It should be noted that the images of the shader nodes have already been updated. rBM8426: Update images to reflect changes in subsurface scattering

Not sure how best to describe this.
The "new" method is mostly a new albedo inversion technique. Essentially, if one provides a detailed Subsurface Color, the Random Walk will aim for that color at every shading point. This means that, contrary to previous approaches, the Subsurface Color gets very little blurring from the SSS simulation, "only" the illumination on it will get smeared out according to the Diffuse Mean Free Path (called Subsurface Radius in blender lingo). You will still get backlighting and the general "softness" one expects from SSS, but you do not need to pre-sharpen a detailed albedo map to compensate for the extra blurring it would do on it. It just is more accurate, truer to the indented painted or acquired albedo map (Subsurface Color). This is really good for artists, as they can migrate without much effort, a lookdev using a diffuse brdf into SSS, with the same inputs.
Additionally, the albedo inversion takes into account the IOR and the phase anisotropy of the medium, which are important for human skin. And obviously these additional properties are also reflected in the sampling strategies used during the actual random walk.

Almost there.
Could we change "adjusts" into "modulates", ie Behaves similarly to *Random Walk (Fixed Radius)* but modulates the *Radius* based on ...
(modulates implies more what happens internally: we do not really adjust the radius per say; we adjust the reduced albedo to match both constraints, Color and Radius)
And last sentence, change "also" into "thereby", ie This method thereby attempts to retain greater ...

Thank you @Christophe Hery (chery) for the help you have provided.