JD-EE
Back in the days of the first Jets control runs became more complex and at higher speeds their weight became too much for comfortable flying so servos were introduced. The problem with servos was that they did not always feed the control surface pressure back to the pilot so an artificial "feel" was added to control columns. IIRC it was called "Q Feel" and was considered rather revolutionary at the time.
Q Feel was necessary so pilots could judge, ( based on the old direct systems) how much their control surfaces were deflecting or pressing on the air and secondly, as the control pressure became significantly greater, would not deflect a control surface beyond what was safe.
Then true FBW came along. The control inputs are fed into a computer which judges the amount of control surface deflection required for the indicated manoeuver and applies it. The computer also ensures that the control surface is not deflected too much for the conditions so 'Q' is not necessarily required.
Now personally this worries me a bit, but, as someone pointed out on another thread where we were discussing this, the difference is not the risk of electric failure but the difference in risk between electric failure and physical component failure eg. cable or rod bearing etc., in systems which are both complex. There is little evidence so far in statistical terms that FBW is any less safe than the old physical connection controls.