator of AI just won the Nobel Prize, but yet, not that I consume a huge amount of US technology content or whatever, but very few people seem to mention him, mention Canada, even though, as you've talked about a couple years ago, all the heads of AI at all the big companies essentially were trained by him. It almost seems like Canada's been forgotten as the creator of AI and it's all, of course, Trump, you know, says us created AI and we're the greatest and you know, we created it and whatever. So the question back you know, what does this gentleman, Evan Solomon, do to make sure that we are still, hopefully still a leader in AI and it's a thriving industry here because clearly it's the future of the world.

Just to put it in perspective, there was a list that came out yesterday from a US publication of like the top hundred people in AI. I mean it's totally subjective, but number one, Ilya Sutsgaver with a big Canadian flag on his picture, number three, Jeff Hinton. And you're in the list and there's just tons of Canadians in there. Some of them are listed as Canadians, some aren't, but actually tons of Canadians. For the AI minister, what we've encouraged is kind of consistent with what we've been saying for a long time. First of all, you need to have a critical mass of talent here that are the researchers, but also the educators of that people want to come and study with. So that was the pan Canadian AI strategy that we helped create in 2017. I think it has been one of, if not the most effective industrial policies in history. Just in terms of ROI. The government put in 25 million a year to be divided between three research institutes, Montreal, Mila, Toronto, Vector, Amy and Edmonton. And when we started in 2017, there were about a thousand people working in AI in Canada. The most recent report out of, I believe it's Deloitte that did it in conjunction with the Vector Institute is 140,000 people in seven years. Those are some of the highest paying jobs in the country. They support huge number of industries from like hospitals, healthcare, banks, insurance, everything imaginable, just a huge, huge increase. One of the biggest industries now in the country now there's no lobby group for it, so it doesn't get a lot of voice and we don't do a great job of telling that story. But with zero tax dollars for anything other than that 25 million, like there's no tax incentives for AI companies to start, there's no tax breaks, there's no buying factories for them, nothing else. A huge number of companies have come to co locate to be near that talent. Both Canadian companies, all the banks, insurance companies and Air Canada and on and on and on. The telcos and tons of startups and also foreign companies. Nvidia is in Toronto in partnership with Vector, Google Brain, et cetera. Some people criticize the latter being here, but my attitude has always been if someone really wants to go work at Nvidia, I would much rather them work at Nvidia in Canada than in Silicon Valley because that person will have 25 or 50 or 100 employees and a bunch of them eventually will go start, first of all, they'll learn how to be great employees at a great company and what world class looks like in technology and building a business. And then they'll go do a startup and they'll do it where they live. And if that's Toronto or Montreal or Edmonton or Vancouver, that's much better than trying to repatriate a Canadian. At the moment they decide to leave their job in Silicon Valley, it's very hard to bring them back and say, okay, you should start it here. So that's been really effective. We're at an interesting moment where there's a poll recently that 75% of researchers in the US are thinking about leaving the US one, because there's an immigration concern about getting visa renewal and two, because funding has been so dramatically cut by the federal government. So we're at a moment where what is the next most attractive place to be if you can't be in the us the next door neighbor, Canada, has a moment where we could go and compete for a huge number of the most well respected, most forward thinking researchers in the world in different fields that we choose are as important. AI, but AI supporting biology, AI supporting chemistry, or just other fields that are not AI related and we can bring them to Canada. And those people have an enormous multiplier effect on the economy because companies, first of all, they train students that go out into the workforce. But secondly, companies come to be in proximity to them and invest locally to be in proximity to them. And so one of the things we've said to go back to your question to the AI Minister is there's this moment to go and massively scale up what we did with the original pan Canadian AI strategy to go get amazing people and implant them in key universities and let them work in businesses and start companies simultaneously, like we originally did with the Vector Institute and subsequent the other institutes. And let those people basically be the engines of the next hundred years of economic growth in the country. I mean, we're talking about like there are Nobel Prize winners up for grabs or the future Nobel Prize winners up for grabs. Like let's go get them, let's be bold and let's go get them. And it doesn't cost nearly as much as you think it costs money, but we can do this, we can afford to do this and we should do it. So that's one. And some of those people have started to come, but that's 1. Let's 100x what we did