


Canada’s government invests $1.6B in Cohere, framing it as “AI sovereignty.” The episode explains Cohere’s enterprise-focused large language models (ROI over AGI/superintelligence), why Ottawa chose Cohere as a national champion, and critiques of the deal.
Nick Frost (Cohere co-founder; also frontman of a Juno-nominated rock band). Evan Solomon (Canada’s Minister of AI). Murad Hamadi (reporter following Cohere; Globe and Mail/previously The Logic). Jim Balsillie (former BlackBerry CEO; chairs Council of Canadian Innovators). Also referenced: Dominic Barton, Jordan Jacobs, Patrick Pichette (linked to Cohere and public AI policy).
Cohere builds secure, efficient business LLMs for regulated sectors (e.g., telecom, banking, government) rather than consumer chatbots; “sovereignty” means control over models/data, not isolation; Cohere avoids frontier-style safety harms by deploying inside customer environments. Critics argue sovereignty is undermined by reliance on non-Canadian infrastructure (e.g., CoreWeave), deal terms are opaque, and the bet is overly concentrated/political.
RBC analyzing quarterly earnings; filing expense reports; Anthropic shutting off access to Mythos for foreign nationals; open-sourcing a smaller coding model; Artificial Analysis ranking Cohere model near bottom (118/129); OpenAI lawsuits and alleged harms; government funding for “sovereign compute” routed to CoreWeave.
Written by AI. May contain mistakes.
Canada's investment in AI and its implications for sovereignty.
“We are not waiting for the future of AI to be decided somewhere else by those we don't know, or by whom we don't trust.”
The founding and development of Cohere in the AI landscape.
“Why has Ottawa chosen them as a national champion?”
How Cohere builds language models differently for business use.
“Yeah, in some ways it's an odd question.”
Cohere's approach to AI safety compared to competitors.
“Because unlike the other big AI companies, they don't make a general-purpose consumer-facing chatbot.”
Exploration of what 'sovereign AI' means for control and independence.
“When the independent organization Artificial Analysis created an intelligence index of 129 large language models, they ranked Cohear's best model near the bottom at 118.”
Explore the concept of sovereignty in AI and its implications for Cohere.
“It was a tiny 30 billion parameter model.”
Examine criticisms regarding the government's investment in Cohere.
“In 2002, Ryan Wedding was an Olympic snowboarder at the top of his game.”
Understand the reasons behind the government's decision to support Cohere.
“So I'm all for the strategic initiatives, but you cannot commence it like this because it's not going to work because you're not doing it right.”
Explore the rationale behind Cohere's focus on practical AI models for various sectors.
“And fundamentally, it seems to me like what Cohere is trying to build is really good enterprise-level business software, which is great.”
Discussion on the nature of the Canadian government's investment in Cohere.
“But they are very much at the frontier of cybersecurity about testing.”
Examining how Canadians can benefit from the government's equity investment in Cohere.
“Like what does that mean that it's a commercial investment?”
Analyzing whether the $1.6 billion investment in Cohere prepares Canada for the AI future.
“I tried asking Nick Frost about this too.”

Get the gist, explore an idea or ask about something said.
On Front Burner, what are the key points of this episode?
Automatic transcript. May contain errors.
0:00Hey, we're Nick and Jack, hosts of The Best One Yet. Behind every episode, there's research, email, spreadsheets, and our shared Google Doc, the Headline Hammer. With Google Workspace with Gemini, Google's AI is built right into the premium versions of tools we already use, like Gmail, Docs, and Sheets. No new tools or extra logins, just personalized AI for the way we work. That means less time on the work around the work and more time finding the stories worth talking about. Get AI tools for work. Upgrade to Google Workspace with Gemini at goo.gle slash workspace with Gemini. That's go.gle slash workspace with Gemini.
