@jackm If it is possible, people will make it. We have no indication yet that there are any other better architectures or materials. Even if better architectures could be implemented it doesn’t mean that people will be willing to implement them. Multi-core processing is hard enough to get people to program for, imagine trying to get everyone to completely relearn programming and computation for a completely different architecture. By no means will this be profitable any time soon.
We still use gasoline in our cars because everything is built around the architecture of the internal combustion engine. We’ve known for a very long time that we are nearing the end of gasoline’s useful life in road vehicles, yet very few companies are making a big step towards renewable resource based vehicles. It’s not that the technology is not there, it’s that the leap to the next architecture is not profitable.
The exact same applies to computing. There may be technology that can surpass what we are achieving currently but it would require a complete change in the structure of the computing world. There will be a plateau, just as there is an internal combustion engine efficiency plateau now. This plateau will persist until the benefits of more powerful computation outweigh the consequences of reworking the entire structure of computation. Since a great deal of everyone’s daily lives is dependent on computation, I’d have to say that it will most likely be quite some time before this happens.