We're pretty rapidly nearing the end of transistor scaling with currently widely used technologies and materials. Something's got to give, from a device manufa

We're pretty rapidly nearing the end of transistor scaling with currently widely used technologies and materials. Something's got to give, from a device manufacturing point the thing _consumers_ need the most is probably cost. Keep the feature size, slash the cost so they can be sold more transistors. Size doesn't even really matter that much anymore, but power consumption for work absolutely does. Possibly better interconnect technology and integration into larger physical packages (smaller overall footprint since all the big wires got taken out). Who knows if someone wins a lottery of scientific discovery for either a more cost effective way to do stuff we already know how to do other ways, or for something completely new that breaks open space for new ideas.