0:33This is a CBC podcast. This morning, Canada is taking a major step forward in artificial intelligence. Hi, I'm Jamie Poisson. We are not waiting for the future of AI to be decided somewhere else by those we don't know, or by whom we don't trust. At the end of October, Prime Minister Mark Kearney went to Montreal to make this announcement. We're building it right here in the right way. I'm proud with Minister Solomon, my colleagues, to announce that the government of Canada is investing$1.6 billion in Cohere. $1.6 billion. It's believed to be the biggest stake Ottawa has ever taken in a single private tech company.
1:21But the people behind the deal say that This isn't just an investment in an AI startup. It's an investment in our sovereignty. Because today the most powerful AI systems are concentrated in a very small number of companies in a very small number of countries. That concentration carries risks. If you only want to regulate, but you don't have any control over building, you really can't do much. You need to be a builder. I think this technology stands to empower people the way they felt empowered when they first got a PC, when they first had access to the Internet. But it has to be deployed correctly, and that's what we're doing.
2:00But not everyone is convinced. Let's start with the models. So they're behind. I'm privy to those who are making these decisions. I can assure you the expertise is not there. The coherence is not there. The focus is not there. And there's just a sea of red flags here. So over the next two episodes, we want to take a closer look at Cohere. In part two, we'll look at how this deal fits into some of the big questions around AI, about job loss, the data center backlash, and accusations of copyright infringement. But today, we want to focus on Cohere. What exactly are they building? Why has Ottawa chosen them as a national champion?
2:42And now that you're about to own a piece of this company, what does all of this mean for you?
2:50In some ways, Cohere has a classic tech origin story. In 2017, a 20-year-old University of Toronto student named Aiden Gomez got an internship at Google Brain, where he worked alongside Jeffrey Hinton, the so-called godfather of AI. While he was there, Gomez helped write the research paper that introduced the Transformer to the world, That's the T in ChaChiBT, and the thing that kicked off the current AI boom. Gomez eventually came back to Toronto, where he teamed up with Nick Frost, another one of Hinton's protégés, and a university dropout named Ivan Zhang. Together, they started Cohere. Like I said, a familiar story.
3:30A famous mentor, a breakthrough paper, a college dropout. Within seven years, they had turned a brand new technology into a company valued at nearly$36 billion. But Cohere's story is also unique in some important ways. For starters, co-founder Nick Frost is also the frontman of a Juno-nominated rock band. Don't you notice how I don't need reminders of how you don't feel the same? And in a space dominated by American hyperscalers, they've managed to resist the pull of Silicon Valley. You are going to build and stay in Canada. Why? I like it here. Yeah, in some ways it's an odd question. This is Nick Frost.
4:23I mean, that's interesting that you think it's an odd question, right? Because some people might think, like, why not go to the U.S. where all the kind of juice seems to be? Yeah, yeah. I think the reality is that we think about this, like we being me and my two co-founders, Aidan and I then, we think about this differently. So we grew up in Canada. We are a product of the Canadian ecosystem. And we want to build our company here for the same reasons. Like we feel like we were built by this country. We would like to be building the country now. So what exactly are they building? I think when people think of AI, they think of chatbots like ChatTBT or Claude.
5:03But as I understand it, that is not what you are building at Cohere. And so explain to me what it is that you're doing differently. Well, we build large language models. Large language models are a type of machine learning, which are responsible for the chatbots that you just mentioned. And for most people, when they think about AI, they think about those big chatbots, because that's probably the way they're most actively using them. We do build large language models, but instead of making them as big as possible and paid for via a subscription for people to use in their personal lives, we make efficient, optimized models for businesses, for people to use in their work life.
5:44So it is the same underlying technology, but deployed very differently, and subsequently built with different things in mind. And I wonder if you could just give me kind of a concrete example of that. How are people using it? Yeah, how are people using it? Oh, man. I'm sure you've got many. Yeah, yeah. I mean, we work across industries. People use them all over the place. But you can imagine anybody trying to do work more efficiently, if they're sitting behind a computer, if there's a thing that they can get a secure and tailored language model to do for them so that it frees up their time to do the things that they're more interested in, things like the strategic or the creative things that they're better at, they normally will.
