This paper presents a theoretical model of meta-cognitive adjustment in intertemporal choice. We consider agents who are aware of their present bias—captured by quasi-hyperbolic discounting—and attempt to correct for it by applying a self-chosen adjustment parameter, λ. While modest correction improves utility, we show that overcorrection, where the agent becomes excessively future-oriented, may result in outcomes inferior to those under the original biased baseline. The model characterizes optimal correction as occurring when the agent faithfully incorporates their time preferences. Attempts to fully "cancel out" present bias (e.g., by setting λ=1/β) often yield second-order inefficiencies due to nonlinear interactions between discounting and utility. These results challenge the common assumption that bias awareness necessarily enhances welfare. We discuss implications for behavioral interventions and policy design, emphasizing the need for calibrated support rather than rigid rationalization. Our findings highlight that self-awareness is necessary but not sufficient for improved intertemporal decision-making.