Why gig workers say yes to some jobs but not others

A neural network model trained on 36,891 gig transactions reveals workers exhibit hysteresis: price cuts suppress acceptance more than equivalent increases boost it, creating a gap where the same wage yields different decisions. The model identifies which jobs can survive 7-31% wage cuts while maintaining 97% acceptance, potentially saving platforms 21% in labor costs while filling 10% more gigs. This only works because workers have psychological thresholds, not rational equilibrium points.