6:25So you can imagine people filing expense reports or people within RBC, which is a big customer of ours, using it for analyzing quarterly earnings reports. Really, any work that's done behind a computer that people find kind of boring, LLMs are a perfect use case for that. This is why you've probably never used one of Cohere's models before. Because unlike the other big AI companies, they don't make a general-purpose consumer-facing chatbot. Now, their argument is that they don't actually need to get good at everything because they're not making models for you and I asking questions about whatever in our lives, right?
7:05Like, the Kohir model doesn't need to be good at meal prep. Murad Hamadi has been following Kohir more closely than any other reporter in Canada. He's with The Globe and Mail now, but when we spoke to him a few weeks ago, he was at The Logic. Where it's best suited right now and where they've really started out is focusing on highly regulated sectors and the big companies in those sectors. So think about telecom, where they've got Bell. Think about banking, where they've got RBC and TD. Think about regulatory. Think about government. These are sectors where they need specific things out of their AI systems and Cohere's Pitch is we can build that for you.
7:43The company has a mantra for this approach. ROI, not AGI. They're trying to build models that deliver a return on investment, not some kind of superintelligence. But superintelligence is the very thing that all the big American labs are chasing right now. Hey, everyone. I want to talk about our new effort, Meta Superintelligence Labs, and our vision to build personal superintelligence for everyone. It's important to move quickly. On our current trajectory, we believe we may be only a couple of years away from early versions of true superintelligence. And then probably by 2030 or 2031, call it five years from now, AI will be smarter than all of humanity collectively.
8:30When we look at these big American companies like Anthropic and OpenAI, they are trying to build some kind of superintelligence, right? And just, I just want to be very clear. Is that something that Cohere is trying to do? We are not trying to build a digital god. I don't think that's, I don't look out in the world and say, oh, gee, I wish there was a digital god here. I do say, hey, I wish computers could empower me. I wish I could use my computer more effectively. I wish I could ask it to code things for me better. I wish I could use it for patching cybersecurity vulnerabilities. And we are building, you know, massively impactful, powerful technology.
9:10But we're not trying to build a digital god. Cohear's decision not to pursue superintelligence might just be their greatest strength and their biggest vulnerability. When the independent organization Artificial Analysis created an intelligence index of 129 large language models, they ranked Cohear's best model near the bottom at 118. But choosing to go after ROI instead of AGI also means that Cohere hasn't faced the kinds of safety issues that have plagued the big American labs. Anthropic says its AI models went rogue during testing and hacked into three other companies. OpenAI, the company behind ChatGPT, is now being sued by eight states over claims that popular chatbot harms kids.
10:01I want to ask you about safety. We've seen kind of a litany of harms emerge from some of the frontier models in the U.S., everything from lawsuits alleging that ChatGPT has helped plan murders and suicides. For the first time, we're hearing how the Tumblr Ridge shooter planned the attack and learned to get around safeguards with the help of ChatGPT. To users experiencing AI-induced psychosis. Seven new lawsuits have been filed against the tech giant OpenAI, claiming their clients have been driven into delusional states, at times resulting in suicide. Too fierce that superintelligence could wipe out the human race.
10:40A new alarming call from inside the house. An AI researcher warning the technology could, quote, kill us all within the next 10 years. Have you seen those kinds of things at Cohere? No. And that's because we deploy them for a very different reason. We deploy our models into a company's environment securely so it can operate on their data without sharing that to people, connected into the right data source so that somebody at a company can actually ask a question about their company in specific and have the correct answers. And we deploy them efficiently so that they actually get ROI. All of that means that people are not coming to our language models to talk about their feelings.
11:22They're coming to our language models to help them do their work. So we don't have those use cases at all, nor do I think those are particularly good uses of language models. I don't think a large language model is a good therapist. I don't think a large language model is a good friend. I think a large language model is an amazing computer, and they should be used as computers. We tracked down two people. He used to work on Cohere's safety team as contractors, and they largely agreed with Nick's assessment. I would say everyone that's actually a part of the core team there really, really believes in the vision that they have of becoming just like this amazing workplace tool.