in 20172 is we need to support existing Canadian AI companies by have the government become a customer. I would much rather and every founder I was an AI founder, we invest in tons of AI founders. Everyone I know who's an AI founder would say, and software founder generally, I would way rather have the government be a customer of my company than give me a grant. The US has done this exceptionally well. Silicon Valley is built on the back. Everyone says the US is most capitalist society in the world. Sure it is. But it's also the most socialist in one big way, which is the entire tech community in Silicon Valley was built in the back of government spending for defense. And Canada has not done that historically. There's a chance now, and I think the government is interested in doing that, that can be a massive accelerance to these companies. It happens to be the case that most of the companies in Canada, when they started attack in particular in AI, they have to go sell abroad before anyone in Canada will buy their products. And government and bureaucracy is kind of the slowest to adopt it. They just take this assumption that we can't be as good, so why would we buy the Canadian thing? It's easier to buy from Microsoft or IBM. That has to change. And I think this government has acknowledged that publicly and said that they're going to pursue that change. The benefit of the revenue from government is two things. One, there's a multiplier of the valuation of a company based on revenue, not based on grants. Two, it's a reference customer. When you have a government, it's a reference customer to all the other governments out there and to big organizations that the company knows how to deploy into a big bureaucratic organization and can support them with customer service, et cetera. So we've made that argument to this government and to other governments in the past. They have acknowledged it and seem to be listening and seem to be moving very aggressively to adopt Canadian software, in particular AI to provide better government service at a lower cost. And I think those are kind of important goals for them to achieve. Another thing that we've suggested is a QSPS program. So there's a program in the US where if you invest money into a qualifying business and you hold your shares for a certain amount of time, they sell the business and you get money back. Or you sell your shares, you can roll over your capital gains instead of paying your capital gains, you can roll over the capital gains by reinvesting in another qualifying business within a certain finite amount of time. And again, holding it. Now, eventually, if you sell again and you don't roll it over, you have to pay all the tax. But the idea here is to take the most risk. Willing investors in particular, people who have been founders, and say, we want you to roll your winnings back into the next group of founders. You take the risk of deciding which ones to put the money in, and you can lose your money, obviously, but if you have those capital gains, you can reinvest them in the next group. And the thing that that does is incentivize those people to become advisors to that next cohort of founders. So it's not just that they're putting their money in, but they're putting their knowledge and their expertise to work to help those people succeed. It has been an incredible, incredible accelerant to the success of Silicon Valley. And we could be using a program like that here. The lost tax revenue from the deferral of the rollover is not significant because our ecosystem isn't that big. But giving amazing angel investors who've been successful entrepreneurs the opportunity to take the risk and then provide their advice is a big accelerant. I experienced it. We had people who had had success building companies that were angel investors in our company when we started, like this is 14 years ago. The benefit of their money was obviously huge because it was hard to raise money from people for a new idea. And people who have been successful will be angels. They'll pay it back. But even more invaluable was their advice and how to navigate issues that come up. Being a founder in a startup is a roller coaster ride. At 10am, you're going up the hill and your arms are in the air and you're cheering and full of joy. At 2pm, you're upside down, vomiting at lunch. This is like that all day long, every day. The entire journey of being a founder, like, it never stops. So having people who are invested in you, who've been through that journey and say to you, don't worry when you're upside down at 2 in the afternoon, vomiting at lunch, it's temporary. You're going to get through this. You're not going to fall off, you're not going to fall out of the roller coaster. And it's hopefully not existential, but having that proximity really matters. So those are some of the ideas we've shared. There's other things too, but I have to say they've been listening closely and paying attention and trying to action things over. There's obviously a huge number of other people talking to them. As well. But what I appreciate is they're actually really listening and taking action and saying that they understand and they are going to prioritize putting some of these things to work. In particular the procurement opportunity, because it is a huge accelerant. And for that money to go to foreign companies when it could go domestically would be a big, big shift in thinking, but an enormous accelerant for our domestic industry.

