L2.01
Openness of Natural-Language Commands and Its Cost
Cards in this group · 8
- L2.01.1Open input does not cue the range of capability
- L2.01.2Not knowing how to say it is the main barrier
- L2.01.3Differences in wording produce differences in results
- L2.01.4An empty box conveys no boundary; the first sentence is a guess
- L2.01.5Open input steers failure attribution toward “I said it badly”
- L2.01.6Synonymous phrasings yield different results, so users invent phrases to memorize
- L2.01.7Openness makes functions non-enumerable; the product can no longer show a complete capability list
- L2.01.8Error messages for open input stay vague because the system does not know what the user meant to do