12:01And I was pleasantly surprised as to how much they cared about making it a safe product. One of them is Deanna Grant. I felt like 90 % that it was safe. I would say like there's a small percentage always that AI is going to not do the thing you want it to do.
12:32Cohere's pitch really has two pillars. One is that they're building enterprise software, not superintelligence, which could mean safer, more reliable products. The other part of their pitch is that they're building sovereign Canadian AI. But what does that actually mean? Yeah, that's one of those funny terms that was not in anybody's Lexicon. Here's Nick Frost again. When I think about sovereign AI, I think about empowerment and I think about autonomy. So to me, you as a person are sovereign, you as a company are sovereign, you as a country are sovereign if you are in control of what's happening to you.
13:07Energy is a good example. We can claim energy sovereignty, energy independence if we make enough electricity in this country for all the things we need and if nobody can turn that off. The same is true with AI. Like we saw earlier this year, Anthropic had a Mythos model, which was a very good model at coding. They announced access for people, and then they decided to shut it off for some countries based on what the government of the United States decided to do. This really occurred over Friday and Saturday. Anthropic disabling access to its most advanced AI models, including Mythos, following an unprecedented order by the Trump administration to keep the technology out of the hands of all foreign nationals.
13:48You cannot claim you have sovereignty if you have built something that a government can shut off if they'd like to. Actually, on the same day that they announced that model, we announced and released very much the opposite model. It was a tiny 30 billion parameter model. So that's a model you can run on a pretty good laptop that was very good at coding but very, very small. And we released that, open sourced it, gave the weights away so that anybody could run it in their environment and nobody could shut it off except for them. To me, that is sovereignty. If we've built the model, if it exists, controlled by the person who's using it, and it can't be shut off.
14:24That definition of sovereignty doesn't necessarily mean Cohere is wholly Canadian. They recently announced a merger with a German company called Aleph Alpha. They rely on American chips from firms like NVIDIA. And when the government gave Cohere$240 million in 2024 to invest in sovereign computing power, That money went to a data center built and operated by an American company called CoreWeave. Does having American components in your tech stack or a partnership with a German company undermine the sovereignty argument that you're making right now? No, I would say very much the opposite. In fact, I think the merger with LF Alpha bolsters that.
15:02How? So to quote Carney, sovereignty – or I forget who said this first. Carney – sovereignty does not mean solitude. It sounds like something he might have said. It sounds like something he'd say. For the record, it was actually AI Minister Evan Solomon. We don't have sovereignty through isolation. We have sovereignty by building resiliency and being in control of what we're doing. The merger with Aleph Alpha, that's a joining of like-minded forces in order to build a sovereign alternative for all the people in the world who care about that. After the break, we'll hear from those who aren't convinced.
16:09to help you work smarter because less time spent on busy work means more time building the business you believe in. Get AI tools for work. Upgrade to Google Workspace with Gemini at goo.gle slash workspace with Gemini. In 2002, Ryan Wedding was an Olympic snowboarder at the top of his game. Ryan's style is really gun it and go. And in less than a minute, it slipped away. But he continued to chase the high. Ryan James Wedding. He controls one of the most violent drug trafficking organizations in this world. I'm Michelle Shepard, and from CBC Podcasts and the Toronto Star, The Most Wanted Olympian, available now on Amazon Music.
16:54Well, I'm having a hard time seeing the benefits from it because of all of the red flags. Jim Balsillie used to run BlackBerry, one of the most successful Canadian startups of all time. Now he chairs the Council of Canadian Innovators, which lobbies for Canadian tech companies. He spent years arguing that the government should be doing more to support the Canadian tech sector. I'm all for Canadian companies growing, and if I was running that company, I'd want the money too. So Balsilli is the last person that you would expect to object to this$1.6 billion investment. But he has three big problems with it.
17:32It's tremendous concern, like I would say to the point of overwhelming concern. The first is that he thinks the government's approach to AI sovereignty is too narrow, too piecemeal. That by focusing on this one LLM company, they've neglected other parts of the AI ecosystem, like where the data is held. When you want sovereignty, the most important thing is to control the data and have that under sovereign governance and rules. And so is this a company now moving into being a sovereign cloud or is it an LLM company? And there are over a thousand open-weight LLMs out there. So I'm sure it's a good LLM.
