The Interview: Cohere CEO on why democracies need sovereign AI

0:00Hey, Yetis. This is Nick and Jack from the Best One Yet podcast. Now, the last company we worked at, they used Paylocity and everything just worked. It wasn't until launching our own media business, this show, that we realized how rare that is. Because Paylocity is one delicious burrito of operational needs. They roll up HR, finance, and IT seamlessly into one delicious bite. When everything wraps together like that all at once, your workforce, your tech stack, your business, you don't need more tools. You don't even need cilantro. You need one solution. And that is why Paylocity built a single platform to connect HR, finance, and IT with AI-driven insights and automated workflows that simplify the complex and power what's next.

0:36Or as we call it, a delicious operational burrito. Yes, we do. Experience a one place for all your HCM needs, besties. So start now at paylocity.com slash one. Paylocity.com slash O-N-E. Thumbtack presents Tile Trouble. Every single time I use my sink, I make eye contact with the uneven grout of my kitchen backsplash. The crooked corners eye me and I'm haunted by more tiling questions than I thought possible. What kind of pro can fix a backsplash? Can I replace one cracked tile or should I replace them all? What? What if I just use Thumbtack? I can hire top-rated pros, read reviews, and compare prices with a tap.

1:15All on the app. Thumbtack knows homes. Download the app today.

1:22This episode of the Times Tech Podcast is sponsored by KPMG. Katie, we've been tracking the AI surge for a while, but some of the latest data is pretty striking. It is. KPMG's latest quarterly AI pulse suggests that 7 in 10 UK business leaders still see AI as a top investment priority, even if the economy slows down in the next year. But turning that investment into tangible value is another question. And so a lot of organizations are still figuring out how to move from endless AI pilots to something that works across the business. KPMG's focus is on that gap, helping organizations scale from the pilot phase to AI-powered transformation they can trust.

2:04To find out more about how KPMG can help your organization make the difference, you can visit kpmg.co.uk forward slash AI. Why? We're existing in a really consequential period in history where we're getting hit by a massive geopolitical change and a massive technological change at the same time. And the only thing that counts in this period for you as a country and your competitiveness on the global stage and where you end up through all this change is do you have sovereignty? Hello and welcome to the interview from Times Tech, our Sunday conversation with some of the key people who are shaping technology.

2:47I'm Katie Prescott. And this week, I am very excited to say that I've been speaking to Aidan Gomez, the boss of Cohere, which is a very rare AI company for several reasons. It's one of the rare companies that's actually building large language models. But unlike most businesses building LLMs, which are based in America, in China, Kohir was born in Canada. And he has a really particular mission to develop sovereign AI. And he's very, very conscious of where AI is built and who owns it. Aidan has a really fascinating background. He grew up in rural Ontario in the most non-tech environment you can possibly imagine.

3:33Went on to university where he met Professor Jeff Hinton, one of the godfathers of AI, and has since gone on to build Cohere, worth now$20 billion. So he has a fascinating story. So here's my chat with Aidan Gomez from Cohere. Aidan, thanks so much for joining us here on the Times Tech Podcast. I've wanted to do this interview for a while now. So I'm absolutely delighted you're here. No, likewise. Yeah, thank you so much for having me. Just walk us through what Cohere does. For listeners who don't know much about the company, how do you explain it to people? We're a model builder. So we're one of the companies that builds these large language models.

4:14We deploy them not into consumer applications like a chat GPT, that type of thing. Instead, we're focused on critical industries, so enterprises. So think financial services, healthcare, telecommunications infrastructure, basically places where it is the critical infrastructure of the economy and the security of those systems and the data that those systems hold is oftentimes like a national security concern, right? If the grid goes down, that's a catastrophic impact. If the financial system goes down, if the internet goes down in a country, these are critical to the operation and functioning of society.

4:56That's the place that Cohere specializes in. There are so few companies doing this, and most of them are based in the US and China. And you're incredibly rare because you've emerged from Canada. Yeah. Yeah. What made you want to get into this space and also think that you could take on the very, very well-funded American market? Canada has a very long history in AI, similar to Britain. So we have the great British Canadian computer scientist Jeffrey Hinton. We have Joshua Bengio. They won the Nobel Prize for their world. They basically invented deep learning, which is the thing that has given us the entire AI revolution that we're seeing now.

