Objective Deception

Objective deception is the problem named by 139. 为什么伟大不能被计划：有志者事竟成，吗？ through [[WhyGreatnessCannotBePlanned|《为什么伟大不能被计划》 / Why Greatness Cannot Be Planned]]: in a complex search space, an objective can mislead because the apparent direction of improvement is not the direction that produces the eventual breakthrough.

The source distinguishes this from ordinary goal failure. For simple tasks, a clear target can organize action. For open-ended problems such as scientific discovery, career direction, image evolution, artificial intelligence, or social policy, the target often becomes an imperfect proxy. Once that proxy is treated as a compass, people may follow a false gradient into dead ends while ignoring strange intermediate forms that could become useful [[SteppingStoneExploration|stepping stones]].

Key Claims

• Objective functions can be useful in bounded, well-understood spaces but deceptive in open-ended spaces.
• The closer-looking intermediate state may be a dead end, while the valuable stepping stone may initially look irrelevant.
• A bad compass can be worse than no compass when it creates confidence around a wrong gradient.
• Complex achievements are often easier to discover forward than to reconstruct backward from the final form.
• Objective deception applies beyond algorithms: careers, reading, research policy, and social metrics can all inherit the same failure mode.

Connections

• Novelty Search - alternative search method highlighted by the source.
• Stepping-Stone Exploration - practical response to objective deception.
• Campbell’s Law - social-metric version where proxy targets distort behavior.
• Non-Consensus Innovation - adjacent product and organization problem where existing metrics may not define the next category.
• Human Judgment Under AI - adjacent theme where judgment remains necessary when formal targets are incomplete.
• AI Engineering Thinking - AI work still needs experiment design while remembering that metrics can be incomplete.