StarTalk Radio - Artificial Intelligence and Sports

Episode Date: August 14, 2020

Will AI change sports? Has it changed sports already? Neil deGrasse Tyson, co-hosts Gary O’Reilly and Chuck Nice, IBM’s Noah Syken and Second Spectrum’s Rajiv Maheswaran investigate how artificial intelligence is impacting sports. NOTE: StarTalk+ Patrons and All-Access subscribers can watch or listen to this entire episode commercial-free here: https://www.startalkradio.net/show/artificial-intelligence-and-sports/ Photo Credit: Second Spectrum Subscribe to SiriusXM Podcasts+ on Apple Podcasts to listen to new episodes ad-free and a whole week early.

Welcome to StarTalk, your place in the universe where science and pop culture collide. StarTalk begins right now. This is StarTalk Sports Edition. And our special topic today is the role of artificial intelligence and machine learning in sports. What is it? How is it manifesting? How is it going to change what we experience, not only as viewers, but how it might change with the athletes themselves? So, who do I have with me?

Of course, my trusty co-host, Chuck Nice. Chuck. What's happening, Neil? Okay, you stand up comedian and you think about sports sometimes. Occasionally. You know, every once in a while I have a dream about it. Okay, so for this we go big guns and of course we have Gary O'Reilly.

Gary, ex-footballer, ex-footballer. What team did you play with? Tottenham, Crystal Palace and Brighton. So my idea is that I think about sports a lot and occasionally I think I'm funny. Okay. But it's only me thinking I'm funny, apparently. So that's the combination of that that we need here.

Very good. Well, today we have a special guest, Noah Sykin. Did I pronounce that right, Noah? Yes, Neil, you did. Perfect. Noah you're VP of Sports and Entertainment Partnership at not a sports company

not an equipment company not a logo sneaker company at IBM. Wow. So this is Revenge of the Nerds. What's going on? Why do you even exist?

How about that? Well, sports uses technology and data, so why not IBM? Well, there you go. There you go. We're done here. And ladies and gentlemen, the show is over. Thank you all for tuning in. I think we've covered it all.

It's all about quality, not wits. Okay. Well, I just wonder, because I think we did have to run through a generation of those who had gotten wedgied and their lunch money taken by the football jocks to let that sort of work its way under the bridge. And now we have the next generation of techies

who the sports community embraces, such as yourself. What's your formal background? I'm actually an advertising guy. I started in advertising and media. And as I came into IBM, I learned all about technology. I learned about the intersection of how do you take great technology, how do you mix it up with something that people love and they're passionate about and tell stories? And so, like I said, the mix of technology and sports is actually very natural. So IBM needs the person who is sensitive to the technology, has the target objective with sports,

but also IBM is one of the leaders in AI in this world. So they need somebody like you. Yeah, well, look, you know. Or at least you tell yourself that. Well, I look at it as, you know, sports are business enterprises. They have the same needs as every other business, right? They have to get customers. They have to please their customers. They need a great technology infrastructure.

It needs to be secure so that nobody's, you know, breaking in and doing bad things to their technology infrastructure. So sports are just businesses with a different name, and they don't call them clients. They call them fans. And so we're very tuned in to how to help businesses operate better. That's what we do in sports. Well, Chuck and Gary think about AI all the time.

Every time we have one of these shows, they're bringing it up. So I'm sure they got a million questions for you. I'm waiting for the day that AI is thinking about me. It already is. You just don't know it. In fact, it just did your gig last night at the comedy stop.

So Noah, I mean, I think people realize in the back of their mind somewhere that ibm has a connection to sport but i don't think they realize just how holistic and embedded you are because it's not just tennis it's not just golf it's not just we can protect your tech. You're in so many different areas of sport altogether. Can you kind of expand on where IBM has gone in and moved the game forward? Sure. Well, I mean, our history goes back to the Olympics in the 50s and 60s. And we started just keeping track of the data with punch cards, you know, a couple generations ago. Explain punch cards, you know, a couple generations ago. Explain punch cards, please. It's when you get really mad at a birthday card.

Well, you know, Neil, but it's those little cards that you used to get at toll booths, right? You used to slide the card in and it has all the little holes in it. So mechanical computing, right? It used to slide the card in and have all the little holes in it. So mechanical computing, taking those little punch cards and scoring the rowing contest or the fencing contest or the basketball contest in the 50s and 60s. And that just accelerated through the 80s and 90s into the digital age and the Internet. So Noah Sykin is 90 years old. That's what he's telling us.

My gray hair, yeah. My gray hair, yeah. Okay, so you guys have been, I guess it's obvious in retrospect that you would be there because it is a lot of data. But I'd heard that certain Olympics among them, but Wimbledon as well. So you guys have a focus right now, no Olympics this year. So what's your focus?

Yeah, so in the modern age, we partner with entities like Wimbledon, the U.S. Open, the Masters, the Grammy Awards, fantasy football. So Wimbledon was canceled, but the club came up with a great idea to actually play out Wimbledon over the course of the two weeks and still present four or five new matches each day of the fortnight of Wimbledon. They're classic matches, but if you think about the video content of your favorite match,

Borg-McEnroe 1980, broadcast standards weren't where they were today. So we actually created some custom algorithms using our AI and updated about 15 of those matches to modern day broadcast standards. So being able to actually see Bjorn Borg's hair, you know, fiber by fiber, standard. So being able to actually see Bjorn Borg's hair, you know, fiber by fiber, and really bringing a modern day standard to all of that broadcast footage and presented it on Wimbledon.com. So in just a matter of six weeks... Oh Noah, what you've done there is you've given AI the chance to give us data on Borg McEnroe final that never existed because AI can't use the grainy old footage. Oh, did you release those into the wild or is that just something you're keeping back? No, no. If you go to Wimbledon.com today, you'll actually see modern day statistics that weren't

captured in 1980 that were able to be infused into that content today. So you're right on point. It's not just the video, it's all the supplemental data that we have ability to update. But that begs the question then, since these guys cannot play any longer, could the AI take the data, the history of these matches, compile all that data, extrapolate what the outcome would be under certain conditions for a match that has never even taken place. That's what, in fact, that's what I'm waiting for. I don't care how many strands of hair Bjorn Borg has.

