기술의 세 층

# Craft 의 세 layer + humble-vs-over-confident axis.

class PatternRecognizer:
    """세 구조적 attribute + 한 craft attribute."""

    def __init__(self, name):
        self.name = name
        self.generic_skill = True               # layer 1
        self.no_single_answer = True            # layer 2
        self.weapon_for_seers_only = True       # layer 3
        self.confidence_interval_visible = True # craft attribute

    def emit_prediction(self, case, matching_attributes, divergent_attributes, confidence):
        if not self.confidence_interval_visible:
            # Over-confident class spotter — recognizer 안 Fourth K trajectory.
            return f'{case} 는 THE {matching_attributes[0]} 패턴.'
        return (
            f'{case} 가 {matching_attributes[0]} 에 '
            f'{matching_attributes} 로 {confidence:.0%} confidence match; '
            f'{divergent_attributes} 에 divergent; '
            f'6개월 후 recalibrate.'
        )


humble = PatternRecognizer('humble')
over_confident = PatternRecognizer('over-confident')
over_confident.confidence_interval_visible = False

print(humble.emit_prediction(
    '회사 X',
    matching_attributes=['emperor pattern'],
    divergent_attributes=['founder 가 1-on-1 에서 여전히 dissent 요청'],
    confidence=0.65,
))
print(over_confident.emit_prediction(
    '회사 X',
    matching_attributes=['emperor pattern'],
    divergent_attributes=[],
    confidence=0.99,  # over-confident 한테 visible 아님; output 에 surface 안 됨
))