Cohere's Aidan Gomez: The Canadian Founder Taking AI Models To The World, Not Just Big Tech

0:00There's been a sort of slowdown in growth for these developed democratic economies. When that starts to taper off, you see the emergence of things like xenophobia, like that immigrant is taking my opportunity, they're taking my slice of the pie. A lot of people talk about like AI destroying the world. I really don't think AI is going to destroy the world. It might be like our best shot at saving the world.

0:27Aidan Gomez:Welcome to the Upstarts podcast, our weekly show where we talk to emerging startup leaders about their upstart moment. Upstarts are challengers who punch above their weight and take on the status quo to improve the world, all while building a big business too. I'm your host, Alex Conrad, founder and editor of Upstarts Media. And I'm delighted to be joined today by Aiden Gomez, co-founder and CEO of Cohere. Thanks for joining, Aiden. Thank you for having me. This podcast is brought to you by Mercury, banking redesigned from the ground up. Aiden, you have run Cohere now for six and a half years.

0:59Aidan Gomez:You guys have raised$1.7 billion. You work with some big governments, some big companies. When you think about sort of what Cohere looks like today, is it early innings? Is it everything you hoped for six years ago? I think we're in the takeoff phase. So we've been around for a while, but this past year was really the breakout year for For us, in terms of enterprises just adopting AI at scale across the entire company, and the importance of all the things that Cohere is good at really sort of hitting the front of mind of executives at companies all over the world. So privacy, sovereignty, and this focus on critical industries.

1:39And so last year, all of that came together, and we sort of started our takeoff. And then this year, we expect it to get even bigger as enterprises and the broader global economy really, truly adopts AI at scale.

1:55Aidan Gomez:So with some of these big companies or G7 countries that are working with Cohere, what is the problem that they want you to solve? Or where is your technology actually helping them in sort of a simple overview? view. Yeah. Well, it's super broad because we're working with a really diverse range of customers from different industries, from like manufacturing to public sector to financial services, energy. And it's as simple as like, you know, get the model to every morning at 6am, check my unread emails and send me a text with a summary of like the most important ones that I need to know. So that's like a super general, you know, any company would want that.

2:35To a hyper-specific, you know, within a specific ship, I'm going to plug north into the telemetry within that ship, do continuous monitoring on different sensors on board, and build automations around if sensors cross certain thresholds, what to do, who to notify, what sort of remediations.

2:56Aidan Gomez:But I know it's a wide range of use cases, but generally speaking, why often do they need Cohere to exist? Like, why is that an unsolved problem for them? Yeah, I think it's the privacy piece. I think the secure deployment of this tech, of this platform, is something that our customers couldn't do without. It's something that is a hard requirement in their settings. They can't send data out into a cloud. They're highly regulated. They're in these high security settings. And so they need something that's deployed on-prem or even air-gapped on their infrastructure. There's a deployment-level differentiation, and the other piece is within the product itself.

3:37It's super easy to use. It's very customizable. There's tons of administrative enterprise controls that are unique to the North platform.

3:44Aidan Gomez:So should we think of Cohere as sort of the base layer, this infrastructure layer that then the company is either building internal apps on top of or it's running these agents doing sort of all sorts of other work with sort of the info that your horsepower is providing? Or what would be the best way to sort of visualize this? Yeah, I think it's infrastructure for these high security enterprises to adopt agents and integrate it into their secure systems. That's what we do. At Coher, obviously, we're internal customers of our own product. And many of the best automations have been created by or discovered by folks who don't code.

4:22And they just, as part of their day-to-day work, they're like, oh, I could totally use North for this. And they build an automation and they share it with people. And so that's what we want to be able to enable is people just discover killer applications of the system. And then it goes viral within an organization and spreads out to all these different teams. and North really enables that.

4:43Aidan Gomez:We are in a moment where we're being told you can Vibecode anything and there are some really successful companies that are helping people come up with an app out of an idea. Is the difference here that we couldn't really run a Bolt or a Lovable or one of these tools internally in a sort of sensitive or secure environment with all our data so using North is just sort of much more viable or what would be sort of the distinction there? Yeah, I think the vibe coding stuff is very helpful for technical folks who want to get 90 % of the way there, or even 95 % of the way there. It can carry you quite far, but there's still a last mile that you need to close.