Show me. Right. Neil, how dare you? Well, no, because I see what Neil is saying. It makes great sense because here's the deal. Invent a match. Invent a match. Invent a match. Borg versus Federer. Could we find out Borg versus Federer? Dude, you've got the data.

Yeah. Who would win? You've got the data. I mean, okay, what metrics are you going to need to construct a Borg versus McInerney? And you know what you can do? You could take a modern person, put them back in time, slow down their racket, the ball speed off of the wooden bamboo rackets, and just to equalize that. So that's what we want you to do.

Go back now and do that, and we'll call you when you've got something to report. Oh, you guys are great. You guys should come work for us. You have all the time. All right. People are always trying to match up people who never competed against each other,

especially in boxing. You know this is a pastime. Neil, it's the classic barroom conversation. Yes. And now we have it. Now we can actually produce. But providing we set the metrics and continue as a constant in the equation, we should be able, now I'm talking as if I know about AI,

we should be able to do that. What's an example of the kinds of data you would collect in a tennis match? Well, so let me actually talk about the data in two ways. You guys are talking about two different aspects, right? Statistically, who's going to win that match, right? But you're also getting at who's the best player of either one of those players. And that's a more nuanced conversation because that gets into your heart and that gets into what everybody believes, right? And so we can do this statistical modeling

of, you know, who would perform better, whose speed is, who serve is faster, who has more, you know, less unforced errors, all that kind of structured data. We think about structured data, rows and columns of data. But then we think about unstructured data, video, language, right? Language. And that gets at the second part of the equation, not just statistically who would win, but who's the favorite in the eyes of the fans. And more and more, we're thinking about how do we apply AI to that question? How do we get people talking about the matches, their favorite matches, their favorite players, expressing themselves? Because at the end of the day, there's no real

answer to who's the best. The best player is who you believe is the best player and who the people believe is the best player. So we think AI actually has the ability to go, we know AI has the ability to go out, listen, learn, and compile what people say about the players and actually synthesize that into a crisp, articulate point of view as to who the best player is or who would win that match if it was played today. Structured data, unstructured data, both have a role. So you're using language as A, a metric, and B, as a way for AI to learn sport? Yeah.

Okay. Or to learn the fandom of sport. Yeah. That's it. Also, I mean, if it can analyze language like that, Neil, imagine the coach's conversation with a player, the caddy's conversation on the tee or on the green in the Masters, which IBM are involved in.

Imagine all of a sudden, they can analyze that data to bring forward. Because we're looking here historically, this Borg-McEnroe thing takes place in 1980. That's historic analytics. I guarantee now, Noah has spun that 180 and is looking with AI going forward in a more predictive role. Am I right, Noah? Yeah, absolutely. I mean, we think about predictive statistics already. We produce keys to the match in tennis. What do you have to do? What does this player have to do when there's these circumstances to win this match? But we are very much looking at predictive using natural language. I talked a little bit about tennis, but in the world of

fantasy football, we're using the same kind of capabilities. So we're out there reading the blogs, reading the local beat reporters, working with ESPN and the projections that fantasy football players will get on ESPN fantasy football app are actually augmented through AI based on our natural language understanding of what people are saying about the players. So this notion of language is pervasive across sports and more and more we're tapping into it. Think about sound right now. You have all these competitions going on with different levels of sound. In a team sport, going on with different levels of sound. In a team sport, right, if you eliminate that piped-in sound, you're able to hear the interactions of the football players on the pitch, the hockey players on the ice, right? So getting value out of that, how are those players

interacting with each other? What are the coaches saying? What are productive interactions between coaches and players that lead to success versus unproductive interactions? So now how exactly would you be able to implement that even from a fan perspective? Because if I'm a coach, I'm looking at the information you just cited as proprietary. I do not want you on the field listening to the communication patterns of my team because that gives my opponents the opportunity to see inside my head. So at that point, I'm trying to punch Watson in the face. I want to see that.

Chuck, Chuck, you do that. There goes your cable subscription. There goes everything. There goes your... Chuck, you're right. You're right on the right point. You know, these are the gnarly questions that the sports organizations have to sort out right now. And there's no question. I mean, go watch a hockey match tonight, one of the hockey games tonight,

and you'll see they have this crowd noise and you can't hear the voices. And, you know, in many cases, I suspect that's very purposeful because of the exact reasons. But a great organization who wants to compete more effectively is going to get their hands on that data. They're going to get their hands on those interactions and understand those interactions, just like a business wants to communicate with their client and understand what their clients are saying. So what you're saying is Bill Belichick will definitely get his hands on this. I said nothing. I'm a New York fan. This is the next generation of sign reading like we saw in baseball. Basically.

No, I know you and IBM have a strong connection to the U.S. Open tennis, right? But there ain't nobody in the stadium, and it's a cavernous. The Arthur Ashe Arena is just a magnificent space, but it's going to be empty. Am I wrong in knowing that you have mapped the sound of every round going through to the final? And depending on who's playing who, you can actually recreate the noise from a first round with an unranked, unseated player winning

to create that soundscape to, and this game's never taken place, I don't believe, Nadal Federer final. So you could do all of that? Yeah, so for the past five years for Wimbledon and the US Open, we've been creating AI-powered highlight reels. And so what we did was we'd listen to the sound of the crowd. We would watch the players in terms of fist pumps, gestures of excitement or not excitement. Understand the particular position. McEnroe would need a whole separate anger track. He needs his own algorithm. Understand, you know, is this a break point?

