have a look at the I-V curves for diodes, especially zeners. they helped me understand - when a Zener is in breakdown, its slope is VERY steep. This means that for a LARGE change in current (vertical), there is a SMALL change in voltage (horizontal), which is why they are good regulators.
Look at the datasheet for a 1N4148 diode. It does not suddenly conduct when its forward voltage is 0.7V. The graph shows that current and temperature affects the forward voltage.
At a current of 1uA the forward voltage is only 0.275V at room temperature.
Why do you say that diodes have a maximum reverse voltage of usually only 50V?
Look at the 1N4148 and the 1N4007.
Not great amplifier mind you! Diode's have a linear region during their pass from non conducting to conducting, you can bias the diode with a fixed voltage and capacitivly couple an external signal into it. Distortion is horrible though, not particularly useful, though I've heard some crystal radio makers use a bias voltage to increase sensitivity.
No, a diode will not conduct until it's forward voltage threshold is crossed. Typically .7v. LEDs are higher (1.0-2.0v) and Schottkys are usually less (~.3v)