5:27And oftentimes, you do need to be technical to close that last mile. And for north, you don't need to be technical to do that work. You have a sort of like builder view and you can just chat to the model and there's no need for you to be a developer yourself to do the work. So I think that's one thing. The other thing is that for a lot of these applications, you don't need to create a product. You don't need to create a whole new piece of software to solve it. All you really need is like it to have access to the right tools with the right authentication. And then you need to be able to build a series of steps that you want it to accomplish.

6:03and that's a much simpler problem and there's an easier way to solve that rather than for each one of these that I want to create I'm going to have to create an entirely new code base and program and ask an agent, coding agent to go build it for me. So I think it's a more simple and elegant solution to the problem.

6:23Aidan Gomez:Very cool. There's an interesting question of are startups reflecting the personalities or sort of backgrounds, preferences of their founders, or is it vice versa? And I'm curious, is this the kind of business that you would have always imagined running or building? Yeah, I don't know. I don't know. I definitely appreciate my own privacy as an individual, but I try to be open. I think for people to trust you, they have to know you. And so I talk about where I grew up, how I grew up, my family. So I try to be somewhat open, but I'm certainly, I'm not the type, I'm not a great marketing or sales guy.

7:04Like I'm not out there screaming, you know, my own name or coheres name constantly. My background is like, I'm an academic, right? Like I like came into this through research. And I think there's a very different type of personality that thrives in research than in company building. And I've been, you know, doing my best to adjust to being a CEO and having to be a spokesperson and a salesperson for the company. But you can't run from your nature, which is like, I'm a, you know, shy nerd.

7:39Aidan Gomez:So that doesn't go anywhere. Just a very tall. It's tall, shy nerd. Yeah, got it. Well, so when you were in academia, you were privileged or you had earned your way to be at sort of the epicenter of some of the earliest work with LLMs, you know, this famous Transformer paper. Could you talk a little bit about how you first, you know, got into this field, why it was exciting to you, and sort of how the journey really starts there? Yeah, I grew up in the middle of the woods in Canada without, like, any internet. At some point, we got, like, a satellite dish, because before that it was a dial-up, but this was, like, pre-Starlink, and so it was, like, these, like, geostationary, like, if there's a cloud, because it has to be like super high up.

8:21If there's a cloud, you're screwed. Like goodbye internet. And I think that scarcity from technology, like it was just so hard to use my computer and like I couldn't watch YouTube. It was impossible. I could like watch pixels load one by one on an image. That scarcity made it super appealing to me. It was like just out of reach. And like, you know, my friends were gaming online and doing all this cool stuff. And when I went home, we had a computer in the house, of course, but it wasn't the experience everyone else was getting. So that made me deeply appreciate computers and be very excited about them, invested in them.

9:07I was always trying to make my computer faster at home or let me access the things that other people were accessing through my limited setup. Was it working?

9:15Aidan Gomez:Were you hacking a faster computer? I mean it got me into like web dev and that type of I wanted to understand why this fucking web page wouldn't load faster and so you know I inspect element blah blah but yeah that's what got me into coding like first it was just like going through websites and trying to understand all this text on my screen and then I took a course because I was interested in that then I started a little small business in my tiny little town of Brighton Ontario building websites for the little businesses like there was a really nice lady who had a knitting uh store and you know yarn and patterns and stuff and so i helped her build like an online catalog for her customers and you know she's been an incredible customer of mine but i started that little business and then i just decided this is what i want to do so i went uh into computer science uh at uft that was my local university right like toronto was the nearest big city and i just happened to go to university at like the epicenter of deep learning like jeff hinton was there the guy who won the noble prize uh and the turing award for his work creating deep learning which is like the thing that worked it was like the breakthrough system that let us reach where we are today.

10:41Aidan Gomez:And were you immediately like, this is an area that I want to throw myself into, or was it more grounded? I was obsessed. I just remember every waking minute of my day thinking about it. I would walk around with a stack of papers. Initially, to read one paper in the field would take me a month, actually, because I would be going through and I'd run into like a word that I don't understand. And then I would have to go start the process of, you know, researching what the hell is that? You know, like what's Bayesian or, you know, what's a Gaussian? You didn't have LLMs to short-cut for you. No, no.