Is it a double fault? And we take that and analyze it all with AI and give it an excitement score to each individual point in a tennis match. But remember, part of that excitement score was driven by the audio. So, Gary, as you said, you know, we've been listening to the sounds of tennis for five years, using AI, understanding the characteristics of a first-round match versus a final match. And in this moment of need, we are enabling the organizations like the USTA to be able to turn around and actually generate that sound in their stadium. Whether they will or not is their own choice, but they have that capability now.

So you need the judges to come in to say to AI, quiet, please. Quiet. To the AI. A question. There's no sport like baseball with regard to statistics. And I think it's because there's so much downtime between pitches and events that the announcers

have to talk about something. Translation, it's boring. Go ahead. Go ahead. Go ahead. This is the third time he's stolen second base in late innings when they're down by three runs. On a Wednesday.

So the statistics are so rich that for every player in every event, in every situation, it seems to me AI can just create baseball games that statistically and accurately represent actual players and actual teams. Is anybody working on that? Well, we're not. That's for sure. All right. Well, I'll go you once. Did he not have an attitude? He had an attitude.

He copped a toot on that. Yeah. Let me tell you something. Yeah. I must ask you, because we always kind of end up landing on this particular spot. We can historically analyze things. We can predict to a certain extent based on history. And these are kind of facts that we have,

but what we haven't got yet is understanding what's between the ears of an athlete. Has IBM actually gone into this brain learning space or is that just too much of a leap forward right now? Well, you know, I think, Gary, again, I think for us, at least, it comes back to language, right? And the ability to understand language. If you were to take Tiger Woods as an example, or take Tom Brady, take your favorite athlete, over the course of their careers, they've done how many hours of interviews? Post-round interviews, pre-round interviews, games, Super Bowls, pre, you know, all that content, hundreds and hundreds, right? And so if you can actually understand what they're saying, listen to what they're saying, decompose what they're saying, track it over time. What's

the sentiment of their conversation? What kind of language do they use? I think you're going to be able to actually get a pretty good feel for their mental state over the course of time if you did that work. We're not putting, you know, electrodes on people's heads or anything like that quite yet. Maybe other organizations are. But for us, it's really about people expressing themselves. And people will tell you what they think and what they believe if you have the right tools to be able to get at it and actually, you know, track it over the course of time. So again, for us, language, natural language, understanding, using those capabilities to understand the language of players, coaches,

owners, the organizations, the fans themselves, that gives you a pretty good surrogate into the mental state of sport. Once again, I'm going to just say, I hope you're not counting on Bill Belichick for, for any advancement in that area because every single thing will just be like, yeah, well, okay. So Noah, you're describing a brave new world. Cause that's a parameter space that hardly anyone is talking about. And of course it's one of the strong points of Watson. It's facility with languages, multiple languages,

and people's ability to express themselves in that language. Interesting point though, Neil. Are you working with it in multiple languages? Because there's a definite contextual difference between every single language. And, you know, there are things that create different meanings, even though it's the exact same thing. We, we, but the course. Well, we're, we're definitely using AI in multiple different languages. And I think one of the things also to keep in mind in this time is eliminating bias from our models, right? AI comes with the same kinds of bias potentially as the people that are creating it. So we have- Did you just tell me that Watson is racist?

I did not. No, you did not. And let me just say this, Watson is not racist. Okay, people? Thank you. But just going back to an earlier example, when we're watching a tennis match and we're listening for crowd noise, Roger Federer's crowd noise is significantly different than the number 25 player in the world, right? Absolutely. And to be able to normalize that and not have Federer, bias the models when we're trying to equalize the AI highlight creation is really important. So that notion of de-biasing AI models into the future cannot be understated. And the point is, Chuck, not all bias is racist.

That's all. There are other kinds of biases as well. The only ones that count are. Okay. We got to call it a break right there before we go to our next segment. I want to thank Noah. Noah Syken.

Noah, S-Y-K-E-N. Noah Syken, it's great to have you on. Good to know that IBM is still at it, still working that Watson thing. And it'll be great to get further developments in the future from you. Yeah, thanks, guys. Appreciate you having me.

Excellent. Thanks for being on StarTalk. When we come back, we're going to further explore the role of AI in the fan experience on StarTalk. We're back. StarTalk. Sports edition. AI. AI. Just, it's both terrifying and exhilarating to think about what it could be for the future of sports. And we want to talk about now what might be the new emergent fan experience that exploits all the trappings of AI.

And we have as our special guest, Rajiv Maheswaran. Rajiv, welcome to StarTalk. Thank you. It's a, welcome to StarTalk. Thank you. It's a great pleasure to be here. You are CEO of Second Spectrum. And, you know, they didn't tell me this. I really prefer the CEO of the First Spectrum. Can we swap you out? We don't do seconds here. Your background is in electrical and computer engineering, University of Illinois. That's the place to do that sort of thing, leaders in computer science. And what is Second Spectrum?

What do you do there? So I think the name came up because we needed to come up with a name for the company. It was like we wanted to be the next way of seeing things, the next way of seeing things. We were wanted to be the next way of seeing things, the next way of seeing things. We were doing variants on the next way of seeing things, and we landed on Second Spectrum because we needed a name for the company.

