Deep Mind by Foxymoron

 Google acquired the company Deep Mind in 2014. A chess-player who once attended coaching sessions I ran, Demis Hassapis, founded it. The programs Alpha Zero and Alpha Go have become the best "players" of  chess and go respectively, by teaching themselves. They have avoided the number-crunching approach of previous AI software and just learn by playing millions of games on their own. Already humans have no chance at either of those games. Curiously, bridge (and to a lesser extent poker) are still defended by the human race. I recall composing a chess study in 1981 and it is coming up to its 40th anniversary in a month or so. I plugged it into the best chess program today, Stockfish, and it still cannot solve it. Something to do with the horizon effect. It is worth referring to briefly. The full solution is in Wikipedia if you are interested. Just search for Grotesque (chess).

The interesting aspect of this position is that it does not require any great chess ability to solve it, although such luminaries as Kasparov, Speelman and Nunn only cracked it in about half an hour. But it does require a lot of logic. 

In bridge, they also lack something called "intelligence". However, today, Robot West and Robot East surprised me by having a remarkable auction. If I did not know better, I would think they had seen each other's hands.

I regretted not trying something esoteric as North, such as overcalling 1NT, when the robots galloped into the good Six Diamonds here. And on a club lead, the Robot played it well too. It won with the ace, cashed the ace of diamonds, and crossed to the ace of hearts. Now it cashed the ace, king of spades pitching losers from dummy, crossed to the king of hearts and ruffed a heart with the nine of diamonds. All the defence could score was the king of diamonds. This line works about 74% of the time, whenever there is a singleton diamond honour or the diamonds are 2-2. Could the robots have really conducted such an amazing auction knowing what they are doing? Well, I think that Robot West was stuck after the negative double. Now 2NT shows a weak NT and 3NT "shows" 18-19 so it felt that it had to "make up" a 2D bid. It explained it as 5+ spades, 4+ diamonds, and one could almost see Robot East having an orgasm when it looked at its seven-card support. And when it bid 5D, West now cued 5H, presumably looking for a grand, but Robot East decided it did not have enough for that. 100% for EW, unsurprisingly.

Some rough edges, but the robots are getting there.