The first mirror is partially transmitting to allow light into the interferometer. The second mirror, in our case, will be the object whose position we want to detect.
In the best case scenario (corresponding to the maximum <ahref=https://en.wikipedia.org/wiki/Interferometric_visibility>visibility</a> in which the fixed and moving mirror have the same reflectance), the total reflectance $R$ of the cavity as a function of the distance to the object will have the form of a sinusoidal oscillation with a maximum reflectance $R_{max} = 0.5$ and a minimum $R_{min} = 0$:
To detect changes in the position of the object, one shines a laser into the cavity and detects the power of the laser light that is reflected. By adjusting the first mirror to the correct position so that the interferometer is on the "slope of the fringe", as illustrated above, any change in the cavity length will result in a change of the reflected light due to the slope $dR/dx$. In this way, tiny changes of the position of the object can be translated to large changes in the laser power detected by the photodiode.