And I think in general, we basically are trying to bring AI in general to sports. It was founded by myself, a colleague of mine, we were professors at USC in the computer science department in AI. USC, University of Southern California? University of Southern California. We didn't call it AI because it's, you know,

there was a time where AI was a bad word. Now it's a good word. And then a friend of ours named Jeff Su, and the three of us sort of started the company to bring sort of cutting edge AI to sports. Which of you, any of your three were athletes yourselves? Well, it depends on who you ask. In our own...

That's my line. Don't be stealing my lines. None of us played professionally or in college, but it was the three of us and two 19-year-olds, a 20-year-old and a 21-year-old who sort of dropped out of school from our research lab. But that being said, I think we, at some point, we have sort of over 30-some people in the company who played sports in college and then several professionals.

Okay, but they're not in charge. The ones in charge are geeks. That's what you're telling me. Pretty much. Okay. Thank goodness. I don't think we'd like to have it any other way. Guys, guys, I'm still here.

I'm still here. Oh, Gary. We got Gary here. Former pro footballer. Okay. Gary, no. Except for Gary, Chuck. Exactly.

We're definitely a big shot. There's no doubt about that. So you are partnered. It's one of your foci, if I said that right. It's one of your focuses, soccer. Absolutely. So we started out working in basketball and we were a basketball company for many, many years. But very recently, you know, we've always wanted to get into soccer.

And I apologize for calling it soccer for your international fans. But, you know, one of our board members said the top three sports in the world are soccer, soccer, and soccer and we have definitely found that out very recently. I like him. I like him a lot. So we made a big splash in soccer in the last year. We sort of got into, we sort of are league-wide partners

of both the English Premier League and Major League Soccer in the US and so we're super excited and we're going pretty hard into soccer and it's been very- And what do you do? What are you doing? So I think we do a lot of things in the sports space.

So I think one of the things you can think about it is there's, you know, people watch, historically, people watch games with their eyes, they understand it, and then they sort of create content, whether it's a reporter in the old days who went and watched a game, understood it, wrote a story, and then distributed it to people. And so TV came along and there was things like video where a person went with a video camera and then somebody edited it and told the story. I think for us, watching the understanding and the creating

is now we have sort of three parts that are done by AI. So watching is what we call player tracking or ball tracking where we have all these cameras in basketball and soccer stadiums where a machine can tell you like, this is Raheem Sterling, and this is Marcus Rashford, or this is Carlos Vela, and this is where he's standing, and this is how they're moving, and this is where the ball is. It just gets the positional locations of all the players, their identities, and the ball, you know, at 25 frames a second for every single game. Okay, but my brain does that, so you're not impressing me yet. That's right.

I don't need AI to tell me who my favorite player is, where he's standing, whether or not he has the ball. Right, and that's the biggest thing is that it was recorded, and people were recording all this location. A person could watch it, but they can't write it all down. But it was recorded, and it was pretty useless because the people said, I have a big pile of numbers that I can't do much with. So we came around and said, we're going to turn that into things that have meaning. So we can turn all this data, and we call it semantics, so we can turn it into the understanding

of what previously only a human could do. It's like, that's a pick and roll. That's a blitz defense. This is a between the lines pass. This is a progression, a buildup, organization, transition defense in the different sports. And we were starting to create words out of these numbers that everybody in the sports landscape started to understand. And then we started working with teams and leagues across the world. And the third part is creation. And we said, well, if a machine can understand, it can also create. And not only did we sort of start creating content, we could say a machine can understand, it can also create. And not only did we sort of start creating content, we could say a machine understands the full video.

And so it can put information, it can put probabilities or names or explosions or characters into a video, basically add special effects into video completely automatically. And so we basically do three things. We sort of watch the game with machines. We understand the game with machines, we understand the game with machines, and we create, at this point, augmented content for these leagues. So your experience could very easily translate into a completely different fan user experience. You could end up watching a game that with your technology, someone else sitting at home would not see the game that I'm watching via your AI augmented transmission. Exactly.

And I think one of the biggest things that we are trying to do is the fact that, you know, everybody wants what they want, when they want, how they want it. And the world's just getting better and better at delivering in it. So when you talk about content, you know, when you look at even the radio, it's like you sort of back in the day, you know, you, you go to a radio station, you like a song, you hope they play it, you know, two hours later. And then sort of things start to become digital and then machines start understanding it. And right now, nobody listens to it that way. In fact, whether it's, whether it's a music service or podcasts,

you basically build up what you listen to personally. And this is true whether it's content on the web or whether it's music or whether it's photos. There's lots of things that people experience in very, very different ways. Sports is this one last thing where it's like everybody gets to see the exact same thing, whether you're a person who's been rooting for a team for 30 years or it's your first same thing, whether you're a person who's been rooting for a team for 30 years or it's your first game watching or if you're a person who's into data or fun. Regardless of who you are and what your life is,

it's the same content for everyone. And the only way you sort of really scale personalization is to have a machine understand it and give everybody something else, something different. Wait, wait, wait, wait, my team. Wait, wait. I know exactly where Neil is going to go right now. No, no, just wait, wait. Go for it. Just stop. Go for it, wait, wait, wait, wait, wait. I know exactly where Neil is going to go right now. No, no, I'm just, wait, wait. Go for it. Stop. Go for it. Go for it. I don't understand

what you're talking about. I pay good money to see Gary O'Reilly perform on a soccer field. Let me think forward here. Linear TV is getting pushed over the edge by COVID plus everything else that's going on right now. We all want our own content universe, Neil, specific to us. The sort of, I get what I want when I want it. Explain to me how, what, I don't understand. I think that's what Neil is saying is, you got to give, and I think many of our listeners right now are having the exact same response, that's what Neil is saying is you got to give and I think many of our listeners right now are

