hard not to get carried away here...
...since the implications of reducing to practice this longstanding
memory effort are enticingly large.
i will try to take your advice to be patient re: the intercompany & people arrangements. who knows, perhaps intel's interest is still a factor. the capital is certainly there, which could dissolve much bad blood between micron and the new sentient being about town. conspiratorily, perhaps lowrey's new freedom is a strict requirement for some intel-size combine to solidify interest.
on the tech front, i did not mean to imply that 10**14 cycles is anything to sneeze at. just one huge market, DTV frame buffers, does not need such overkill (display memories only need updates 100/second or so).
ecd's david strand et. al. lays out the progression to trillions of cycles [they stopped the experiment before observing any failure!] pretty well in #5,536,947. intriguingly, the denser the memory the better the performance, so better results go hand-in-hand with lower cost per bit. this is a refreshing change from battery science.
we're a long way from the 1960's where the ovshinsky effect was denounced first as a "thermal" process, with attendant limitations, to a mere electrical curiosity. it must be sweet vindication to see that optimizing active materials for submicron optical processes shows the amazing discovery to be virtually immortal w.r.t. cyclability, no doubt because the once highly-controversial perception that the effect involved gross atomic rearrangement has been fairly shattered. |