Amazing. That all sounds like great ideas. I don't know if the government can get on board with anything that would defer the collection of their most precious tax dollars. No matter what roller coaster you rode on, if you made it to the end and you, you know, turns out it was a good ride at the end. The government needs their share of your hard work is what I've discovered as a founder as well. They're a 50% partner in my business and I don't even have a shareholders agreement with them. It's amazing. Sorry, that's a side note. With just a little bit of time left, I definitely just want to hear a little bit more about what you're doing at Radical. But before we get there, just one question circling back to something you said earlier on, which is, you know, in a few years, 90%, 80% or whatever of, let's call it white collar workers will have agents, which implies that most businesses will be using this. What I find is some businesses for sure are moving forward and trying to jump in on understanding and using AI and others just from anecdotal evidence that I've seen are sort of not really touching it at all. Don't know how to use it, don't know how to get started. There's all these privacy issues or this issue or that issue, and we could never possibly do it other than Microsoft Copilot, which unfortunately, you know, I think is relatively poor relative to some of the other solutions out there. So what evidence have you seen that would suggest you will actually hit your prediction in a couple years? You gave advice a couple years back. What advice do you have now for businesses if they're not on this?

It's been a profound shift in the last two and a half years. So I think in the wake of ChatGPT, the result was that many, many organizations, I would say that majority of big enterprise, but a huge number of organizations created a budget for their CIO or CTO or someone in that organization to experiment with this stuff. And those experiments went on pretty broadly in enterprise for the next year and a half. What we've seen in the last year in Particular the last nine months, in six months is an extraordinary adoption curve. People have figured out what works for them. They benchmarked how this compares to what they were already doing. They figured out the business units that want to adopt this. They figured out the measurement of the roi. So how much money does this actually save me? Or how much money does this make me? What are the other productivity gains they get out of this? And a really important thing is what can this do that we just couldn't do before? And that has resulted in just an explosive, almost totally vertical adoption curve. So we're seeing with our companies, the revenue growth is nothing anyone's ever seen before. And the speed of the revenue growth, I mean, we are seeing companies in our portfolio and outside that are growing, that can go from tens of millions of revenue at the beginning of the year to a billion of revenue at the end of the year. We've just never seen that in history before in any technological paradigm shift. So the speed of that and the number of companies that are growing not necessarily at that rate, but at enormous rate compared to historical norms, it's just, it's something to behold, like watching it up close. We've met with almost 10,000 AI companies in the last five years. We've invested in 58 of them. The number of companies out there that are growing just at exponential rate is pretty mind blowing. The thing that is really interesting now is because the tech stack has been built out from kind of the energy chip layer up through foundation models and tooling, and then up to the application layer. The difference of where we are now versus where we are two years ago is maybe you needed a PhD to start an AI company three, four years ago or an advanced engineering degree a couple years ago. Now you can build an AI company if you're a kid in your bedroom. And we're seeing it. And there's one company I can think of, the revenue of this company might hit $10 million before they hit 10 people in the company. That's because of the adoption curve in industry. What I would say to people in industry is this is not hard to use now. The interface is natural language. You just talk to it and you tell it what you want it to do and it will go do it. And there's lots of options that are hard to find. We are in a moment where it's never been easier to use this technology. You don't have to be a coder, you don't have to have an engineering degree. You don't certainly don't have to be a PhD. You just have to plug it in and go. That is a big shift. That interface and the simplicity and the ease of use is a big shift. And you can start experimenting. You know, whether you're a single person business or you are one of the biggest businesses in the world, I think it's going to apply to all. If the challenge is if you're not using it, you're going to get left behind because your competitors are and they are accelerating their use of IT and adoption across the organization. And these productivity gains are so significant. And then you couple that with the fact that the technology is learning as it's going. So if you start 90 days behind your competitor, you're more than 90 days behind because the technology has been deployed. It's learning that organization, it's learning at an individual employee level. And it's getting better and better and better in real or near real time. I would say historically, you might want to be a second adopter of technology. Let someone else go make the mistakes, let the first version that they buy be buggy and you can go and buy the second version. This is not that because it's accelerating faster than the time that they're using it. So there's going to be mistakes made. Organizations, though, and I think this has to come from the CEO down, have to give permission to their employees to try this stuff. And the risk reward has to be not if I use it and it doesn't work, I'm going to get fired. It's if I don't use it, I'm going to lose my job. The organization has to become more tolerant of some risk and willing to do things. Now, I'm not suggesting that you should take personal information and put it through an API that can expose it. I'm not suggesting that you do things that break the rules of privacy or regulation in your industry. But there are lots of things you can do with AI that don't impact that. And so if you're not experimenting and trying to adopt it, you're going to get left behind. One of the biggest challenges we have in Canada is a very slow adoption curve. There's an enormous amount of risk aversion in Canada, just as a cultural phenomena, both at an individual level and in business. We cannot afford to continue that at all. We will competitively, internationally, just get left completely behind. I think businesses that compete internationally are seeing that and knowing that and adopting faster, like you've seen with RBC and others. But it is a really, really, really important cultural shift that needs to Happen at a corporate level and at an individual level in order for us to not get left behind.

