It seemed the most logical route to have started down. I would not have thought given our current level of technology that moving robotic 'bone like' parts with tendon mimicking material was all that difficult.
Well mimicing an existing organic system does have merit to it, but actually pulling it off is a lot more difficult compared to what can be done with mechanical/electrical components. The article itself says that even the most 'basic' movements, like raising an arm and moving it about, is fairly complex, especially when you compare it to the motors and pistons that already available and have been used fairly extensively by current robots.
A relatively similar analogy would be to why we don't have more walking robots. Sure we could just copy what already exists, and has existed for a long time in nature, but then again nature can't give living organisms wheels for the most part.