Unverified50% confidenceFactExact time
In chess, the 'centaur' (human + AI) model was once regarded in the late 1990s to mid-2000s as the optimal solution for board game play, but with the emergence of self-learning systems such as AlphaZero, pure AI playing strength quickly opened a gap over any human combination, and by the 2010s this model had basically lost significance in top-level competition.
1
Sources
50%
Confidence
Long-term
Relevance
10/13/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn chess, human-AI collaboration as a whole surpasses pure AI or pure human performance78% similarUnverified自我博弈作为AI训练范式的著名案例是DeepMind的AlphaGo Zero，通过与自身历史版本对弈无需人类棋谱即超越人类围棋水平73% similarUnverifiedEquating game completion with AI already possessing general long-horizon planning capabilities is a misreading.68% similarUnverifiedExtrinsic integration is essentially using the game as a reward to drive learning; the two are logically separate, and this is the shared reason for the failure of a large number of early educational games.67% similarUnverifiedChess is a single domain with closed rules and clear objectives, whereas the real economy consists of countless heterogeneous tasks, each with different human Elo thresholds and different crossing points in time.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1018221API
curl https://kongchang.com/api/v1/knowledge/claims/1018221MCP
get_claim(id=1018221)