Lengpudashi, or “the cold poker master” is an artificial intelligence program that was designed by Professor Tuomas Sandholm and a graduate student Noam Brown from Carnegie Mellon University. Lengpudashi is currently located in Hainan, China, where he is “celebrating” a smashing victory against Team Dragon, which consists of three top poker players and three computer scientists. Alan “Yue” Du, the first man from China who managed to win the bracelet in 2016 $5,000 No-Limit Hold’em World Series of Poker (WSOP), was the head of Team Dragon and he included engineers, computer scientists and investors in his team in order to defeat the cold poker master, but, as many supposed before the exhibition match started, the humans never had the chance. After 36,000 hands over the course of five days, Lengpudashi defeated Team Dragon with nearly $800,000 in chips and took home $290,000.
This is not the first victory of AI against team of humans, because the same duo that developed Lengpudashi also created Libratus, which managed to defeat Jason Les, Daniel McAuley, Dong Kim, and Jimmy Chou in the Brains Vs. AI challenge that took place at Rivers Casino in Pittsburgh in January. However, the earlier version, called Claudico, failed to defeat the human team in 2015.
Kai-Fu Lee, head of Sinovation Ventures company, who organized this exhibition match, also had the chance to taste the strength of Lengpudashi, since he challenged the cold poker master to about 30 hands and got butchered. Lee said it was fun to be defeated by an Artificial Intelligence program, but chances are we will stop laughing soon enough, because there is a possibility that in the not so distant future machines will take over. The story from Terminator does not seem so unbelievable now, does it?
The biggest problem of AI programs is control. Lengpudashi and his predecessors are currently safely locked away in CMU, but no one can guarantee that the cold poker master won’t decide it was locked away for too long and it is time to leave. Furthermore, any day now Sandholm and Brown, or some other computer genius will create an artificial program that will decide that the best way to defeat humans is to destroy them. This is why computer scientists must figure out a way to control artificial intelligence which can be superior to human kind sooner than we think.
The best thing about poker is that there are different kinds of challenges players must face in order to win. Unlike chess or similar games, where everything is on the table, you have to battle against bluffs, bets and folds without even seeing your opponent’s hand. We can understand how an AI program was created in order to master the art of chess, but how in God’s name does a computer program know when to fold? After being unsuccessful with Claudico, Sandholm and Brown developed two smashing AI poker programs that defeated the best poker players in the world, so they literally found a way to make their computer children think like humans. A scary fact that made us wonder what is the next step in the Artificial Intelligence evolution. If we face the danger of being controlled by robots, all of us better have an Arnold Schwarzenegger for protection.