Superoscillating signals are band–limited signals that oscillate in some region faster their largest Fourier component. While such signals have many scientific and technological applications, notably in optical superresolution, their actual use is hampered by the fact that an overwhelming proportion of the energy goes into the part of the signal which is not superoscillating. This phenomenon is quantified using the concept of the superoscillation Yield, defined as the ratio of the energy in the superoscillations to the total energy of the signal. In this talk, the problem of optimization of such signals with respect to the superoscillation yield will be discussed. The basic challenge is to determine whether the belief in the predestination that superoscillatory signals have an 'exponentially low yield' is justified. In particular, we search for fundamental bounds on the energy loss related to the generation of such signals. It turns out that the optimization method we developed can improve the superoscillatory yield by orders of magnitude, and while finding optimal solutions requires high computational precision, the outcome is surprisingly very robust to noise. [1] Yield-optimized superoscillations, E. Katzav and M. Schwartz, IEEE Transactions on Signal Processing 61, 3113 (2013). [2] Yield statistics of interpolated superoscillations, E. Katzav, E. Perlsman and M. Schwartz, J. Phys. A: Math. Theor. 50, 025001 (2017). [3] Optimising superoscillations, E. Katzav and M. Schwartz, in Roadmap on superoscillations edited by M. Berry and N. Zheludev, Accepted to Journal of Optics (2018).