18:12I'm sure they're doing the very best. I wish them the very best. But the probability that these could all be shut off and this is the make or break of sovereignty doesn't hold water. Put another way. If it's a Canadian LLM, but they're using an American cloud provider, like CoreWeave, is that really sovereignty? You have to remember, roughly a year ago, the government gave them$240 million for, quote-unquote, sovereign compute that went to CoreWeave, which is kind of like cigarettes for the health program in the school. And again, I don't know what they're thinking here. Balsilli's second criticism is that the terms of the deal are pretty opaque.
18:51We know that Ottawa will be investing up to$1.6 billion in Cohere, and that in return Cohere will keep its headquarters in Canada, create at least 1 ,000 new technical jobs, and spend$1.7 billion building computing infrastructure here. But beyond that, the deal is light on details. For example, we don't know how big of a stake the government will be able to buy with their planned investment. There's a lot of opacity, and the company's happy to say they're raising$7 billion on a$35 billion valuation or something like that. But then when somebody wants to ask what's Canada getting on it, they say that's proprietary information.
19:29And that's kind of like saying, I'll show you my left thigh to get you all interested, but my right thigh is proprietary. The federal Conservatives have raised another concern. And the House of Commons Conservative MP John Brassard pointed to some people with close ties to both Cohere and the government, and questioned whether they stand to benefit. A member for Barry South Innesville. This is not about Cohere. That includes Dominic Barton, who chairs the federal agency responsible for attracting foreign investment, and who is also a partner at Radical Ventures, one of Cohere's founding investors.
20:04Canadians have seen this self-dealing, Barton liberal playbook before. They shape public policy, then benefit from the results. Patrick Pichette and Jordan Jacobs were both appointed to the Prime Minister's AI Council and both sit on the board of Cohare. Jacobs co-founded Radical Ventures, where Barton is a partner. Together, they back Cohare in its early stages, advised the Prime Minister on Canada's AI strategy, and now the company they invested in has received$1.6 billion in taxpayer funding. When will the Liberals stop using taxpayer dollars to align public policy that enriches well-connected Liberals?
20:44We reached out to all the parties mentioned in that clip and the government. A spokesperson for the Minister of AI told us that Dominic Barton, Jordan Jacobs and Patrick Pachette were not involved in any discussions on the investment in Cohere. Investing Canada, that agency, the Barton chairs, reiterated that he had no role. Pachette, for his part, said he's no longer on Cohere's board. His spokesperson said that while Cohere is in his investment portfolio, Pachette was not involved in the government's decision to take an equity stake in the company. Then there's Balsilli's third criticism, which is that Ottawa seems to have put all of their eggs in one very large Cohere-shaped basket.
21:24You could do this in a way where you're not necessarily picking a winner. You could have an RFP for two companies out of a set of proposals. You know, there's a half dozen really good agentic layer companies, and you could pick two. It's a very concentrated big bet when we have a lot of bets. And so what justifies such a concentrated bet when there's apparently so much funding there at that valuation? And again, this is not a criticism of the company. I wish the company well. well, I'm all for Canadian companies, but it doesn't hold up to scrutiny for public investments. And we're in an era where we need lots of public investments.
22:07But this isn't how you do it. I want them to do this properly so that you get the political and country support that this is a nation-building time. Canada was built on national initiatives. A hundred years ago, we were building grain co-ops. We were building telecommunications and the original Bell Canada. We were building railroads. We were building seaways. This country was built by strategic initiatives. And we went to sleep on that in the last 25 years. And we have to wake up and do these things because we're seeing that the world behaves a certain way, that you have to be strong and resilient as a nation.