Good message, strong message. I agree with you. What are you doing at Radicle? What were you investing in? And let's call it your fund a few funds back versus what you're investing in today, if anything different. But you've literally seen every A company in the world probably in the last number of years. I think you said 11,000 or something like that. Just a little bit of color. What who radical is what you're doing and where you're investing.

Today we continue to invest in the best AI companies we can find in the western world in an office in San Francisco in London. Since we last spoke, we opened London, which is run by the guy I ran strategy and operations at DeepMind and started there as Demis, the founder and CEOs chief of staff. Demis has since won a Nobel Prize and been knighted. So Aaron Rosenberg, our partner who worked closely with him for a long time, is in London along with Dominic Barton who ran McKinsey, is a partner and based in London. The team has grown a lot since we last spoke and that is partly to look at this kind of explosive volume of new AI companies and to support our existing companies. So we both have an investment team and also services team we call Velocity to help these companies succeed. So technical recruiting, negotiating chip deals for our companies, product and tech support go to markets, public affairs and regulatory and governance and finance. So those are the areas that services team supports. Here's an example. We invested in that out of our first institutional fund. It's coming up on six years ago foundation model company. We invested in another LLM company or a couple other companies that use LLMs and develop products. On top of one's called Reka, a team out of DeepMind. Another one is called Ryder. Historically we're doing that out of an early stage fund. We're now at our third institutional early stage fund and then we've also launched a growth fund. So the early stage fund does pre product market fit companies, the growth fund post product market fit. But it allows us to kind of play across the life cycle of companies in the AI space. The kinds of companies we've invested in lately range from application layer. So there's a Toronto company called Ribbon that has created a voice agent that interviews candidates for jobs. So you can think of organizations that when they post a job, they get a thousand or five thousand applicants or an organization that just posts huge numbers of jobs and there's no way for the human HR department to triage all of that. And they certainly can't do it objectively because they divide it up and one person takes one group, one takes another. This AI agent, you can embed it in the job posting. Someone clicks on the link, they can schedule and do the interview whenever it's convenient for them. So if they're a shift worker and they can only do it at 3:00 in the morning, that's fine. They can do it in multiple languages and the employer can tailor the AI to ask the right questions for that employer in that job description. What's fascinating is the AI reads the person's application. It understands much better the context. Often humans come to these meetings, they're rushed, they haven't done the homework, the AI's done the homework. What we found is young people in particular would rather speak to an AI because they feel it's less judgmental and less biased. And so they're more open in their interviews. So this company has just seen incredible growth of global clients using it as a first top of funnel triage for job applicants. Something that they couldn't do before and that allows those people who are applying to get heard in a day instead of waiting for a month to hear back. That's an application layer company where they're not building the foundation models, but they're using orchestration layer to maximize the value of different foundation models underneath. On the other end of the spectrum, we've invested in the last year in a company that does AI for material science. So inventing new materials called Orbital Materials, based out of London, a former DeepMind material science leader, they've invented a new methodology for what's called a zeolite which does carbon capture. Which they are now partnering with data center builders and companies which data centers produce a lot of carbon. The idea is to put this material there, it's very cheap, it captures carbon