Optical imaging through scattering media is a long pursued yet unresolved problem. Researchers propose and develop kinds of methods and techniques, from the earliest time gating and spatial filtering techniques, which utilize only ballistic photons, to subsequent wavefront shaping, scattering matrix measurement and speckle autocorrelation imaging by taking advantage of scattered photons, to the popular deep learning methods latterly. Although all methods and techniques are verified by proof-of-concept experiments with thin scattering media, such as diffusers, zinc oxide layers and slices of biological tissue, they all fail rapidly as the thickness increases. The longstanding thickness challenge is still a bottleneck. In the commentary, I summarized and compared the methods and techniques in the field, reviewed the mainstream viewpoints, i.e., wavefront is completely scrambled after scattering media, analyzed the reason for the failure of existing methods and techniques through thick scattering media, and discussed the potential research directions to solve the problem ultimately in future.