having the exact same response which is can you please paint for us a couple of user experience examples that make this a worthwhile experience for me to be interested in I think that's what he's really saying thank you sure let's go from the simple to various so one is sort of a new fan and sometimes even experienced fans might not even know who the players are right and so just having gary's name over his head while he's playing it's like oh that's gary right so these are these right now you're describing google glasses okay go on basically like you make all the video like that and somebody who's watching a game and maybe you you don't you can't read their drills you don't know who they are you're

a new fan you might only know who the stars are you might you might only want to see who are the stars like gary and like put put that over their head and float them over their head so i know plus in games where they have helmets like football yeah that's even better you have to know their number or their body shape okay go on in the uk who like their job is to watch all primary league matches all day and they're like i don't know who number eight for Huddersfield is. And I wish I could, because I had to do a story about them. And so that simple thing would just makes everybody way more educated about, you know, who's actually playing and knowing when you're having a piece of content,

knowing who the characters in the story are, are kind of important. And right now that's a hard thing to do. And now the reason you might want to personalize that is maybe some people know all them. Some people only want to know a few of them. Some people might want to only know the ones who are on their fantasy team. So just even knowing who people are is a value add because that's such a big deal for somebody coming into a sport or getting to learn a sport a lot more. So that's level one. Like we can just tell you like who the characters in the show are. Right.

Then level two is like describing the conflict of the show so in basketball like there's these two people are trying to score and the basically the the fight is over the probability of the ball going into the basket based on all the different things you are like well we can float these probabilities over people's heads so if you're a data geek you might love to see all these probabilities floating over these these heads but somebody who isn't might not want that. So you can turn that on for yourself. And somebody who's not allowed these numbers, like turn that off.

So what you're saying is if I'm dribbling the ball, there are different parts of the court where I have a better field goal percentage than others. And you can watch that. You can watch that in real time. Come on. Okay, so keep going. And so somebody just like, and then a third person is like, you know what? I want to have, I watch sports just to have a lot of fun. I'm on Snapchat.

I'm on Instagram. It's like, why isn't it that every time somebody dunks a basketball, there's an explosion or lightning hitting the backboard? It's like, well, we can do that. And we can have all kinds of fun stuff that like, you know, NBA Jam effects and all kinds of cool special effects. Yeah, but will Pikachu show up? Pikachu? Ah. Yeah. We have many characters.

It will do if there's a Pokemon movie coming out. You know that it's going to... We need some augmented Pikachu. Oh, my God. I just realized that, Gary. But that's amazing. Chuck, so now who gets to monetize this? Exactly. But still have...

He just gave me level two. Give me level three now. All right. Buckle up. Buckle up. Well, can I give my level three? Can I give my level three? Because I'm sure that you might be able to do this.

Okay, so one of the things that I really like when I watch Amazon is the fact that while I'm watching a scene, I can find out everything about that actor. That actor, I can actually either pause the scene or in real time while I'm watching it, I can find out other stuff they've been in. I can look at their resume. I can see how they got this job.

It's when you're watching movies on Prime. Yeah, when I'm watching movies on Prime. On Prime. And I at first thought that that was kind of stupid, to be honest, because I'm used to watching television a very specific way. But then after like two or three experiences doing it, I find myself looking for that information.

I'm like, I'm pulling up. Who is this? What else have they done? What did it, you know, because, you know. You can't wait till the damn movie ends. I can't. And honestly, honestly, what you just said, Neil, is why I thought it was a stupid idea. I'm like, why can't you wait till it?

Just wait till the damn movie. Regine, give me level three. Forget Chuck. Give me level three. Wait, couldn't that be a level three? Chuck's right. Chuck's right in the sense like what we are doing is basically x-ray supercharged for sports. Like, really, that's what we are doing is basically x-ray supercharged for sports like really that's what

we're doing right there's another another which is like show i want to understand the game more right like you're right basically so in basketball like we can draw the play diagram under the players as it happens in in soccer by the way let me just interject real quickly there i don't think it was until um john madden when he became a sportscaster, that that game was ever unpacked by any sportscaster. Because he was the coach. He was an actual coach. And he empowered you to pay attention to players

that never otherwise got anybody's attention. What was the center doing after they hiked? What was the first blocking person? And he opened this hole and he came through and he came through. And it's, wow, the whole game was transformed for me hearing him comment. And I got to tell you, Neil, I am convinced, even though NFL will never say this, I am convinced that that is a huge reason why football became the number one sport in America. Because after John Madden started

doing that, every single analyst started doing that. And then the network started hiring former players to become the color analysts, and they all started doing the same thing. And as a result, you had a much deeper insight into the game. Okay, so now Rajiv, you're doing this with AI. Exactly, because I think, you know, John Madden is exactly right, because John Madden explained a very complicated game to a lot of people that brought them into that universe, right? So, for example, if we're watching TV shows, if I'm watching like a detective show or a romance, like I know the plot, I know the characters, I know what's supposed to happen. Like for a person watching a sporting event, unless you played the game or in,

you don't really understand it. And for games like basketball and soccer, that are more continuous, you don't have a lot of breaks to sort of explain a lot of the strategic elements. And so for us, when we can- Baseball, you have tons of breaks. Baseball is all about the-

Don't you go there again. Hey, can you put some explosions in baseball, please? Please, please? Please! Or a statistical chance that the base will blow up when you step on it. For you, I'll look into that. So we can draw that in soccer. We can draw the...

Soccer's so geometric. There are these lines and there's formations and overloads and all these things that you can point out. And Gary knows so much of soccer happens outside the ball that people just don't get. And Rajiv, to this day, I still don't understand offsides, just so you know, between you and me.

You can explain that by drawing a line that explains what offside is to people. It's like, wait, there's this line, right? They keep changing the rules, Neil. So if you don't feel that you understand it, just join everybody else. That's very kind of you. Yeah, okay.