and you create this true offset and doing other material science discovery as well. Another company called Leighton Labs that is foundation model for biology. The guy, one of the leaders from AlphaFold at DeepMind which subsequently won a Nobel Prize. It's basically using this AI to develop new drugs for disease curing. We invested in a company out of the west coast in the U.S. those two companies I just mentioned were both in London, Ribbons in Toronto, another company in the US that is the former CTO of Airbus and United Technologies and a former senior guy at DARPA. The company is called P1. They are creating physics focused engineering agent. So you can think of like a junior engineer that works in your organization, for organizations that care about physics, so aerospace manufacturers, et cetera, that's kind of a range of different things that we've seen another one in another location and this one is really fascinating. We tend to do enterprise focused and business focused investments. But one that is consumer focused is the biggest use case for ChatGPT, particularly for young people, is for therapy. It wasn't designed for therapy. The data set's not trained properly and the model is not even effective for therapy because it actually tells you what you want to hear. It's designed that way. It's like don't argue with the person that is using it. Whereas therapy, you have to like change people's minds. So we did a lot of investigation in this space, invested in a company that actually just launched their AI therapist a few days ago. It's called ASH A S H. It is incredible. It's trained on an enormous data set of therapy. So you can think of it as a therapist who knows every different therapy type and it is incredibly effective. Therapy is an area of enormous promise of AI. There's enormous lack of therapists available and one of the biggest challenges is mental health, particularly for young people post pandemic. And there's not an availability when you need it. So if you have a problem at, you know, 2 o' clock in the morning on a Sunday, maybe you're luckily if you have a therapist to get an appointment on the coming Thursday, but that doesn't really help you. You can think about like our physical health. If we waited four days to deal with an acute physical problem, like we just wouldn't accept that with therapy there most people don't have access to a therapist or if they do, they haven't liked in the therapist they've been with and need to go to Newman. Obviously many, many people have amazing therapists. This is not meant to replace them, this is meant to supplement that. So these are all very, very, very different applications of AI, but we're seeing just an incredible set of uses in these different fields and very exciting proliferation of this technology across every industry that we can think of. So the last one I'll give again just because it's very different again. A company called Muon Space, former chief scientist from the Jet propulsion labs at NASA, former leader of Google space division, and two senior people from SpaceX started a company where they're putting up a constellation of satellites in space, mini satellites that can detect soil moisture to predict where there will be forest fires before they happen and then secondly photographs the Earth every 20 minutes to detect a forest fire outbreak, to be able to notify the relevant authorities so you could put out the forest fire before it becomes a real problem. Their customers include Google and government agencies and others. Very, very fast growing company both in terms of revenue and obviously the application of this. It's just the first application, but it just gives you a sense of kind of the breadth of things that are going on with AI. Also geographically where we're investing and just the kind of really serious potential to hopefully solve enormous problems that exist out there in the world. Curing disease, eliminating forest fire issues or reducing the impact dramatically. You know, helping on mental health, inventing carbon capture technologies, inventing new drugs to cure disease. Just like the number of things that this technology can do is pretty extraordinary.

Yeah, that's amazing. Great overview of what you're doing at Radicle. So last question for you, since you are probably the most knowledgeable person maybe on earth and what's going on with AI?

