Member-only story
Computers now exceed people in detailed memory capacity
In the past, computer programs for playing chess relied on being able to look ahead farther in moves than most people can. That is, starting with the current configuration of the chess board (a.k.a. the current board position), they could generate many possible moves of their own and many possible responses by their opponent, and then look at the resulting new board positions, evaluate them, and repeat the look-ahead process, before ultimately making their next move.
Now with much larger memory capacities, computers can also be programmed to remember huge numbers of specific chess board configurations and moves that have occurred in chess games that have been played previously by people or computer programs and have been recorded in written form in texts available on the internet. (In that respect, such databases are like large language models for generating English text.) To make a new move, the computer can simply compare the current board position with that huge database of board positions.
If the current board position is found in the database, the computer can then make a move from that position that frequently led to a winning outcome (information that could also be recorded with each board position and move in the database). Only if the current board position is not found in…
