jwblue,

I'm assuming when you said the article "isn't optimistic" you were referring to the prospects of shrinking chips to the 14nm node, and not that the plight of POET isn't optimistic.

The fact that you think an article from a year ago is ancient is entirely the point. It isn't really ancient because the problems from a year ago (indeed from 6 years ago) are still present today. All that has changed is the level of urgency to produce a new solution.

What would be the ultimate invention would be a chip that could start at larger feature sizes, as Solesius suggests, leaving lots of room for further scaling. Perhaps transistors on this chip could also eventually share space with optical thyristors which only need to be scaled down to 500nm, allowing them to act as lasers and to also interact with other chips stacked vertically. It would be really neat if these chips were rad hard so they could be used in space without extra shielding and could be immune to the effects of solar flares here on earth. Perhaps they could also use the inherent properties of the thyristor to replace every form of memory and hard disk used today with a single type of universal memory, which could also be stacked on the logic chips.

I know it's no use dreaming of such a perfect solution. Surely a monster of R&D like Intel or IBM would have invented such a thing if it was at all possible. Surely we wouldn't see a 50cent company come along and blow the lid off this problem . . .