11:20It's just Wikipedia pages and like research papers. And then gradually I just got more and more familiar and comfortable with the material. And I went to a startup in Vancouver where I worked on, it's called polyphonic transcription, which is like listening to music and transcribing music into sheet music. So you listen to someone playing and you write down the notes that you're hearing. Super difficult problem at the time, but we used neural networks to approach it and got some pretty good results. And I had to, this was before, you know, before Jax, before TensorFlow, before PyTorch, before any of these frameworks really existed.

12:02and I had to write kernels in metal, which was like Apple's GPU language for forward and backward propagation. So I just had to learn everything at like the lowest level imaginable.

12:17Aidan Gomez:You're like, you know, basically writing machine code. Not really, but, you know, just super low level, stupidly low level. And then I came back to Toronto and I messaged Hinton and I was just like, hey, I have this idea, this activation function that you've made. I think there's a better one. And he actually responded. And then he sent me in the direction of Roger Gross and others inside the U of T ML group. And from there, started doing work with them, eventually ended up at Google down here in Mountain View and was on the Transformer paper. And the rest is kind of history. Then I went off to Oxford.

12:56I met my co-founders when I came back from the Valley to Toronto. Ran into Nick and Ivan. Nick was working for Jeff at Google. It was like Jeff's first report at Google. And then I went off to Oxford to do my PhD. And then from Oxford, I called up Nick and Ivan and just said, guys, we have to start something. I saw the first few results of language model scaling. And it was very much a zero to one. You know, we could barely get these models to string a sentence together to this was like plausible, fluent, ostensibly human-written text coming out of a machine. And now we take that for granted. It's so normal.

13:39It's like this is kind of not surprising. But back then it was like it was surreal reading a machine. I think people sort of take for granted the fact that like we're talking to some rocks and metal. We're literally talking to our pet rock at this stage, which is, it's so insane. It's normalized. We do it every day. But good God, like back then it was just mind-blowing. And so like, what else could you want to work on?

14:10Aidan Gomez:We talked about the Transformer paper, which was called Attention is All You Need. I don't know what the best comp would be. Maybe the Magna Carta of AI research. Very seminal paper that you were a part of. For folks who are not deep in AI world, what would be the TLDR of what this paper was? Yeah, well, it was done when I was at Google. And in particular, it was done as a collaboration between Google Brain, which was like the AI research department under Jeff Dean, and Google Translate. And the two teams came together. And really, the objective was to improve translation, improve the state of the art of translation.

14:50The architecture itself was designed to better and more scalably model language. And what I mean by that is, you know, language is a sequence. It's a sequence of characters. It's a sequence of words in order. The order matters. And you want to be able to model that or represent that data in the most efficient way possible. So the transformer was, that was the problem we were trying to solve. How do we just make this a more efficient model for language? Turns out the transformer became extremely effective beyond language. And you can phrase many different data modalities like video, et cetera, as a sequence.

15:31The fact that the model was so efficient made it a very good option as we were going through this process of scaling up models. And they're called large language models because they're really big. the transformer as an architecture just scaled better than the alternatives because it was so efficient and so as we went through this process of the field building bigger and bigger models training on larger and larger computing clusters it was the best thing available and so people poured effort into optimizing chips for it and coming up with better uh you know libraries for training these models and then over time the transformer just kind of took over everything It took over the domain of images, took over the domain of audio, it took over pretty much the whole field.

16:17But it certainly was not designed for that. It was just designed for this one narrow problem of, like, we want to make translation 3 % better.

16:25Aidan Gomez:And now people are saying, what comes after the transformer? No, I've been hearing that for a long time now. Yeah, I don't know. Like, I've gotten so excited before. I thought SSM, state space models, which was like this thing that came out three or four years ago, I thought that was going to kill the transformer. I get excited. Discrete diffusion models. But would that be like killing your own baby? No, not at all. No. If you're a researcher, you really hope humanity comes up with something better than what you've created. Otherwise, progress has ended. When you published the transformer paper, the group of you, did you think this is going to be, like, we know this is going to be huge?

17:04Aidan Gomez:we're all going to end up rock stars in this category now. I mean, we're like the world's most niche celebrities. There's like 13 people that know us. But no, we didn't know the impact it was going to have. Or at least I didn't. I don't want to speak on behalf of all the Transformer paper authors, but we can't take credit for what happened with the Transformer. Like we put it out there. Our effort was around improving translation performance. That's the thing that we set out to solve and to come up with an architecture that was better suited to language and sequence modeling. And I think we did that.

