Facebook’s poker playing AI can beat professional poker players

AI has become a market-boom recently and Artificial Intelligence Market will worth about $190.61 Billion by 2025. Facebook and Carnegie Mellon University their joint research has changed the way of playing poker. It will never be the same.

Facebook and Carnegie Mellon University have announced on Thursday that, they have developed Artificial intelligent power software which can play poker and it is also able to beat any poker professionals in six-player game no-limit Texas Hold’em poker.

Why an AI poker player?

According to the researchers, developing the AI poker player was quite challenging. Poker is complex, there are multiple players and there is very little information too and that’s what excites researchers to push the AI to its extreme limits.


Besides, the AI projects will reshape the future and that is the need for making such software. They have also given the player a name, it’s called Pluribus.

How will it impact the game?

According to Facebook officials, they won’t release the source code of the AI poker player because they are also concerned about the impact on the poker community.

It is certain that this software will impact the poker community as well as the way of the game. It might be positive or negative, it will be seen in the future. The players may learn from the software up to some extent, they will learn new strategies or the approaches, according to the anticipation.

What extra does Pluribus do?


AI is all about the neural networks, where the computers learn itself or sometimes with little help from humans. So it develops its way.

It is seen that Pluribus plays in a new way rather than the conventional way. As an example, we can mention that It is well known in the poker community that the Donk betting is a mistake but Pluribus seems to disagree with the community, as it uses the donk betting quite often.

The other example is, Pluribus balances its bluffs which makes it almost unpredictable to understand that when it will bluff. Besides, it varies the bid size widely. All these are quite unusual for a professional poker player.

Pluribus has played against human pros, for more than 12 days and 10,000 hands of poker. But that isn’t enough to make Pluribus. The more it will play, the more it will learn. The strategies of the poker players, as well as the Pluribus, will get better gradually once it becomes widely available to everyone.

The system requirement


Running the AI player needs cloud computation which costs only $150 which includes the training data as well as the computation. Thus it is quite cheap. Besides, as this software runs on the cloud, the software can be used in any device with little computation. Thus the system requirement is minimum.

What is the possibility?

We can use the AI poker player for the training and practice purpose by varying the hardness level. Which will help the poker community for sure? As an example, we can take chess. The grandmasters of the new generation have used the AI chess players to improve their own game. We can name Magnus Carlsen as an example.

Leave a Reply

Your email address will not be published. Required fields are marked *