On the latest episode of Pablo Torre Finds Out, Pablo is joined by Daryl Morey and Sendhil Mullainathan to explore how AI is already transforming the sports world.
Artificial Intelligence is here to stay, but how is it being used in sports? To better understand, Torre asked Morey, the president of basketball operations for the Philadelphia 76ers, how long he has used large language models and AI in his executive role. Morey explains that he had been using large language models for about three years before OpenAI. When he reflects on it, he was unaware of how popular it was until the public caught on. In terms of using artificial intelligence, Morey shares that in his front office, he has used a version of AI and machine-learning models for a long time. He contends that AI is just forecasting, and his job is based on predicting and making good decisions. The 76ers typically use AI or machine learning to forecast draft picks, utilizing forecasting techniques such as traditional and linear regressions. He uses the example of the pick-and-roll and how computer vision is used to identify it. He jokingly adds that even humans can’t agree on what a pick-and-roll is and what it is not. He acknowledged that we generally understand it as one guy hitting another guy, which keeps him from guarding a guy, and then something happens.
76ers president Daryl Morey says he asks A.I. for input when making team decisions.
"We absolutely use models as a vote in any decision." pic.twitter.com/RaERlg1s7N
— Pablo Torre Finds Out (@pablofindsout) March 13, 2025
The reality of phantom pick-and-rolls, where a player will act like he is screening and then goes away, makes it difficult for humans to arrive at a precise definition. However, Morey contends that computers do a pretty good job at identifying such a fake but are stil not perfect. He added that they also utilize LLMs for coding and speeding up the writing process. Torre followed up by asking plainly when Morey is making a decision for the 76ers, is he is asking the LLM for input? Morey’s answer was twofold. He acknowledged the 76ers use models as a part of any decision, but how much of that decision depends on how much the assumptions behind those models don’t change and what is their past success rate. He added that LLMs do fairly well with predictions but are still not beating human super-forecasters. They add signal over scouts, and Morey says that they treat them like one scout, and over time, if they have a successful track record over scouts, he will allow for multiple LLMs to have a vote.
Torre flips the conversation over to Mullainathan and asks him about the ambitions behind artificial general intelligence thinking like a human. Mullainathan says that AGI sounds unambitious. He says we already have people who can think like people. Mullainathan believes we need algorithms that can do things that people could never do, algorithms that can notice things we never could. To his point, the AGI component is too tricky because we have had algorithm vectors that have identified things we can’t and stressed that is the path to improve performance. However, when it comes to using AGI to do exactly what humans do with very high fidelity, he does not believe there is much to gain from that.
MIT professor Sendhil Mullainathan calls Artificial General Intelligence (AGI) — a form of AI that would possess human-level cognitive abilities — "a shockingly unambitious goal." pic.twitter.com/XzDJBTifyZ
— Pablo Torre Finds Out (@pablofindsout) March 13, 2025
Getting back to sports, Torre discusses the divide in sports between entertainment and efficiency. He highlights how the NBA is working to try to keep up with those who innovate. Further, Torre paints a proverbial segmented picture where the players are in one silo trying to figure out how we catch up while Morey and the league office are in there are own separate silos. Torre asked Mullainathan if there was a way for each entity to unite and be genuinely aligned on the goal of entertainment. Mullainathan admits that as a fan, he understands that winning matters, but he gets pleasure from other things outside of that, whether it is the underdog story or the player scoring in a way that no one could predict. He added that the happiest he has ever been as a Sixers fan was during The Process years. He would watch every game, and he wishes an algorithm could help us understand what makes for a great storyline.
Morey’s point of view is different as the sun rises and sets on winning championships. Still, he reflected on two seasons ago when the Sixers were up 3-2 on the Celtics and eventually lost the series in Game 7. Sixers fans were not happy, as you can imagine, but fast forward to the Knicks series in which the Sixers lost in six games, and even though the Knicks were not as good as the Celtics team, Sixers fans were less angry. Mullainathan put on his behavioral science hat to speculate that negative emotions are acceptable and that being disinterested is dangerous. He added that it is easy to dismiss sports as entertainment, but what else do we have that brings people from all walks of life together to communicate and have a meeting of the minds? He believes sports’ core contribution to society is that it provides conversation. He went back to that Celtics series where the Sixers fell short and noted the overall sentiment from Sixers fans like himself was frustration and anger, but in the end, fans are bonded by that emotion.
Check out the below episode of Pablo Torre Finds Out. Stay tuned, everybody, stay tuned!
SHOW BREAKDOWN
How Artificial Intelligence Is Already Changing Sports, with Daryl Morey and Sendhil Mullainathan
- Everyone is talking about A.I. now, but what we haven’t heard is an updated, informed discussion of what, exactly, it will do to the sports world. So in our first on-stage episode, at the MIT Sloan Sports Analytics Conference, Pablo hosts an actual verified genius… and a pioneering NBA executive who gets to be sarcastically referred to as one. We discover how the Philadelphia 76ers already consult large language models, why artificial general intelligence is actually unambitious and much more. But rest assured: At this moment of tension between entertainment and efficiency, the forecast is Kyle Lowry with a chance of Gasol.
VIDEO:
Check out the Pablo Torre Finds Out YouTube page for more and subscribe to the show on Apple Podcasts, Spotify or wherever you get your podcasts. You can also follow the show on X at @pablofindsout and Pablo at @pablotorre.
