Think of heat treatment of steel (and cryo-treatment) something like baking candy with two materials that have a different melting temperature. You heat the material above the melting temperature of both materials (to get them mixed well) and then you cool the mix down.
If the mix is cooled slowly: the material with a higher melting temp begins to crystallize first, leaving chunks of the material in the final product. The size of the chunks is dependent upon the time the mix stays between the range where only one material precipitates and when the mix falls below the range where both materials crystallize (the dwell time between when the single material solidifies and the base remains liquid).
Cool the mix fast, and both materials crystallize at almost the same time, leaving the smallest chunks.
What drives the speed of cooling? The temperature difference between the molten material when fully mixed (the high temp well), and that of the cooling medium (the low temp well). The greater difference between the two wells, the faster the cooling effect, and the smallest chunks resulting in the cooled mix.
How does this relate to steel? The alloys in steel (tungsten, nickel, chromium, magnesium, etc.) work like the chunks in the candy mix: the faster you can cool the mix, the smaller the grain structure of each piece of alloy material in the total product.
What is so good about smaller chunks in the mix? The different alloy's molecular shapes create a more rigid structure, as they prevent the steel's cubic grain structures from zipper strain failures, shear plane failures, and constrained stress failures as the odd sized grains rub against each other (as the steel stretches before failure). The microscopic cracks split a few steel grains before smacking into a rigid alloy grain, that stops the failure crack. It's much like when a frozen chip stops your spoon digging into rocky road ice cream. The problem reveals itself when a large chip lets go, it releases fast (the spoon flies).
Keep the alloy chunks small and you keep the grain structure failures small, resulting in a more durable material without excessive brittle properties. You can design the total material mix for a harder result, with less fear of large grain structure failure.
This controlled cooling action is called quench hardening. Traditional quench hardening used room temperature as the lowest (low) temperature well to drive the speed of the mixture cooling (the water or oil bucket in the corner of the blacksmiths workshop). Cryogenic treating simply lowers the low temperature well a few hundred degrees below room temperature, to accelerate the speed the total mix is cooled. The result = smaller grains in the final material.
Blacksmiths learned that quenching red hot steel in water improved it's ability to hold an edge, with the risk of getting the steel too brittle (great edge, but the sword snapped before the battle ended, and the beta-tester died). They also learned that if you reheated the brittle steel (go collect the swords after fire raged through the battlefield) the edge still held, and the swords were more durable (not so brittle anymore, battle hardened swords = good).
Reheating the steel, below the melting point of the majority of the alloy, allows some of the softer alloy to flow and fill in stress cracks formed when the steel grains were rapidly cooled (in the quench hardening). The result is a material with more resistance to deformation (less internal cracks to displace, because they have now been filled-in like mortar in a block wall), and because the former cracks now have a softer dampening material between the rigid chunks.
This process of reheating the material is called tempering. The magic in tempering is predicting how hot and how long to elevate the quench hardened mix, to gain the desired result. Different alloy mixes need more, or less, dwell time at different temperatures to gain the desired result.
The Blacksmith used to read the polished steel color change to judge the reheat temp, and then place the piece a standard distance away from the forge to allow a controlled (tempered) cooling rate. The closer to the forge, the slower the cooling rates. You can imagine Apprentice Blacksmiths were busy much of the time, moving tempered steel pieces around the forge.
Room temperature quenching and tempering resulted in some great steel alloy mixes: traditional chrome-moly, tungsten-carbide, etc. (after a few thousand years of trial and error, and a few decades of applied science). Sub-zero quenching and tempering was to expensive (not too much liquid CO2 or N2 around for Blacksmiths to experiment with in 1540), and except for rumors of superior Viking steel quality not much motive to experiment (you want me to leave this toasty forge, to dip the sword into a frozen lake ... I'll dip something into the lake, but it gonna be a roasting sword).
Cryo-quenching is still so new that exploiting new mixtures to gain a better result is still on going. Using the accelerated cooling process on the best traditional mixtures has (for the most part) provided good results (small grain structures with good tempering properties), and there is much more to be learned.
Cryo treatment has the tempering rate of cooling to explore (after doping steel with low temp alloy like silicon and arsenic) as well as allowing a combination of alloy (alloy with wildly different melting temps) for the quench hardening process. There is a lot of room to play, with the different quench rates provided by different cryo mediums (liquid CO2, N2, and even Helium).
Replacing the needle bearings with bushing seems to be of a great value (with or without the cryo-treatment).