5:38and that was a product of early Canadian investment in R &D. So when I went to the University of Toronto with my co-founders, we were sort of raised into AI. Jeff was teaching there, he had built a lab. Those folks were the ones that really built all of this. And so it was natural, going into AI was just a product of my environment. Of course, a product of my interests. And from a very early age, was fascinated with AI, like everyone else is, you know, reading sci-fi and playing video games and the notion or the idea of intelligence or even like consciousness. It felt like over the past hundred years, we have gotten such a strong understanding of physics and like the natural world.

6:28We can predict things down to like arbitrary precision. We understand it so well that we can know what will happen. But almost none of that has translated into something useful to understanding what intelligence is. And so it felt kind of like the last mystery. And so as a field of pursuit, it was always very attractive. And then I just happened to be born in the place and at the university where it was all happening. Of course, when you say it like that, that is where it was all happening. I'd just rather forgotten that was part of the story because so much of the story now is concentrated on the West Coast.

7:04I mean, do you feel like there is still that rich seam of expertise in Canada today? Oh, absolutely. Yeah. I mean, the legacy carries on. There's extraordinary researchers that come out of Toronto, Waterloo, Mila, Montreal. And yeah, I remember when I went to Google, I was in undergrad and I walked into the HQ in Silicon Valley in Mountain View. and on my very first day, I had just gotten my badge, my computer, I'd gone through orientation. I was walking to like my building where my team, the Google brain team was sitting and I walked past the kitchen and I hear French, Quebecois French, you know, like Canadian French.

7:43And I look over and it's Sammy Bengio, Joshua Bengio's brother. And like, I realized I was in the heart of Silicon Valley in what was at that time the most advanced AI lab on the face of the planet. And I was surrounded by Canadians. They were everywhere. And so I was kind of like, Canada invented this. And I want to talk about sovereignty shortly because it's become such an important issue this year. But let's take it right back to how you got to university in the first place and what drew you into AI because you didn't grow up in a busy city where more tech was going on. Tell us about your upbringing.

8:23Yeah, yeah. I grew up in - Non-technical, maybe. Yeah, yeah. I was born in Toronto, but at one year old, my parents moved into a cabin in the woods that my dad had built. So my dad and grandfather, they're immigrants from Spain. My mom's, she's an immigrant from Britain, from the UK. And my dad had bought this big plot of land, like 150 acres in rural Ontario. And then on that plot of land, he read a book on how to build cabins and started cutting down trees on the property and like stripping them and drying them out and then coating that, like doing the whole thing. And he built a little cabin for my grandpa.

9:05And my grandpa moved in there and he lived there for ages. There was like no running water, no electricity. He just like totally lived off the grid. And then the two of them, my dad and grandpa decided, oh, we're going to build a larger cabin and it's going to be like our cottage. And so then they built a much larger cabin still with like trees from the forest. And then at some point my mom and dad decided, you know, that's going to be our house. And so we moved into this, what was supposed to be the cottage, which was like a log cabin built by my dad and grandpa. There was electricity and running water and a well at that stage, but it was very disconnected.

9:41I didn't have neighbors for like kilometers in every direction. And yeah, you're just in nature, you're in the woods. And we didn't really have any, we didn't have a lot of tech. We didn't have satellite TV. We didn't really have internet. We had dial up forever because there was no high speed. And eventually satellite internet started to get a little bit good. But by that time I was like 16, I almost moved out of the house. So for, for a long time, I was so disconnected from technology. I think it, it developed in me as this affinity towards it. It was like this thing that was like just out of reach that everyone else had access to that, that was like transforming the world, but I couldn't have it.

10:26And so there was this deep appreciation for it. Like the, I guess like being constrained from it or pulled back from it gave me an appreciation for just how incredible it is. And that sort of like fostered the obsession with it. And do you think that's what took you on to university and sort of move away from the cabin and never alive? Yeah. I mean, like I knew then that, you know, what I wanted to work on was technology and computing and AI was a curiosity at that point. But when I started at the university and was sort of like embedded into this ecosystem that was so, had such a deep history in AI, had contributed so much.

11:08It just pulled me in. I got obsessed with it. I was like every waking minute of my day, reading papers, trying to understand the state of the art in AI and what makes it tick and what has caused progress. And then trying to help push that forward, like reaching out to professors. I remember I sent an email to Jeff Hinton and I didn't know who he was. Like he hadn't won the Nobel prize at that point. He hadn't won the Turing award. And I just kept seeing this name on important papers and it was Jeff Hinton and he was a professor at U of T. I was like, Oh my God, he's here. And so I was like, I can email him.

