So it's really matured in the same direction only further.

The most effective & consistent mass-production processes do not always produce components as identical as would be ideal, and not all equal in ultimate performance capability, especially in the most demanding applications.

This can be expected to be more problematic when production is first initiated, whether there are bugs to be worked out, or optimizations have yet to be accomplished.

With resistors the percent deviation from the target ohm value is a simple tolerance rating.

Not everyone needs resistors within 1 percent of their labeled value, but those that do can not settle for anything less.

When initial production results in a maximum deviation of 20%, during QC/QA each component (or batch) can be measured and binned into the 1%, 5%, 10% and 20% tolerance ratings.

And most importantly priced accordingly.

Interestingly, if the binning is most comprehensive, then the consumer of the 20% parts never gets a component closer than 10% of the nominal value.

Vacuum tubes are not quite so simple as resistors, but it was well-established that the military would gladly overpay for the very top-performing tubes meeting the tightest tolerances.

In earlier decades military tubes were often physically enhanced and produced separately from the lesser consumer versions.

By the 1950's things were very mature in this respect.

And it can be a lot less costly to manufacture everything the same.

So binning it was, to select the closest conformers, which were then labeled with military part numbers destined for a supply chain where price is not as significant as it is for consumers.

The remaining tubes then were marked with everyday consumer part numbers.

This really became prominent for dual-use items like radio tubes.

Once the manufacturing process was fully optimized and every tube met the most stringent specifications, the cost to make each tube can decline dramatically.

At the same time, freshly made tubes were no longer individually labeled, they then all went into one bin.

So the most overspending customers could be supplied the same parts at higher prices than everyone else from that point without anyone knowing the difference.

Unless the customer really checked the performance independently in great detail.

Depending on the customer ordering the parts, they would be labeled accordingly right before shipping, sometimes with a little more QA and/or written guarantee but not any additional costly manufacturing QC which might have been necessary at the begining of production. They could then better meet fluctuating demand between premium payers versus consumer usage from the same manufacturing line and a single ultimate bin. And with a different part number to inhibit direct substitution even after identical performance could be recognized in some cases.

So this was well underway before solid-state semiconductors became a commodity.

As the processes improve it ended up with price-sensitive consumers who figured out when the lesser-priced CPU's had become as high-performance as the premium-labeled "alternative".

With the complexity of modern CPU's it was basically trivial to circumvent this consumer effort with the last-minute fusing & labeling approach seen today.

We've come such a long way it's not like the 1950's at all.

Lots fewer consumers expect to get their money's worth, and some vendors are only looking for these type customers any more.

Eventually all they make is i7's and if you don't want to pay the price, they further customize the component to be especially crummy, just for you.