Daryl Morey throws AI under the bus for terrible Sixers season

There have been a lot of memorable AIs in Philadelphia 76ers history, from Allen Iverson, to Andre Iguodala, to Al Horford if you used an uppercase “I” in a typo instead of a lowercase “l.” But as the Sixers’ deeply cursed season winds down, general manager Daryl Morey might have just revealed the AI that will go down as the most infamous in Philly: Artificial Intelligence.

As the 22-43 Sixers wait for their injury-riddled, accidental-tank campaign to wind to a merciful end, the man who put the team together was speaking on a panel at the MIT Sloan Sports Analytics Conference, and when asked by host Pablo Torre if he uses any AI Large Language Models (LLMs) when making basketball decisions, Morey said that they are unequivocally a part of his process:

When I saw the first half of the above post, I thought Morey was admitting they Allen Iverson was doing some consulting for the Sixers, and thought “oh, that’s pretty cool, I bet a franchise legend has some unique insight. Maybe he can help turn things around.” ... and then I saw the second half and realized he meant he was asking computers to help him figure out which basketball players are good.

Now, to be fair to Morey, he does say that said LLMs only get a vote in basketball decisions, similar to a scout. They don’t get to make final calls, and how much they are trusted is constantly reevaluated based on tracking their success or failure. Even as big a nerd as Morey isn’t handing the keys fully to the computers, at least not yet.

But with that said... this Sixers’ implosion actually makes a lot more sense if you see the team through the lens of AI. After all, what do large language models do? Aggregate previously held opinions and facts written out by others and spit them back out like a high school student lightly plagiarising an internet article to complete their homework assignment.

And if you look at this Sixers team as if it was built by typing “are Andre Drummond and Eric Gordon still good at basketball?” into ChatGPT... of course the computer was going to think they were decent role players based on past production. Obviously an AI model was going to think Paul George and Joel Embiid were worthy of max contract extensions because of their prior production while underrating future concerns, because it can’t project those with any guaranteed certainty. It can just look at what players have done in the past and predict that if these trends continue, disco sales will rise forever and Joel Embiid will eventually be bound to make a Conference Finals. Really, the horrendous roster the Sixers built might actually be the best argument that AI isn’t all it’s advertised to be yet.

Unfortunately though, just like with the eventual dystopian societal effects AI, it won’t be the people like Morey that replace humans with computers to create a crappier product that will pay the price; it will be Sixers fans. It’s always Sixers fans.