i don't know if it was rose coloured to attach relevant context...

i don't know if it was rose coloured to attach relevant context to the quote... I certainly don't have the "jaded glasses" glasses, was really just trying to be objective and do a little fact checking there.

The focus on PL5's lack of announced metrics is understandable, especially given the excitement around PL4's results.

Just got to keep reminding ourselves of the value of PL5 on what is a long development path. While we didn't get to read about the impressive metrics we saw with PL4, PL5 served as a crucial experimental stepping stone. The most significant aspect of PL5 isn't its individual metrics, but the knowledge it contributes to manufacturing a working 20nm platform - this aspect was covered with some detail in the PL5 annoucment. For shareholders it might be harder to interpret the relevance of those statements than by reading key metrics. But as PH explained in that interview, this type of iterative development, involving "variations or experiments" to inform the next stage of development, is standard practice across memory manufacturers "whether you're 4DS or TSMC".

We should also keep in mind that in this type of innovative engineering, not all iterations yield immediate breakthroughs. Failure and experimentation are inherent to the process. We should anticipate that PL6, like PL4 and PL5, will involve multiple design variations as the team aims to create a functional 20nm chip that can be properly characterized and provide those sought-after metrics. Success for PL6 could simply mean achieving one working 20nm variation.

Looking ahead, the critical question is whether there will be market demand for a working 20nm memory solution. Interestingly, a paper from Microsoft Research Lab - Cambridge on "Managed Retention Memory" outlines the need for a technology to solve bottlenecks for AI... If 4DS is successful, they would have answers this need perfectly. The existence of this paper suggests a potential strong interest from major consumers in this type of advancement.

https://www.microsoft.com/en-us/research/publication/managed-retention-memory-a-new-class-of-memory-for-the-ai-era/