Tag
This paper systematically evaluates human creativity tests for LLMs and finds they fail to predict scientific ideation. It introduces the DRAT, a new test that combines convergent and divergent thinking to reliably predict scientific ideation ability in language models.
Research published in Science analyzes 12 million scientists, finding that while capacity for connective novelty increases with age, the ability to produce disruptive innovation declines.
The article argues that true AI creativity may require subjective experience and intrinsic drives similar to human emotions, raising significant ethical questions about creating sentient-like systems.
A user shares their experience spending ¥3,000 worth of tokens to create an AI video, suggesting that creativity, rather than compute power, is the true bottleneck.
The article comments on how AI video generation technology has reached an extremely high level, highlighting that aesthetics and imagination have become the new key values.