Yeah. So Rajiv, someone is going to monetize this. Because as a viewer, I'm sat there, and as a viewer, they know everything about me. If I want a beer, if I want nachos, if I want a taco, if I want a burger, if I want to take a plane ride, if I want to buy ice cream or whatever it is. So, while I'm sat there, all of a sudden, all of my favorite things are going to explode in the soccer goal or come out of the hoop. It's the Powerade Power Dunk of the Quarter. And

Chuck's auditioning for the voiceover, which is brilliant. So who gets to monetize this? You, the teams, the athletes, the league, or all of the above? Hey, why can't we all eat, Raziv? Why can't we all get a little piece of this pie?

Yeah. I think in the long run, like, you know, it's a matter of how we get there, right? I think, you know, the sports landscape is changing significantly. I think, you know, that the way it's been delivered through linear TV into a television is going to change from being delivered over the internet into interactive personal devices. And I think you will see that same transition as you have seen, let's say the music industry or some other kinds of

video industries that are not sports change, where everybody in the ecosystem makes money in new ways as the distribution mechanism. Yeah, but Rajiv, they were biting and kicking and screaming in the transition. That was a bloody transition. I think sports will be the same. I think they will. But I think like in the end, you sort of scientifically have to go back down to first principles in the sense of like there is a very uh competitive market for people's attention and uh that is it's much more competitive now than it ever was you know when i was a kid i

would spend from like friday night until sunday night watching nothing but sports because that you know i was in the middle of wisconsin i grew up in the middle of wisconsin you know i was there i watched it uh but now it's like it's there's so much good stuff to to spend your time on like the star talk podcast and many other things where you have to fight. Thanks for the plug. That was very nice of you. It's true, right? Can you please add some explosions to our podcast? I mean, I think of it like

the world is much more competitive. It's a lot better for me as a consumer. I'm a science geek, so I love sci-fi shows. For me, when Next Generation came back, there was one show a week that I would live on. But now, if I look at all these whether it's podcasts or whether it's like for me like when next generation came back there's like one show a week that i would live on but now if i look at all these you know whether it's podcasts or whether it's you binge it i can i can spend my whole life doing nothing but sci-fi and i can spend my whole life watching nothing but robots and explosions and spaceships and it's amazing because i like the

world is much much more competitive and i think sports will have to go there sports basically had it and i'm a big fan okay so who pulled you out of your parents' basement? Somebody said you need a life. By the way, what was her name? I'm still in the basement. I'm just working different stuff. Excellent answer. We've got to close that segment.

But in a third segment, Rajiv, if you can hang out, it's when Gary, Chuck, and I just sort of chew the fat, and you sound like a good fat chewer. You can just join us. Unscripted, we just kind of go at it. That's great. Let's do it. When we come back, more StarTalk

Sports Edition. StarTalk, we're back for the third and final segment of StarTalk Sports Edition. AI. I've got Rajiv Maheshwaran. Help me there. Maheshwaran, but it's getting better all the time. Okay, Maheshwaran, thank you. I'm automatically approaching my name.

Very thank you. And you and all the Maheshwarans in Wisconsin, where you're from. Yes, that's surely a long lineage of Maheshwarans in the Midwest. Also in Wisconsin. So are your parents from Wisconsin as well? No, I grew up in Sri Lanka, which is a tiny island near India for those who don't know. But it's not so bad. Can I tell you one of the main reasons why I know Sri Lanka is because that's where Arthur C. Clarke hung out.

You know, Arthur C. Clarke, I know that he almost ran me over with his car once. I think that that's my closest I came. Really? Okay. Is that what we call a brush with fun? No, it's a brush with death. It is.

It's a brush with death. All right. But then from there, we moved from Sri Lanka to Wisconsin. So from a tropical island to Wisconsin. I used to make a joke that I could handle any temperature except room temperature. And then I moved to L.A. That's funny. I moved to L.A. and then I can cover, I've got room temperature covered as well.

There's room temperature all the way. There you go. In L.A. So we're just going to chew some fat. This is where we just shoot the shit. I'm just wondering, Rajiv, are you really transforming our experience? Are you the birth of something that in 20 years we'll look back on and say, I remember when? Absolutely.

I think I use that both internally and externally a lot. I think we would like to internally and externally a lot. I think that I don't, I think we would like to be the birth of it. I think that we would like to be the birth and we're doing everything in our power to make it so that like, remember when people used to say, remember when we all used to watch the same game

the same way? That seems crazy. And we- No, but wait, wait, wait. But sports is still event television. Yes, yes. No one DVRs their football game, okay? That just doesn't happen, really.

No, it's true. So there's both live and non-live. I think there's two things in terms of personalization. In live, you know, everybody can watch a live game in different ways. And even the numbers show that not everyone watches 100% of a game, right? Right. And so it's like everybody's watching different subsets of a game. So even live...

Right, but before we broke, you said that one manifestation of your company's inventiveness is to have, you'll have your own headset, so that's a very personal, private experience, because it will know what you need and want from the game that you're watching. However, isn't it true that sports is more communal? Absolutely.

So I think one of the things that we're going to do is like, the first thing we're doing is the one way to personalize it, have everybody have a different visual experience that's tailored to what exactly you want to see. But the next step is, you know, shared experience. So if you're just going to have a different visual experience, why can't I create that for myself? Why can't I share it with,

the next step is why can't I create that visual experience with some of my friends or many of my- Absolutely. So what you end up doing is kind of like Spotify-ing sports in such a way that you're creating or curating the visual experience and then you're inviting other people into a community to share the experience that you-

They're listening to your playlist. Right, they're kind of listening to your playlist while watching the game with you, which, by the way, already happens on Twitter during live events now. You have these small groups that pop up and they're talking to one another and they're sharing memes and they're showing replayed videos and things. That's Chuck during the Oscars. Chuck is like a thousand tweet comes out of Chuck's panel in the Oscars. What was she wearing?

