The most basic reason why Bitcoin cannot and will not persist long term is that it requires large processing power as well as storage capacity but offers a tiny number of transactions per second at 7. The exponential nature of technology doesn't concern me when it comes to the processing power and storage requirements. What concerns me is how little it delivers for so much consumption. Meanwhile NXT which is an early 2.0 is over there featuring 100tps right now and eyeing 1000tps. SPV nodes don't really offset this either.
I wrote a whitepaper back in May describing a simple solution to this problem that also has the advantage of providing instant acceptance of transactions. One nomadic mint agent creates the new blocks on a non-branching blockchain. Peers verify the result and copy new blocks to their own copies of the canonical blockchain. A single writer to the immutable blockchain should be very fast. Full nodes share the block rewards without proof-of-work mining effort. Fewer than one hundred lines of code are modified in Bitcoin Core, provided that suitable software agents operate the distributed network.
All core devs I speak to are skeptical, but some encourage my project anyway to see if there is a better way than Satoshi's proof-of-work.
Sounds interesting. Could I have a link to the whitepaper?