Hello Reddit! We are PhD student Noam Brown and Professor Tuomas Sandholm at the Computer Science Department of Carnegie Mellon University. We do research on developing AIs that can reason about hidden information (which is widespread in real-world strategic interactions). Earlier this year we built Libratus, the first and only AI to defeat top humans in no-limit poker. We played four of the world’s best pros in a 120,000 hand, 20-day Brains vs. AI match of heads-up no-limit Texas hold’em, with a prize pool of $200,000. The AI won the match decisively, winning a combined $1.8 million (at $50/$100 blinds). The victory was statistically significant with a p-value of 0.0002. The details of the bot were just published in Science Magazine! We’re here to talk about Libratus, the competition, what this means for the future of AI, and any other questions you might have. We’ll be back at 9 am to answer your questions, Ask us anything! EDIT: We’re closing the AMA. Thanks for the questions everyone!