17:43But then the community took it and applied it to everything. And they've taken it so far. They've optimized the hardware for this specific architecture. Sort of set the ball in motion. but the work has been done by thousands of others who have actually turned this into what it is today. We still call all these models transformers. I think, I'm not trying to be like overly modest but if it wasn't us, something else would have come out that kind of looked and felt the same. There were ideas along these lines with like WaveNet and ByteNet and all these different architectures that were emerging at the time.

18:21I'm so grateful that the community leaned in on the transformer but someone else would have done it three months later.

18:28Aidan Gomez:I do think that sounds kind of modest to me, but I respect it. I think it's true. I think it's true. Modest, but true. On your entrepreneurial journey, so it sounds like at Oxford you're like, okay, we've got to build something here. Would you have kind of always had in the back of your mind, maybe I'm going to do a company, or did it need to be the right moment, the right idea, to feel excited to really go into company building? So Ivan, my co-founder, he had been messaging me being like, man, like, you know, I'm kind of done working where I'm working. He's working at some other startup. I want to do a startup.

19:05And I was in the middle of my PhD, so I was like, good luck. But I was like, dude, I really don't want you like wasting, like a company is such a huge commitment. I don't want you wasting, you know, years of your time on some stupid idea. Let's come up with a cool idea together. And so we were sending each other ideas back and forth. We had a bunch of ideas. And then eventually there was this one which I think precipitated Cohere. And it was train a model of the web. And we didn't know how that was going to be a business. But we had seen these, like internally at Google, there were results on modeling Wikipedia with language models, produced incredibly compelling text.

19:49GPT-2 came out. We saw the result of training on a much broader source of text. And so we just said, okay, what if we treat the entire internet, like all the pictures, the videos, et cetera, as a data set to train on? I bet you something interesting is going to fall out of that. And that eventually is sort of what led to Coher. We started to see these results like GPT-2, and it just gave us the conviction that there was some direction of travel here which was going to be pretty important. We went from a world where you couldn't speak to computers in natural language and certainly they couldn't speak back to you in natural language to one where that seemed like a plausible future.

20:32Aidan Gomez:As a year-old startup, we experienced a lot of firsts at Upstarts Media. So when it was time to process our first international wire following a London event, I braced myself. This is going to be a painful lesson. Instead, Mercury made the process easy. Issuing an invoice, sharing routing numbers, and processing the wire with just a couple clicks. No phone calls, no paperwork, no learning curve. We're still figuring plenty of stuff out. When it comes to our banking, Mercury's already got it all figured out for us. Visit Mercury.com to learn more and apply online in minutes. Mercury is a fintech company, not an FDIC-insured bank.

21:08Aidan Gomez:banking services provided through Choice Financial Group and Column NA members FDIC. What was so exciting about the idea or the mission of Cohere that you were like, let's commit a lot of time and opportunity here? Because you predate and saw, for example, the team leave OpenAI and start Anthropic, and then all of a sudden Anthropic is growing super fast and Cohere is continuing along the journey. Why not just be like, we can do a better OpenAI ourselves or let's go for sort of the most obvious piece of the pie here? Well, the attractive thing for us was the technical project. Like that's what was the most exciting.

21:49We wanted to train this model of the internet. That was really the objective. We didn't initially know how we were going to monetize that. Quickly, we did decide we want to be B2B. Like the impact that we want to have is driving this technology into the economy. That was a very, that was like within the first 12 months decision that we made. And I still think that is like the most compelling application of this technology. That is like the best impact we can have is by driving growth for the world and driving growth for the global economy. If we can make people more efficient, that in itself is a good.

22:29A lot of the issues that I think are very prominent right now, have sort of been stewing over the past 15 years. I'm from Canada. I live in the UK. The same is true for Japan, Korea, much of Europe. There's been a sort of slowdown in growth for these developed democratic economies. When that happens, when GDP per capita or pick your metric, when that starts to taper off and the pie isn't growing for everyone, you start to see like zero sum thinking start to dominate. If the pie isn't growing, then my slice isn't growing unless I'm taking it from someone else. And vice versa, I feel very threatened that someone else is eating into my slice of the pie.