22:48So I'm all for the strategic initiatives, but you cannot commence it like this because it's not going to work because you're not doing it right. All this raises the question, why did the government choose Cohere as their national champion? I pose that to Evan Solomon, the federal minister of artificial intelligence. There are hundreds of AI companies in Canada. The U.S. Department of Commerce says there are 1 ,500. Why has the government chosen Cohere as their national champion? I think it's really important for Canadians to realize that there are really only four countries in the world that have the possibility of really designing these large language models, doing what Cohere does.
23:37There's only the United States. They have, you know, chat, GPT, Anthropic, and others. China, there's one model in company in France, and then there's Cohere. And so Canada is one of four countries that really has an opportunity to shape this. And so it's really important. We believe this is a Canadian success story that we believe Canadians can now participate in that. And it gives us an opportunity to shape how AI is built, who it benefits, the values that it's built. and it puts Canada with a seat at the table. You hear that four countries line a lot from Minister Solomon and the Prime Minister.
24:20It's observed that Canada is one of only four countries in the world with a large language model. There's only four countries in the world that have a foundation model. Only four countries in the world that have this capacity. France, Canada, China, and Canada has Korea. It's a catchy line, but it's not entirely accurate. Stanford's 2026 AI Index Report lays out which countries released notable AI models last year. They say there were 59 from the U.S., 35 from China, 8 from South Korea, and 1 from Canada. The same as France, Britain, Singapore, and Hong Kong. And even if you do buy that there are only four countries in the world with these frontier models, Journalist Murad Hamadi says those countries aren't exactly on the same playing field.
25:09Frankly, to call it four countries is stretching the truth. There are two countries, and then there are two other countries. There's the U.S. and China, which are like the runaway leaders. And then there's France with Mistral and Canada with Cohere. And I should say, there's quite a lot of people in Canadian AI who think this is the wrong bet, that Cohere gets too much attention, gets too many of the resources, is just favored too much. The argument that you hear when you talk to people sometimes in AI is, we've lost the foundation model race. Sure, we have a contender, but they're so far behind.
25:44We should focus on XYZ instead. We should focus on being the best builders of applications, because actually that's where the money is going to be made. We should focus more on companies that can get to profitability sooner. We should focus on infrastructure because that's where our advantage is. We should use Canada's natural resources. And it's not so much people being like, oh, they're getting an unfair advantage. It's more like people thinking, this is the wrong bet. So what explains the size of this bet then? Murad thinks that the answer might have more to do with politics than technology.
26:18Is there some FOMO here? Absolutely. Look, if you're going to be the Canadian Prime Minister, who is trying to build a coalition of middle powers, who is trying to set up something that at least on paper can look like a viable alternative to our dependence on the United States, who is trying to say to the world, you don't need to be absolutely beholden to them. Then you look down and you say, what are the cards in my hand? And the cards in your hand are Canada's energy resources and a bunch of other stuff. But look, AI is the thing of the moment. And like, it's kind of nice if one of those cards in your hand has the word AI stamped on it.
27:05So if you're looking down and you're trying to pick that card in your hand and you have something that allows you to say with a straight face out in the world, there are four countries in the world that have a foundation model and we're one of them. You pick up that card. Maybe it's that simple. The thing that nobody, including Cohere's co-founders, really disputes is that their frontier model is way behind the ones coming out of China and the U.S. I just want to talk to you about what it is that Cohere is doing. You spoke earlier about how they're working in building these LLMs. I asked Minister Solomon about this.
27:42Cohere has repeatedly said they're not trying to build artificial superintelligence, which is what OpenAI, Anthropic and Google are all trying to do. But the reason why investors are pouring trillions of dollars into these other companies, the thing that will supposedly revolutionize the world, is some kind of superintelligence. And fundamentally, it seems to me like what Cohere is trying to build is really good enterprise-level business software, which is great. But how do you rationalize investing$1.6 billion in a company that, by its own admission, isn't trying to build the thing that's supposed to be revolutionary?
28:20Well, first of all, this is an important investment to build very powerful AI models that companies and businesses can use. And there's lots of ways to do this. It's not just a matter of who spends the most money. It's what is the most useful, what is the most practical, what is the most pragmatic model to use. and Cohere is developing practical models that's meeting the real needs of organizations and hospitals, small businesses, banks, telcos. You know, like I always like to say that, you know, you don't use a Ferrari to drop your kids off to school and go to work. Most businesses don't need to have Ferraris as delivery vehicles.
