I think its favourable for those who has high level, large budget & classified research project such as Boston Dynamics or maybe like some laboratory in Japan, as my Professor laboratory...but still not in other area such as agricultural, construction etc. like in the late 1990's
Legged/Multi-legged robot and wheeled robot are fit for different environment walking. Legged/Multi-legged robot is fit for uneven or complex environment walking. Wheeled robot is fit for even environment walking relatively. Compared with wheeled robot walking, there are more difficulties to overcome for legged/Multi-legged robot walking, such as contradiction between good walking stability and high walking speed.
I think that the current level of advancement of artificial neural networks is essential as the robot is moving, but if he can make autonomous decisions. Way to get around is secondary, much easier to implement in relation to cognitive mechanisms.