What was he doing? Did he say that? Did he really mean? Where's Janet Jackson when you need her? Anyway. Oh, dear. This is a sports show, right? Yes. Yeah. oh dear this is a sports show right yes yeah so rajiv okay we are talking about being on the cusp of a new dawn a new age in the way that we consume but you guys don't sit on your laurels you go again so are we into voice

activation are we into touchscreen are we into a 360-degree soundscape? Can you tell us all your IP? Well, I think that you don't want to go too far ahead. I think you want to do is, you know, I think we are already sort of pushing the envelope quite a bit. And I think we do find that, like, we have to sort of stay with the pace at which the world is willing to change.

But I think, like, you don't want to go too far away from a core experience in terms of watching. And I think one of the biggest things to happen in the world is that everybody has these rectangles that they carry around with them. And some of them are small, medium, and large, but they're all these rectangles, and you can touch them, and you can interact with them.

And it's fundamentally changed the way people interact with computing and with content. And so for us, people are used to being, have the ability to be active. And so interactivity, being able to touch and interact with video is going to be a big thing. I think we've actually already built some prototypes where we can show what's possible. We're just waiting for the world to catch up a little bit. Because we're waiting for the world, more of the video to go on to digital than linear. It's already happening for people under a certain age,

but it's going to happen to the whole world very soon, and that's going to really unleash interactive video. Is there sufficient bandwidth to move video at the rates you need to accomplish your goals? Absolutely, because we've already delivered this stuff into the world, so it already exists with lots of our partners, and so we're already doing it. No, I meant literal bandwidth.

I mean, are you going to stream HD, you know, soccer game on a device of someone who's on a mobile platform? I mean, is this, this is real? This is real. It's already happening. It's only going to get more, better and more powerful.

So the degree of personalization and the degree of interactivity is only going to grow over time. If you think about it like the first time somebody streamed a game ever versus what the world is now, or they streamed a piece of video content to what the world is now, it's like we're on the tip of the tip of the iceberg. Okay, so in five years, what are we doing?

Oh, in five years, everybody sort of just watches games. Wait, in five years, you're the first trillionaire. Okay. No, no, no. Somebody's a trillionaire. But I think we just want to be part of the birth. I mean, I don't have any assumptions other than, like, we would like to move the world forward.

Like, we want to make a dent, and as big a dent as we can for as long as we can. I think it would be worthwhile if in five years, every game everywhere, not just the professional games, but pickup games are watched by machines, understood, edited, augmented, made interactive for people who care about that, whether that's a super fan or grandma watching my kid. I think that's the vision, is that it's not just for the pros. It's going to get cheaper and every game of everything played everywhere on the planet is going to be edited, augmented and streamed and watched. I think you're right. I think you're right.

And here's the only reason why I say that. I know five years ago, you saw the emergence of individual player profile videos showing up on YouTube and they were being used by scouts. And what happened was the players themselves, the coaches, started making the content. Instead of waiting for a scout to say, let me video this guy, they'd be like, no, let me take a video of all the stuff that I do great. I'm going that up on youtube and they would put together these compilations and and now it's almost like it's almost like its own little espn now you know the way it works so i could totally see this being the next incarnation but i still want to know i need i want give me some detail five to ten years from now and like the yankees are in the World Series,

let's say. That's not a stretch, of course. What do I experience? What am I... So, I mean, I think what you can do is, you imagine, like, I'm going to take the Clippers, because they're one of our... We launched this augmented view with the Clippers, and so you're at home,

the Clippers are on, you turn on your iPad, you come in the middle of the second quarter. The AI basically tells you, here's what's happened on in the first quarter and a half of the game. I catch you up at the important highlights. Then I take you to your story. Maybe you're watching the first quarter with your kids,

so you watch a mode with a bunch of explosions and characters and things like that. You go to bed. Second half, you get up. Pikachu. You need Pikachu. You need Pikachus in there. And then second half, you get on with your buddies, and you're a bunch of data geeks, so you turn on all the stats. The Riley. The Riley or whoever your favorite

is. Pick your favorite, right? And so you're watching it with your friends, and it's interactive, so maybe you play a bunch of games, and like you say, you're playing a bunch of games by interacting with each other on the screen, by tapping the screen and it's interactive. So maybe you play a bunch of games and like you say, you're playing a bunch of games by interacting with each other on the screen, by tapping the screen and making things happen. You have a social experience.

Then after the game, you know, maybe the next day or that night, maybe you're like, you watch your kids highlights from the Y game that she played earlier and you want to see her highlights, you share that with grandma.

Grandma makes some comments on that video. The next day you show it to your kids like, hey, I've cut up your pickup game with the Clippers game from last night. And you will have all these sort of mixing and mashing and sharing and augmenting. So the game is this living data entity. Yep.

Is that a fair way to describe it? That's a great way to describe it. I'm going to take that if you don't mind. Okay. Our lawyers will be in touch. So, just saying, do we have lawyers? Don't tell Rajiv.

Do we have lawyers? What I'm hearing is A, firstly, this can only improve the baseline of sporting talent and elevate it because you'll be able to see and then not repeat mistakes and learn from your successes. The other thing, Rajiv, you've talked about stories, your story, the story so far. And now what I'm thinking is if I want the John Madden character to be my narrator, I dial up that. If I want an NBA superstar to be my narrator,

if I wanted some guy from the hood, or whichever way, I want to go cartoon graphics. Gary, you never came from the hood. That sentence coming out of your mouth doesn't really work. Sorry, just thought I'd say that. I am not talking of a personal experience. I'm talking about someone who did. The hood coming out of the hood.

