The rise of the robots continues following the development of an artificial intelligence capable of taking on, and beating, humans at the game of poker.
Or, at the very least, a computer program capable of deep enough strategic reasoning based on statistical analysis to be able to adapt and respond to human players at Poker, one of the most human of games.
It follows a BrainsVsAI tournament arranged by Carnegie Mellon University against four of the world's top poker pros. Dong Kim, Jason Les, Jimmy Chou and Daniel McAulay.
They took on Libratus (Latin for "balanced but forceful") at heads-up, no-limit Texas Hold'em in a follow-up tournament to one held in 2015, when the humans humiliated Libratus's predecessor, Claudico.
This time it was the humans who were not only humiliated, but even demoralised after losing $1.7m in fantasy chips (ie: not real money) to Libratus.
The AI was developed by Carnegie Mellon University professor of computer science Tuomas Sandholm who built Libratus with PhD student Noam Brown. "This challenge is so huge and complicated that it's been elusive to AI researchers until now," said Sandholm.
Unlike in conventional poker tournaments, every night the players shared tips and ideas about how they might defeat the computer, while Sandholm and Brown used a meta-algorithm to analyse the weaknesses that the poker pros had exploited in the AI's own game, and tweaked it accordingly.
Nevertheless, Sandholm says that he wasn't confident that the AI would be able to crush the poker pros quite so easily.
Les had played Libratus's predecessor, Claudico, just two years earlier. Playing Libratus, though, turned out to be an entirely different experience. "Libratus turned out to be way better than we imagined," said Les. He told The Guardian: "If you play a human and lose, you can stop and take a break.
"Here we have to show up to take a beating every day for 11 hours a day. It's a real different emotional experience when you're not used to losing that often," said Les, who is probably now mulling alternative careers that AI won't be able to do for the time being.
While poker might seem like the ultimate human game, in that even if you have a poor hand you can bluff your way to a win, it is ultimately based on statistical analysis: the probability that one hand might be better than your opponents', and knowing when it's just best to fold.
What is key, though, said Sandholm, is the AI's ability "to do strategic reasoning with imperfect information" better than the sharpest humans.
Mark Vartanyan was working for Norwegian e-healthcare firm Dignio when he was arrested
Samsung can't see a way to profitably compete against Amazon and Google
Fix being rushed out - but not quite as quickly as an ambulance to an emergency
Massive miner Rio Tinto claims 20 per cent of pit-to-port train kilometres in Australia are now driverless
Rio Tinto today, TfL tomorrow?