Inside the AI Revolution: Together AI’s CEO on Where Artificial Intelligence Goes Next

I'm here with Vipul Ved Prakash. He is CEO and co-founder of Together AI. Thank you so much for being here, Vipul. Thanks for having me. Of course. Tell us a little bit about what is going on in the world of AI today, right? We're seeing so many changes, so many things people worry about, and also so many things people are excited about. What's going on? You know, one of the things that has happened is that models have become good enough to do intellectual labor. They are being deployed widely in enterprises, you know, doing really productive work. And this has led to, you know, token maxing and really deploying deployment of AI across various, you know, functions inside companies. And one of the big trends that we are a part of is, you know, open models, open source models have now caught up to close frontier models. So these are being deployed widely as well. Well, and they are It's dramatically cheaper in terms of token cost relative to closed models. So we've seen our token serving volumes go from 100 billion tokens a month last year to 600 trillion tokens a month last month. So you're seeing this diversity of models now emerging and this model ecosystem of both closed and open models that are being deployed. as part of agents and agent swarms that are performing creative work inside companies. It's a very exciting moment. Yeah, it really is. There's just so much going on. Again, these jumps that you're talking about here and what is able to happen now with these open source models. Talk to us a little bit about what's behind the idea of open sourcing so much now. Again, it used to be this kind of closed idea of don't steal my ideas. This is completely proprietary. And now there's been this shift. So talk to us a little bit about that. You know, I would say there has been an open-source thread in most software infrastructures. If you look at operating systems, right, you had Linux, which became, really became a foundation of cloud computing. You have open-source browsers, so as open-source databases, right, and as technology becomes, you know, very fundamental technology, which gets deployed in a variety of ways around the world. It's generally quite difficult for a vertically integrated company to cover all those use cases and this kind of gives sort of naturally gives space for open ecosystem to exist. And when we started together four years ago, we had this belief that intelligence is going to be very fundamental, like, you know, bandwidth or energy and there will be a open and custom sort of. portion of this ecosystem, which has really materialized this year. And that's what you're seeing. Yeah, it's really interesting thinking about that evolution and also where we're headed next. Tell us a little bit more about what led you to start together AI, especially with your vast experience, right? You've started companies before you were at Apple before. So tell us a little bit about what led you to start it and how that process was in this world of increasingly more AI companies doing all sorts of different functions in this space. You know, one of them excitement about AI. This is like really the, I feel it's the terminal state of computing. This is what we've always wanted computers to do since, you know, since they were invented. And we could just see that the scaling laws that suggest you will keep getting more intelligent models as you apply more data and more compute to it. So that's a pretty remarkable result. We've not had such a result. you know computer science before so part of it was that it was very clear to us that this is going to be start of a new era of computing and we wanted to be a part of it and the other aspect was that you felt there will be concentration in the space you'll have you know because it's required so much capital you will have a few companies that may come to control the entire kind of AI ecosystem And, you know, we almost felt it was a moral imperative for us to create a platform that allows for open and custom models and it's really exciting to see that's actually happening. It is. I mean, just to think about starting this company, growing it, having it be so successful so far. How do you balance all of that and wanting to just move forward, forward, forward while also thinking about the safety concerns and privacy concerns and everything else that goes along with the concerns that people have about AI? Yeah, I think those are, you know, as models are becoming more intelligent, those concerns are coming to the fore. And we think, you know, we deeply believe that having open models, which are very transparent, is very important to that safety conversation. You can examine these models, you can benchmark them, you can understand exactly how they're working, and not just us, but, you know, everyone around the world, from companies to researchers, can examine these models, and it's, you know, we think that's a pretty important part of the safety conversation. Yeah, it really is always keeping in mind that balance of excitement to move forward, but wanting to keep everything safe as it is now. As we think about the future, where do you think this is all going when we look not just to the end of this year or next year, when we look five, 10 years down the line, what do you think this kind of AI landscape is going to look like? You know, I think models will keep getting better. You know, you are We are driving towards having super intelligence that is sort of vastly superior to intelligence that we we have today and will really become part of the in some ways a labor force. We'll have you know not just humans but humans and machines working together to solve the kind of problems we have not been able to solve before and you're seeing you're starting to see it's like at the at the conference today. Talking about how decision making and execution side of companies has already become a lot faster because of AI and I think you will have new industries that are just sort of not economical or possible today, which will become possible with the help of intelligent systems. Yeah, it's so exciting to think about where all of this is headed. Do you have any specific concerns as you head forward for your company when it comes to the biggest challenges, when it comes to all of the geopolitical factors that are kind of weighing on everyone, whether you have a company or not? What are your kind of concerns about all of that as you try to move forward, again, growing the business while thinking about everything else that you have to weigh with that? Yeah, I mean, I would say this is definitely a technology that's moving way faster than You know, regulation or even society's capacity to absorb it, although humans are very good at absorbing new technologies and just, you know, changing along with it. So I think there is going to be, and it is disruptive, right? It's going to change industries, and I don't think anybody quite understands the nature of those disruptions, and so... And that is part of building something new. There are unknowns that you will encounter and have to work with as and when they happen. Yeah, definitely. Well, you've given us so much to think about today, Vipul Ved Prakash, CEO and co-founder of Together AI. Thank you so much for being here today.