liveBook · Manning

Adaptive imitation

As explained in Machines that Think

Adaptive imitation is behavior that preserves an appearance of capability by adjusting successfully when conditions change.

Adaptive imitation describes performance that remains effective as conditions change, rather than an appearance of competence limited to familiar cases. Because flexible behavior can persist while the underlying mechanism differs, it can be difficult to distinguish imitation from ability. section 9.1.1 section 9.6

Chess and language models

A chess engine can adapt to unfamiliar positions without playing as a human does. Large language models likewise adapt across topics, formats, styles, and tasks, though hallucinations and rule breaking remain limitations. section 9.1.1 section 9.1.1

Figure 9.1 Brittle and adaptive imitation. A brittle system may appear capable when tested on familiar cases, but its performance collapses when the situation changes. In chess, this would be a system that plays well only in familiar positions. An adaptive system continues to perform in unfamiliar cases, making the difference between imitation and ability harder to draw.

Contrasts performance that collapses in unfamiliar cases with performance that adapts to them. section 9.1.1

How Machines that Think covers it

How deep it goes
Defines adaptive imitation through contrast with brittle imitation and illustrates it with chess and language models, without resolving whether it amounts to intelligence.
Examples
Uses a chess engine facing unfamiliar positions and large language models adapting across topics, formats, styles, and tasks.
What you can do after
Evaluate/understand flexible performance without treating it as proof of human-like understanding.
How it connects
Connects brittle imitation, machine intelligence, large language models, and the distinction between observable performance and underlying mechanism.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Machines that Think. The section links point to the parts of the book it draws on.