29:05They need practical models that get the job done. But just if this is about protecting Canada and protecting our sovereignty, if we're not doing the super intelligence thing and the U.S. and China are both doing the super intelligence thing, does this really protect our sovereignty? You know, the question about how close to the frontier, you know, right now at this moment, Cohere is not what you call at the frontier, the farthest end of the frontier, where there's a lot of concerns about alignment and where this stuff's going. But they are very much at the frontier of cybersecurity about testing.
29:50And this allows Cohere to keep moving towards that frontier and to be practical and safe and sovereign models to give Canadians choice. I also wanted to ask the minister about the actual terms of this deal. The government has said that they're making a commercial investment in Cohere. So what can Canadians expect to get back in return? I just wonder if you could say more about what you mean by that. And just particularly maybe where like the profits will flow. Like what does that mean that it's a commercial investment? It means that the Canadian government is making an investment and taking equity ownership in this company so Canadians can benefit from it.
30:29And how would they benefit from it? Like if they start making lots of profits, do Canadians get dividends? Does that money get put into a fund to help people who, I don't know, maybe lose their jobs because of automation? So when you make an investment like this in a company, an equity investment in a company, obviously our duty is to make sure that Canadians' money is invested wisely, that there's a benefit to Canadians to build sovereign and to build safe AI and to give us a seat at the table and to give us sovereign control and to keep these companies Canadian, which is important, strategically important for our autonomy, and to give us choice.
31:11And of course, there's lots of ways. I'm not going to get into what might happen. We will always make decisions for the benefit of Canadians. But absolutely, we think that a company that is a Canadian success story, that Canadians should be able to participate in that success. I tried asking Nick Frost about this too. What can you tell Canadians who I suppose now own a part of your company about the terms of the deal? The round is still closing. And so the details will come out when it is closed. I know Minister Solomon has said that revenue and job targets are key. And just like, what can you tell us about those targets?
31:50All those details will come out as the round closes. So the$1.6 billion question is, is this a smart bet? The answer probably depends on where you think AI is heading. If the frontier labs are the rocket ships, the rocket ships are going to make a lot of money. There's a lot of money in rocket ships. And some stuff you do need the rocket ship for. But some stuff you need, like, a small passenger plane for. And some stuff you need, like, a delivery truck for. There's probably a lot of money for Cohera to make in selling delivery trucks and passenger planes without having to sell rocket ships. If AI turns out to be a normal technology, a new kind of software that makes businesses and governments more productive, then perhaps Canada has made a sensible investment in a rapidly growing company.
32:41But if it's the revolution the American labs keep promising, if the future really is rocket ships, will a$1.6 billion investment in Coher equip us for that future? Let's say you're wrong. And superintelligence is possible. Yeah. And somebody else builds a machine gun. Yeah. Would we not need a machine gun?
33:07That's coming up in the second part of our series here on FrontBurner. I'm Jamie Poisson. This episode was produced by Mitchell Stewart. Thanks so much for listening. We'll talk to you tomorrow.
At the end of October, Prime Minister Mark Carney announced that his government would be investing $1.6 billion in the AI company Cohere. It’s one of the biggest bets Ottawa has ever made on a single tech company.
The government says the deal is a critical step towards securing Canada’s AI sovereignty – but not everyone is convinced. Jim Balsillie, the former head of Blackberry, told us he sees a “sea of red flags”.
There’s also an even more fundamental question. Cohere has said they’re pursuing return on investment not artificial general intelligence. Which means that, unlike Anthropic and OpenAI, they’re not trying to build superintelligence. Is that enough to keep Canada in the global AI race?
This is the first of two episodes on Cohere. In part two, we bring these questions to the company’s co-founder, Nick Frosst.
For transcripts of Front Burner, please visit: https://www.cbc.ca/radio/frontburner/transcripts