Apart from the obvious (where it explains the benefits of scaling, etc. and where it re-hashes the stated specifications) the article is unfortunately non-sense for the purposes of assessing 4DS, and is basically echoing back some confirmation context to the leading questions.
AI should be asked open questions as opposed to leading questions which will just confirm
1) 4DS hasn't present to an industry conference nor published any papers in the past decade - which as the AI points out is one of the main ways for industry peers to know about a company and their tech. This is something I have repeatedly called 4DS out on. Peter went to MemCon 2024, rubbed shoulders... didn't present though and thus far nothing has come from it.
Perhaps 4DS may like attend and present their revolution discovery to the 16th IEEE International Memory Workshop, its on May 12th-15th 2024 in Seoul, South Korea. Guessing we missed the deadline though..
2) As Moose pointed out in his AI chat, "
The level of independent verification for 4DS's claims will significantly impact their value. Extensive testing and validation by reputable institutions would command a higher premium.", again, something that I have called out that 4DS has failed to do. Approach some tech universities and let them go to town on the memory chips to test, validate, experiment, extrapolate data. Failing to do this gives the impression 4DS has something to hide.... or is trapped behind a NDA preventing the company from being meaningful.
3) Most of those companies listed would have zero interest in memory cells. For the IP to be useful to anyone, it has to be acquired by a company that can design and package the memory cells. Only NVidia and Intel have that capability of those companies mentioned - both companies are already exploring alternative options in the AI space.
4) It references the acquisition of Activision Blizzard by Microsoft as supporting a high valuation - what does the acquisition of a game development company have to do with the valuation of a memory semiconductor IP? Nothing.
5) It states NVidia acquired ARM. It didn't - it failed. Instead it has a ~$150 million stake in the company, not the 40 billion it states.
(20min delay)