Okay, this one affected me personally.
Google's Deepmind team adapted their AlphaZero approach for the games of
chess and shogi and dropped the bomb already on the 5th of December.
For chess they trained the Deep Neural Network for 4 to 9 hours on an
cluster with 5000+64 TPUs (1st+2nd gen) and reached super human level.
Unlike in Go, they did not compete with humans, cos chess engines are already
on an super grandmaster like level, no, they did compete with the worlds strongest
open source engine - Stockfish, result:
This is definitely a smack in the face for all computer chess programmers out
there. Next stop Neanderthal Man.
So, with thanks to the Krampus,
+1 points for the Singularity to take off.