Adaptive developmental plasticity evolves when cues reliably predict fitness consequences of life-history decisions, whereas bet hedging strategies are expected to evolve when environments are unpredictable. In changing environments, these two strategies could be favoured, though their importance remains poorly understood. Cane toads were introduced in Australia 80 years ago, and provide an ideal model to investigate the effects of environment predictability on life history evolution. Indeed, the colonization of Australia by cane toads is an ongoing process, so that it is possible to compare the life history of individuals currently colonizing new (and unpredictable) environments in Western Australia with individuals from long-colonised areas in Queensland. Here, we measured variation in early larval growth and development in response to variation in density and food level in 10 clutches from the colonization front and 12 clutches from long-colonised areas. Clutches from the invasion front were less plastic than clutches from long-colonised areas in response to density or food variation. In contrast, clutches from the invasion front had higher within clutch variation in growth rate under constant conditions, suggesting that they favoured bet-hedging strategies. We also detected a negative correlation between within clutch variation in growth rate under constant conditions (= bet-hedging strategy) and reaction norms in response to density or food (= phenotypic plasticity), suggesting the existence of a trade-off between these two strategies. Our results thus show that toads on the front invest in a bet-hedging strategy, likely as a way to adapt to unpredictable conditions, whereas toads from long-colonised areas invest more in plastic responses. Driven by the ongoing colonization history of cane toads in Australia, these life history strategies have evolved in less than 80 years.