The hood, yes. However, I got to say, Gary, I wish you did come from the hood because I would have loved to walk through the hood and just heard somebody say, I sailed, man. I'd have turned around and said, someone's speaking my language at last. Fabulous. Let's have tea. So, but, So but Alright Rajiv

So if I've got a coaching Mode Set up On my viewing We literally have a coaching Yeah How You really do right

So if If in a soccer game There's a corner And I need to know What percentage In swinger Out swinger Short corner

Then I look at the players Then I can wonder If the first phase will be a counterattack or whether I contain. Then I need my... How far into the weeds can I get with that kind of knowledge? Because I know as the attacking team, I'm never more vulnerable than when I have the ball.

So now I am prone to counterattack. Or do I, with this information, load up my big guys because I have an advantage in size and putting corners that are to my advantage? How far away or how close can I get to all of that analysis in real time? We have all of that right now. So we basically have that and the managers have access to that information to study that before the games.

Now, from a fan experience, so we work with the coaches in the NBA and the managers in the Premier League. And they get that for the game prep. Now, does the fan need the degree, that degree of analysis? Maybe not all fans are like you, but one day fans will get access to that degree of specificity in real time. For us, we're just thinking about what are the first steps we can take to get people just accustomed to the fact that the game can look different than it did for the history of it, you know, since it's basically, since we started broadcasting games on TV.

It hasn't really changed that much. All right. Can we talk some real money here, Rajiv? Let's get into some real money here. How do we turn this into a gambling thing, man? That's where the dollars are. That's where it's going to happen. No, I think that if you think about it,

gambling was the first way people sort of personalized the viewing experience, right? Somebody basically was like, I care differently about this game than somebody else. Fantasy sports took that to a very degree. I have a very different caring about the set of players than some other person because we have different incentives.

So gambling was one of the ways of saying, I want to care more and I want to care more in a very different way than... Just to be clear so that I'm up to speed in case anyone else needs to be. But in fantasy sports, as I understand it, the performance of your players tracks their actual performance in the actual world. Yes.

And if they get injured, your player got injured. So you might be watching a game where two players on Team A are on your team and one player from Team B is on your team. And there's three players from Team B who's on your opponent's team, and so you have a very unique caring about that. It's a completely different thing, right. Yeah, it's a completely different thing, but it's very popular. It's people saying, I want to care about this game in a different way than people next to me.

So imagine you have a visual experience where the visual just tells you, this is what you care about and the presentation of the game you, this is what you care about. And the presentation of the game is going to reflect what you care about and not what the next person, what Chuck cares about, or Gary Kai is about. And also, maybe all of us are in a room and we want to watch this game together, and we all decide we're each going to pick a player or two and have a fun little game amongst ourselves.

Maybe we want each of our heads floating over the player to know whether I should root for that or not. And so gambling is a way of basically saying, look, I want the experience to be better. I'm going to bet to sort of care more about it, but then I'm also differentiating from other people who made different bets. And so. So it's not just whether it's better, it's, I guess, implicit is better, but more, more specifically, you're more engaged. Right. There's reasons to get deeper. There's a stronger connection to the fan. It's like going to the racetrack. Exactly.

And if someone hands you a ticket for the bet, you're going to be rooting for that horse in that race. Even though you have no real vested interest in the horse. You didn't know, but a minute before that, you didn't care.

I think one of the things and this is important, is that part of the reason people gamble is the raw content is not good enough I didn't make a bet before I went and saw Avengers Endgame in the theater, because I don't need to it's going to be good I don't think

I'm not going to make a bet about StarTalk, it's like, I know y'all are going to bring it. And then so, but I think it's, I just need to make it more interesting. And so like, it's just the fan singing thing is like, I need it to be more compelled, right? And so we can either amplify that, whatever compels you in a very unique way,

or we can find interactivity that helps you gamify in lots of different ways to make it more compelling for however much time and what form of time you're willing to give. By the way, I lost the mortgage payment on Thanos. It was a good bet, one in 14 million, you know? So this reminds me, just to close with some reflections here, baseball, was it in the 1970s?

There was some episode where Major League Baseball said people are too bored with the sport. And so they started adding bells and whistles. They would have contests between the innings where the fans would... I only know Yankee Stadium better than other stadiums, but between innings, there's a subway race, and it's all on the video screen.

And they race the B train against the D train against the 4 train. And it's stupid, but people get into it. And they have caps, and they move the ball around under the caps. And people are betting on it in between. And I'm thinking to myself, this is an early attempt to do exactly what you're trying to do. And it's early.

And to me, it was always clumsy what these were. But if you tune it and you make it sharp, you make it intelligent in the way AI is expected to be, there's no stopping what you've got brewing there. You're exactly right. I think they were just doing what they were doing with the limits that they have. What AI has done is remove those limits and saying, what can you imagine? If you can imagine, I can make it happen. And so when you're a trillionaire, we get 10%. I'll make that same deal vice versa when you're a dreamer deal so Rajiv

thanks for doing this and when you get more developments it would be great to have you back on and just check up on how it's going absolutely

I'd love to totally see that Gary always good to have you there pleasure my friend you bring athletic authenticity to these conversations

just so you know very kind of you to say unlike me or Chuck you to have you there. Pleasure, my friend. You bring athletic authenticity to these conversations. Well, that's very kind of you to say. Unlike me or Chuck. That's damned by faint praise. There you go, man. Chuck, always good to have you. A pleasure. This has been StarTalk Sports Edition AI.

Signing off, Neil deGrasse Tyson, your personal astrophysicist. Keep looking up.