00:00CEO Larissa Horton joins us now. She previously held roles at Cisco and Microsoft. Larissa,
00:06thanks so much for joining us. This technology is really cool. Can you just explain exactly
00:11what it does? Yeah, so we start with really basic data like film, which we convert into
00:18an understanding of where you are on the field or court, so 2D, and where your body is in space,
00:243D. From there, we look at what are you actually doing? Are you running a slant or a post? Are
00:29you running a comeback? What did your linemen do? How are they holding the pressure? How do they
00:34protect their quarterback? What are all the spaces that you can actually take advantage of and become
00:39the new edge for your team? So we really understand the movement of the player and the entire team
00:43together, and we're doing this at the frame level. So we're doing it at 10 frames per second. We're
00:48constantly assessing where you are and where every single part of your body is moving for all 22
00:53people on the field. So, I mean, I think it's super exciting, and I've heard that you can, you know,
00:59use it for predictive, so you can kind of try and figure out what the other team is going to
01:04do
01:04in certain situations based on what they're doing now. You can use it for player development. You
01:10can use it for recruiting. You can use it for, like, stadium planning. But I think the genesis of
01:15this company is even more interesting. I heard that Paul Tudor Jones and his son Jack were playing
01:19fantasy football, and because he's so data-driven, he came up with this idea to have AI help him pick
01:27winners. Yeah, so that is the start of the company. They were in a fantasy league where they wanted to
01:31find an edge, and obviously with their quant background, dealing with data all the time,
01:35they're like, how do we use the same strategies that we have in our hedge fund to then apply
01:40to draft? That entire basis of using data to understand football better is what we grew from.
01:47And so the draft and fantasy was one angle, but as you just mentioned, there's so many other use
01:52cases that are useful. And by the way, Paul Tudor Jones still owns the company. He's the one that went
01:58out and found you. What was his pitch to you? Because you're moving in this, you know, incredibly
02:04high-paced world of big tech and running companies and getting options, right? I mean, what is his pitch
02:11to you to draw you into this new startup? I think the world is changing at a really rapid pace
02:16across all
02:17of sports. From high school, we're seeing that, you know, getting seen, being found is becoming a
02:22little bit of a pay-to-play, and that's something we want to democratize access to. At the college
02:27level with NIL, we are seeing 10,000 kids in the transfer portal. It's a completely different
02:31landscape of challenges for coaches and ADs. And then obviously, we want to keep the pro game as
02:37competitive as possible, but it requires finding talent from high school to college to pro, and even
02:42globally, to ensure that you have the highest, most competitive, exciting games to watch. And so when
02:47we talked about that being the challenge, and not just in football, but potentially across many
02:51sports, that was super exciting. You know, helping kids get to become a pro, finding the best players
02:56for every college team, and at the pro level. Like, who doesn't want to work on sports all day long?
03:00I was going to say, what about for us normies who are never going to become athletes? I mean,
03:04is there an expansion? Danny, you are an athlete. That's so kind of you, but I am absolutely not. She
03:10plays
03:10tennis every single day. It doesn't mean I'm good at it. But I know, I think about that. I'm like,
03:15I would love to have your app in the background being like, oh, your serve is bad because you're
03:19not tossing the ball high enough. What are sort of the areas that you're thinking of pushing into
03:23beyond what you have right now? Yeah, so we actually have an app that is available in the
03:27app store for high school kids, and they can measure their 40 time, their hand size, their wingspan,
03:32their height. And we start to tell them, who do you look like in D3, D4, or D3, D2, or
03:38FCS? And give them
03:39that blueprint of what can you start to do better. So what we found within folks who are not pro
03:44athletes or even trying to be pro athletes, they still want the tips of what are the pros doing
03:48really well, and what are things that I can improve to slowly get there, and give me that difference.
03:52Help me understand how I can make those improvements. And with computer vision, it's actually really easy
03:56to show you versus your favorite pro player. Yeah, I like this. I'm going to have it record me,
04:00and then it's going to be like, you look just like Naomi Osaka. You should quit your job
04:04and go play pro tennis. What else can you do with this? Because event stadium direction or event
04:12planning, how do you use this technology, which is so focused on football or maybe even other sports,
04:18to help control those kind of massive environments? Yeah, so on the fan engagement side, one of the
04:24things that we looked at is how do we take the production happening on the field, the exciting
04:28events of the longest catch, the biggest run, and actually bring the fans into that. And this is where
04:32we actually have been working with Cisco on bringing their Cisco AI pod to take that data and
04:38bring it right into fan merchandise, concessions to be like, hey, did you know this is the first time
04:44we've broke our run record in 10 years? Do you want to go buy this jersey now? Or hey, we're
04:49running
04:49with extra hot dogs on that particular concession stand. Why don't we give them away? Why don't we
04:53just tell people there's all this food that's available, go run down to that stand. And so we're looking
04:58at the same foundation of analytics of understanding where are we in the game? Is it a slow period? Is
05:04it a fast period? Is it a historically breaking record situation? And let's engage the fans along
05:09the way. I mean, you've been innovating your entire career. But when you have this new technology that
05:14a lot of people have access to just AI more broadly, I think, you know, of an AWS, they sponsor
05:19F1 and they always have these like predictive metrics up like you will be this car will be able to
05:23pass
05:23this one in this many laps. How much battery power does this Formula One car have? That's Matt's favorite
05:28topic. We should not get into that because we will be here all day. What becomes your moat for other
05:33entrants into this space that are trying to compete? So there is a couple of things. Obviously, we have these
05:37large foundational models that are getting better at every single industry. But there's a lot of secret data like
05:42that playbook, like the plays that we don't actually know what they're calling. We just see what they're
05:46executing. And so bringing in all of your custom data and your custom strategy into this ecosystem in a private
05:53infrastructure is becoming something that's different. Also, we are generating all of this labeled
05:57information at the frame level. So while the game is happening, we are generating this data. A lot of these
06:02models are operating fast, but not necessarily in this kind of real time nature to the degree of every I
06:08tell my kids
06:09head, shoulder, knees and toes, every joint, every angle, every step, we are actually calculating and we're
06:15predicting what you're going to do next that many times a second.
06:18You know, as I was talking yesterday to Dr. Denish Nagda from a company called Resilient, and they're
06:23doing basically telehealth, but it's really with an AI backbone across the country. And I thought Sumer
06:31Sports could analyze me for medical reasons. You could become Sumer medical. Or if I'm going through TSA
06:39and you want to know, you know, do I have anything dangerous or drugs, you know, you could, please
06:45don't. You could, you know, Sumer safety, like you could branch out into so many different
06:52industries. What's next?
06:54So we want to focus first on getting sports right. But it's funny you mentioned this because we've had
06:59theme park companies come to us and say, wait, I want you to just watch everyone and give them this
07:04like happy summary of their day on all of the rides they went on. So only when they're happy,
07:09because we can find you and you're smiling, do I want those clips automatically, you know,
07:14collected and generated. So you leave the theme park and all of a sudden you have this really fun
07:18video of the different rides your entire family went on. I think there's so many use cases, like
07:23you said, of taking computer vision, but applying it in a really specific scenario. So football, it's
07:27understanding the athlete, the team, healthcare, it's understanding, do you look well? How do you look since
07:31the last time I saw you theme park? It's moments of excitement. There's so many applications and it's
07:36really getting to that last mile of AI adoption. So it's not just the generic use case, but how do
07:41we make it useful for you, your business, your customer? And I think that last piece is where
07:45we've been really focusing on.
Comments