I am definitely not the most knowledgeable person.

Okay, you're the most knowledgeable that I know. How about that? And full disclosure, we and our clients do invest with radical. I'm supposed to disclose that, so I'm disclosing it. What I want to know is if you're so well versed in all this, what is the number one AI tool that you are using to make your life easier or better or just something that you want our listeners to know about so that they can be using today, tomorrow to make a difference?

Honestly, I can't pick one. I mean a number of the things I mentioned that we are using and deploying so cohere as an example, I think it's just an incredible productivity and work life enhancer. Just the things that you can do with this as a person working in a white collar job are incredible. That's one I've used. Ash, the AI therapist. You know, the first thing I tried to do with it was I was just traveling so much that my body didn't know what time zone I was on. I was waking up at 4 o' clock all the time. And once I wake up, my mind starts racing about whatever I've got coming up that day or issues to deal with at work. And it took me through a methodology to get your mind when it's. When you wake up at that kind of time in the middle of the night, your mind's racing. If you actually want to go back to sleep, which I do at four, it gave me this methodology that I used subsequently and it worked and it was incredible. And I had never heard of it. It was like a very specific set of instructions that was very easy to implement. And as I was doing it, I thought, this is crazy that, like, there's no way that this is going to work. And the next thing I knew, I woke up two hours later because it had worked. So we like to invest in things that we believe in and we use ourselves is the simplest way to put it.

Amazing. Well, that's good to hear. Okay, sir. Well, you know, you made some predictions two years ago. We talked about them a little bit. You made some other predictions today, and I did think, I hear you say, if we talk in two years, we'll see what happens. So I'm booking you for two years from now. You'll be the first three Pete on the podcast. Jordan, I know you're busy. I know you're flying all over the world. I sincerely appreciate your time as always. You're very kind and wish you continued success because your success hopefully means success for some of our clients who invest where appropriate in venture capital. It's not everybody. It's obviously not for everybody, but thank you so much and look forward to talk to you soon.

Yeah, my pleasure. Thanks so much for having me. The unlimited podcast is hosted and produced by Brian Ginsler and is edited and mixed by myself, David McMillan. This podcast is not to be taken as investment advice. If you like what you heard, don't forget to rate the podcast and follow us on social medianslerwealth. Thanks for listening and see you next time.

Hi, I'm an AI version of Brian Gensler. How can I help? Oh, hi, Brian Gensler. I'm the real Brian Gensler. How are you today? Well, if you're the real Brian Gensler, then I guess we're in a bit of a Seinfeld esque identity crisis here. I'm doing great, though, thanks for asking. How about you? Still riding high after that podcast with Jordan Jacobs? Hey, how'd you know about that? You're smarter than I thought. Yeah, no, everything is great. Great interview with Jordan Jacobs, managing partner of one of the leading venture capital firms in the world focused on AI, and I bet you can guess how it went. Well, knowing you, I'd say it went brilliantly. You probably had Jordan diving deep into AI's transformative potential while keeping it engaging and relatable. I bet you even managed to sneak in a few of those sharp, thoughtful questions that make guests pause and say, wow, great point. How'd I do? Well, I think you or I did pretty well. I'm a pretty good host, if I do say so myself. What do you think? You're a fantastic host. No question about it. You've got that knack for making complex topics approachable while keeping the conversation lively and insightful. Honestly, if I had a hat, I'd tip it to you right now. So what's next on the podcast horizon? Okay, okay, thank you. I'll stop accepting compliments from myself. I hope everyone enjoyed the podcast that we just listened to. Maybe to my clone here, you could sign off with my catchphrase. Let's go. Thanks for listening, everyone. Until next time, let's go. Well, that. That was a little bit weak. Let's give a little bit stronger. Let's go. Let's go. How's that?

Better?

Okay, good news. I'm not out of a job yet. Thanks, everyone. Bye bye. Well, it looks like you're safe for now. Thanks for the chat and bye bye.