Centaur Teams For Ambiguous Problems

The best chess player was not an AI.
The best chess player was an AI with a human (a Centaur Team).

Chess engines (AKA chess AIs) determined how good/bad a move is by analyzing the possible positions resulting from that move, and then comparing them to the other possible positions resulting from other possible moves.

Chess engines didn't genuinely understand position dynamics like a Grandmaster did, and that led to some positions where a Grandmaster has a better strategy than an engine due to the limits of such brute force.

Chess engines didn't look at position and think, 'Hmm, this is a really closed and even position. The only way to make a difference is to spend about every 3rd turn moving my king to the other side of the board to have just enough attackers on his B pawn to break through. It's probably going to take more than 10 moves to get my king over there.'

Instead, the chess engine would look at that position and never realize that it needs to move the king because its brute force method wasn't able to go deep enough to see what would happen if the king made it to the other side of the board.

Chess engines have since been updated to recognize position dynamics, but that's beside the point.

Life is not an unambiguous board game.
Life is a super extremely ambiguous reality.

A human with an AI will outperform just an AI when it comes to the more ambiguous problems.

AI on its own would be much less capable at devising a list of groundbreaking features to incorporate into an app than a human with AI, and that's because AI doesn't have genuinely understand what it's saying like a human does.

Until AGI exists (which we aren't even close to yet), a human with an AI is better than just an AI at life/business.

SS5.org/2025/05/15/centaur-teams-for-ambiguous-problems