11:43I sent him an email with like my idea. I'm like, Hey, I read your paper on this. You know, I think that's wrong. Like, I think you should have done this instead. And yeah, this is like the hubris of second year undergrad child aidman. And he replied, he responded and he took the time to be like, it's an interesting idea, but I don't think that will work because of X, Y, and Z. And he engaged with it. And then Jeff brought me into the machine learning group in Toronto. I went down to Google. Then I started working for Jeff at Google, met my co-founders, and the rest is history. And is that why you reached out to Google at that particular time?

12:21because you like reaching out to jeff you also reached out to them for a job didn't yes i did yeah yeah again like a lot of these papers it was google brain everywhere google brain google all like the important papers it was like a industry lab that was super well resourced and so they could do experiments that in academia were hard to do and they also had complete intellectual freedom like google the company was not saying hey you need to research these things it was just like let's get the smartest people possible give them some resources and see what comes out and yeah i just started bombarding them with emails not so much to get a job like it was again like this idea would have an idea about their paper and i would see it and be like oh did you try this did you you know maybe we could apply this to that and i would just ask the researchers and they would reply and then eventually one was like oh you know what i would want to work on that i I have an intern slot this summer and that catalyzed the process.

13:20But yeah, it's sort of been like a snowball. And what's it been like, you know, that snowball rolling, but at the same time watching what's been going on at Anthropic, at OpenAI, at Google? The giants that are just, you know, feels like gobbling up so much of this space and building a competitor to that. And are you conscious of that? I mean, of course, I'm conscious of my competitors and not the biggest fans. You know, I respect them a lot, but I think there's a lot that they do that I don't admire. You know, things like some of the fear mongering that has come out of those labs and a certain degree of economic inefficiency that you can see.

14:01With Cohere, we haven't done one of these like mega rounds and we don't do a consumer offering. And so because of that, I think we fly slightly under the radar despite the work that we do is like fundamentally the same as what they're doing. And the other thing that we've done is we've avoided selling huge chunks of our company to the hyperscalers. So those deals, like those are sort of - The Microsoft investment in OpenAI and Amazon and Anthropoc. Exactly. So we haven't done one of those, which has protected our positioning and heritage and the fact that we can be a sovereign champion because we haven't sold off the whole thing to one of the big tech companies.

14:41But it also forces us to exist under scarcity and thrive under scarcity and find ways to make these models more efficient, to build them with less compute, to innovate. And so I think that is healthy. That's how great companies get built is in resource-scarce, high-pressure environments. There are these behemoths, these big guys that are getting funded by big tech. and that's great and they've done a lot of good work and they'll have a big impact but I'm interested in something different like I want to build a generational company and I don't want to sell my company to big tech that's fundamentally compromising to the mission of Cohere like what we want to contribute is sovereignty to the world I was a child during the consumer device revolution.

15:38I was, I guess, a teenager when social media was taking off. And then in university, when I started, it was really when the cloud boom took off. And none of those happened in any of my countries, Canada, Britain, Spain. The democratic alliance was carried by only one, the United States. And that's something that I think is unsustainable. I think technological sovereignty over the past quarter century has not been prioritized. And we've seen our position on the world stage. We've seen the democratic position on the geopolitical stage just gradually weaken over time. If you look at consumer device companies, if you look at social media companies, if you look at the clouds, these are some of the most significant cultural and economic levers that exist on the face of the planet.

16:36These are the potent tools of power on the global stage. And for democracies other than the US, we have none of them. We don't have a consumer device company. We don't have a social media. We don't have a hyperscaler cloud. We've just lost our leverage. It has to change. It's like a threat to democracy? The counter argument to that, which is always put by politicians here is that's all well and good, but it's actually quite difficult to find an alternative to Microsoft, for example, and that sometimes it's better to go with a larger company that we feel is more reliable. I mean, do you genuinely think we will get to a point where we can create realistic challenges to the US businesses that are sizable enough?

17:23I mean, I'm thinking of, for example, Palantir, that's been very controversial. They won a major NHS contract in the UK, but the supporters of that say, well, look, they were the best company for the job. Yeah. I mean, you can take the free market, money's going to go where it's going to go, companies are going to go where it's going to go. That's a strategy that's been pursued by much of the West the past half century. But it's that strategy that's just led us towards is the complete dissolution of sovereignty in technology. And I think that to your question of can we compete, can we build these champions?

