IBM likes to play with lab experiments.
I remember when they spelled out “IBM” by moving gold atoms.
Cute and interesting but totally useless from a mass production viewpoint.
Particularly in memory circuits, it needs to work perfect, 100% of the time for decades.
At Intel, we had a circuit which would work fine for about 30 seconds...you would walk away and it would die.
We told the design team to try again and not waste more of our time.
As Zathras would say: “You are finite. Zathras is finite. This... is wrong tool.”