Artificial Intelligence that has beaten ten different poker players in the US & China since the turn of the year could be on our smartphones within the next five years says, expert.
It’s ironic that the artificial intelligence (AI) used to power our self-driving cars will be used to save lives, while SIRI continues to put mine at risk when I drive as it never bloody works.
SIRI is an example of what AI geeks call Artificial Narrow Intelligence (ANI) or Weak AI. ANI technology is so ever-present in our lives that we don’t even recognise it.
Other examples of ANI are Deep Blue, the chess computer that beat Garry Kasparov in 1997; Google’s AlphaGo that beat Lee Se-dol in March 2016, and Libratus, the poker playing ANI that defeated four No-Limit Hold’em experts in Pittsburgh at the turn of the year.
An ANI is AI that can master a single task. Artificial General Intelligence (AGI) is the ability to perform a variety of tasks as well as or better than flesh and blood on an intellectual level. Finally, you have Artificial Superintelligence (ASI) and when that happens we are all up that creek everyone talks about without a paddle.
Yesterday, an ANI that specialises in poker beat a team of six poker players in the Chinese Island of Hainan. The AI is called Lengpudashi (Cold Poker Master), and it didn’t just beat its human opponents. It annihilated them.
The margin of defeat was shocking with Team Dragon (the humans), losing to the ANI by $793,327 play money chips over 36,000 hands or 220 milli-big-blinds. When Libratus beat the four human pros in January, it did so by 147 milli-big-blinds.
The news of the victory spread through all the world’s major news stations. And the significance is because the poker ANI is leaning into the abilities that an AGI will be expected to have such as learning how to make very human-based decisions such as bluffing.
Lengpudashi’s parents are Carnegie Mellon University Professor Tuomas Sandholm and his student Noam Brown, the pair who also created Libratus. The victory in China has netted the pair $290,000, which they no doubt will invest in their goal of creating an AGI that will spend its time improving the world rather than sitting on its circuit filled ass and playing poker all day.
Talking about the victory in China, Brown told the press that the total cost of the ANI was below $20,000 and, ‘within five years this could be running on smartphones.’
They had better hurry up.
If SIRI has its way, I will be dead by then.
Google’s AlphaGo to Take on World #1 Go Player
Not many companies in the world are investing more money into AI development than Google. A year ago, an ANI created by their DeepMind team defeated the legendary South Korean Go champion Lee Sedol, but the humans are back with a different challenge.
Go World #1 Ke Jie will take on AlphaGo in a best of three matches at the Future of Go Summit in China May 23-27, in what the Google Deepmind team believe will be an ever sterner test than the one delivered by Sedol, and for a good reason.
Rather than the 3,000-year-old game of Go falling apart because an ANI defeated one of their top dogs, analysis of that groundbreaking game has led to improvements within the Go community, with some players now believing that no move is now impossible thanks to the innovative play of the ANI.
I wonder if the same will be true of poker?