17:59It's a choice. We have all the talent we need. We can absolutely afford it. We just need to actually decide to go do it. And I do think like if you look at China, they pursue very aggressive industrial policy to extraordinary effect. They decide we are going to build a capable electric car program. The government perhaps has too much of a handhold on some of their businesses, I think, for a lot of people's liking. Absolutely, for sure. But there's a balance to be struck, right? You don't have to become an authoritarian government to practice industrial policy and develop champions. Do you feel the mood music is changing here from clients, but also from governments that people are looking for companies that are offering sovereign.

18:47Absolutely. We're existing in a really consequential period in history where we're getting hit by a massive geopolitical change and a massive technological change at the same time. And the only thing that counts in this period for you as a country and your competitiveness on the global stage and where you end up through all this change is do you have sovereignty? because if not, you're going to be exploited. Your position is going to be weakened. Those with sovereignty will use it as leverage over you. And so right now, the priority to actually build up that capability, I think is entering into the minds of the folks who have been watching and seeing this and are quite close to it.

19:30In the general public, I think there's increasing awareness. I hope that more people see and start to understand and scrutinize the consequences of the decisions we've made over the past 25 years. The fact that we don't have these champions, that we haven't invested in creating them. It's something that I think has been so underappreciated and neglected, and it's led us into a very vulnerable position for democracy. it's fantastic that the united states has been the torchbearer for democracy and has invested to build all of this technology for us the alternative is a much darker path but we can't just rely on they can't carry the weight of this themselves we're going to have to develop our own capabilities we're going to have to contribute capabilities to the global stage because the alternative is as the world looks i'm talking about the whole world here not just the democratic alliance, but as the whole world looks to purchase technology, they need a resilient plan for the future.

20:34So they can't buy all of their technology from the United States. They will need a second rail to run on fundamentally. And at present, the democratic bloc has almost nothing to offer. There's no second rail. And so where do you have to go? You have to go to China. That is becoming the second rail. Which is happening, isn't it? People are finding that it's cheaper and some of the models are just as good oh yeah i mean like in the model we can talk about electric cars we can talk about the modeling game chips now as well they're catching up but that future doesn't have to come about where we have diversification beyond the u.s onto chinese systems and technology it could be another democracy it could be like british tech it could be canadian tech and german we can create this technology we have all the talent that we need to do it.

21:22And despite what people say, we can afford to do it. We are huge economies. Extraordinary. We have the money to do it. We just need to actually decide to do it, have the willpower to begin a program. What China's done is a continuity of strategy over decades that has led to this extraordinarily strong position that they're in. There's a commitment to, I will develop a capability in X. We need to have that capability as well. We need leadership and a public that is committed to building technology and investing over the decades required over different administrations, different leadership to actually see through that project.

22:18to roll out across a large organization. Issues like governance, workforce readiness, and trust tend to become more important as things scale. KPMG's approach is to look beyond AI as just another technology rollout, and instead, how it could redesign work across the organization and fit into wider processes and decision-making. So it's less about individual use cases and more about how organizations can transform and deliver outcomes powered by AI. To find out more about how it can help your organisation make the difference, you can visit kpmg.co.uk slash AI.

23:17you live healthier. Find out more at vitality.co.uk. While every other channel is fighting for your customer's attention, podcasts are where they've already given it. No one accidentally listens to a podcast for 45 minutes. They choose to be here. They trust the voice in their ears. And when that voice talks about your brand, it doesn't sound like advertising. It sounds like a recommendation from a friend. Acast gives you that trust at scale. Digital precision, post-read authenticity, and performance data that proves it worked. Don't fight for attention. Buy it with ACAST. Learn more by visiting acast.com slash advertise.

24:04You just said earlier that you haven't gone in for the sorts of mega rounds of funding that some of your rivals have, but you have recently done a mega merger. Would you tell me about that and how it plays into this sovereign strategy? because it means now you've got headquarters in Canada and Germany. Yeah. Yeah. So we're thrilled about it. Cohere will become a Canadian German company. And it's part of our strategy of bringing an alliance among democratically values aligned countries to build a capability in this space in AI. So we were fortunate to have the partnership and willingness of Alfalfa to merge with us.

24:39They've been focused on something very similar to us, like sovereign deployments of AI, extremely secure, privacy restricted, focused on critical industries, focused on government in the same way that we are. And we were able to navigate a broader cooperation between Canada and Germany at the governmental level to support this digital alliance between Canada and Germany. That's so fascinating. So it does feel like the geopolitical shifts are being reflected in what you're doing. I mean, you're half European, You're saying your father's from Spain, your mother's from the UK. You live in London, you've got a natural affinity here.

