Mastering the art of data: the skills that will get you noticed

It’s almost a decade since a much-shared Harvard Business Review article declared data scientist the sexiest job of the 21st century. But is working with data still as novel or as exciting as it was then?

The rise of digital currencies in the emerging world

As of September 7, 2021, El Salvador will officially recognise Bitcoin as legal currency. This is a remarkable move, and one that will likely stimulate a huge amount of economic growth as Salvadoreans leapfrog from a largely cash-based society straight to a system characterised by the frictionless, finger-snap efficiency of digital currency.

Trump’s plans to ban Chinese app WeChat blocked

A federal judge has temporarily blocked President Donald Trump’s executive order to ban WeChat over concerns the ban would threaten users’ first amendment rights

Oracle returns to growth

The software maker’s quarterly revenue exceeded analysts’ expectations, as remote working drove greater demand for its cloud-computing services

Artificial intelligence finds a new game to beat humans at: poker

Chalk one more up to the robots as an AI system rakes in the cash from poker pros

A computer beating humans at poker represents a significant achievement - poker is not a game with limited parameters like chess or draughts

AI systems have long been able to out-think humans at games like chess and backgammon, but now they can out-bluff them too. In a tournament against a number of poker pros, Carnegie Mellon University’s AI system Libratus has so far been making short work of its human competitors. As reported by The Verge, at the 80,000-hand mark of a 120,000-hand tournament, humans are down roughly $750,000 to the machine.

Libratus’ playing style in the tournament has so far been characterised by unpredictable aggression and brazen bluffs. The AI system also tends to over-bet on hands it feels are strong in order to win a comparatively small amount. This style of play is uncommon among human players, who are unlikely to risk much more than is necessary. Unconstrained by this thinking, Libratus has demonstrated it is happy to go all-in on comparatively small pots to spook opponents. This unpredictability has left human players struggling to find a ‘tell’ for its bluffs.

Libratus’ playing style in the tournament has so far been characterised by unpredictable aggression and brazen bluffs

An AI system overcoming humans at poker is, in some ways, a more important achievement than doing the same in other games. Draughts, chess and go, for example, have limited parameters and allow players to have the same, perfect knowledge of the game at all times. There is no chance involved or any information that a player does not have access to. Poker, on the other hand, is focused on odds, prediction and an imperfect knowledge of the situation. It’s also a game of interpreting actions and quickly calculating risk versus reward, much like most real-world scenarios.

The researchers that built Libratus hope the system could eventually be used to solve more complex real-world situations, including issues of military strategy or cybersecurity. However, the AI system is still not completely independent of its human programmers. As reported by Wired, the researchers could be altering the programme each night of the tournament to sure up any weaknesses exposed. If they aren’t, their secrecy surrounding the workings of the machine is likely impacting on the behaviour of the human players. Like any poker game, what is happening off the table is just as important as the cards that are being dealt.