25:14But there's also something going on between government. Yeah. I mean, it's a moment. Like Canada, we haven't seen our sovereignty threatened this way before. And so there's broad public understanding of how significant it is to invest in it. And at the same time, we recognize that we can't go it alone. We have to partner. if we don't have strong allies, if we don't integrate ourselves closely with our friends, we won't exist as a country. It's so fascinating hearing how politically you talk about this. Politically? Yeah. As in, you speak to very few business leaders who also have such a strong political mission.

25:56I mean, unfortunately, it's political. Politics plays into it. I do think It's hard to avoid. For instance, these models are extraordinary cyber systems. They can hack into systems better than anything we've created in the past. That has huge geopolitical consequences, huge defense consequences, national security consequences. And so there's no pretending that it's just a private sector, like, oh, we just do business. It has to be something more than that. And I do believe that technology is going to be the most consequential to national competitiveness of any technology we've seen before. So whether more than one democracy controls it and influences it is going to be exceptionally consequential to the future.

26:47Looking back, I mean, the time that you spent at Google Brain and when you were at university studying this spending time with jeff hinton did you think things would accelerate as fast as they have done or i mean is there any sort of surprise there for you yeah huge surprise huge surprise like where we are today if at the end of my life whatever 80 90 years old we got there i would have been a very happy old man you know i would have been thrilled like what uh What progress, you know, the fact that we've gotten here in like 10 years is a huge surprise to me. And is there anything you will now worried about?

27:30Am I worried? Yes. The short one word answer is like, yeah, of course, I'm worried about many, many different vectors of potential risk. I think labor is one of them. Another one is the cyber impacts of this technology. biological weapons is another one that i'm increasingly concerned with i haven't been in the past but increasingly so on labor i'm trending more optimistic at the moment like there's this saying like jevons paradox if you yeah so when like an efficiency enhancing technology comes out like intuitively you'd be like okay everything is easier to do you can do it faster you need less people to do it.

28:12And counterintuitively, what happens instead is like the price falls, demand skyrockets, and you actually end up needing more people. It creates more jobs as a result. And that seems to be happening. Like unemployment numbers are not exploding. And in fact, like automation came for software developers first. Like what's funny is like, we kind of automated ourselves first. Like that was like the most substantial place of progress in AI automating a job. And yet demand for software engineers is still blowing up. So I think Jevin's paradox is proving true at the moment. We'll see if it holds up though, I hope.

28:49And I really believe in Coeer is doing work to support exploration of what if it doesn't hold up? What if there is actually disruption in the job market? What are the tools that we have available to help ease society through that? All the way from extremes like UBI, right? So you don't have to work anymore. You are just paid a salary for being a citizen of the country. To softer schemes like right to work or job guarantees, and then even softer schemes like retraining. I think we need to be researching this and experimenting with them to understand what actually works, what's effective. Thank you so much.

29:26It's been really great to meet you and finally kind of put a face to the story that I've heard so much about. And thanks so much for coming to the Times Tech Summit in November. Thank you for having me. I'm looking forward to speaking to you then. Thanks for listening to this week's big interview on the Times Tech Podcast. I hope you enjoyed that as much as I did. See you next week.

29:51This episode of the Times Tech Podcast was sponsored by KPMG. One of the themes we keep coming back to is the gap between AI ambition and real world application. And that's where KPMG is focusing its work, helping organizations move beyond pilots and think about how AI can change the way the business operates and how it fits into wider processes and decision making. You can find all their insights and how they can help your organization make the difference through successful AI workforce adoption. Visit kpmg.co.uk slash AI. When everything is moving all at once, your workforce, your tech stack, your business, you don't need more tools.

30:36You need one solution. That's why Paylocity built a single platform to connect HR, finance, and IT with AI-driven insights and automated workflows that simplify the complex and power what's next. Because when everything comes together in one place, growth comes easy. Experience one place for all your HCM needs. Start now at paylossity.com slash one. Do you keep hearing podcast ads like this one, for example, but always wonder how you actually get involved with them for your own brand or organization? Well, it's easier than you think. We're Acast and we give you the platform to do it all yourself.

31:13Browse thousands of popular podcasts, Choose the shows that match your perfect audience, set your budget, and launch. And if you want a hand, our podcast specialists are there to help you launch with confidence. This is podcast advertising without barriers. Get started at acast.com slash advertise.