Thanks for raising these points. I’ve been discussing these issues recently in private email, especially in debate with precious metal bugs such as 
but on the ontological level it lags the thermo-dynamically hard-properties of a piece of gold.
The tangible mass of proof-of-work, which is all the mining hardware, is detached from the transferable information value. But I fail to see how the tangible mass is any less tangible than for gold.
Resources are burned to mine both gold and proof-of-work generated tokens. Idealized mining that is 100% efficient would burn no electricity and only burn the depreciation of the mining hardware due to Moore’s law.
Bitcoins correctness is only guarantied by constant validation of the distributed state. It is not quiescent. Even when there are no/not enough transactions, miners still need to validate. Ergo Bitcoins grid-storage is RAM and not ROM. Coins should not be considered as permanent or as Szabo calls it "secure property titles".
Ditto gold. Nothing is permanent. Would violate the inexorable trend to universal maximum entropy.
Gold can’t be used as currency anymore because it requires physical exchange but most of our transactions these days are virtual. And exchange requires market makers. Physical market makers can be rubberhosed by the authorities. Decentralized, virtualized market makers can in theory play Whac-A-Mole with Uncle Sam.
Every extrapolation-based model implies, that the scaleability-trilemma is solved ON-chain... which is a strong assumption.
Bitcoin is not intended to scale. It will be a 666 global reserve currency to enslave us.
Their existence is directly dependent on the miners profitability and the miners profitability as far as I understood is dependent on >>ON<<-chain scaling.
The price rises and mining hashrate can adjust to the profitability of the block reward at any price. The only requirements is that hashrate be higher than anyone in the world can 50+% attack. Also there’s another theoretical problem if the transaction fees become worth more than the minted block reward, but a rather small oligarchy of miners has an incentive not to cheat each other to create a death spiral. They will know which of them has defected.
In fact there are no copies of the blockchain or the coin, because redundancy is bottom up and never top down. A part of the system (the single process/node) cant hold the emergent property of the replicated state machine. So coins only sit in the grid. Mutualy shared information does not make a copy.
Agreed there is only one truth, which is very low entropy and highly ordered. And thus totalitarian and not resilient nor antifragile. Just like fiat currencies (they also don’t fork).
The one-chain-to-rule-them-all of proof-of-work as designed by Satoshi also has a poison pill game theory economics to kill all forks. I covered that in the Long-term advance notice and Rogue Wave threads at bitcointalk.org.
Note I think I have a design for a decentralized solution to this issue. I don’t see Bitcoin as the solution for mankind, only for blue-pilled zombie-kind. We red-pilled, awake, sapient humans will need a more resilient and relativistic form of proof-of-work.
Second point is the mathematical rigor: Markets are fat tail distributed right? Hence convergence to the real mean is not observable in real time. Why you bother with those "predictions" at all? Even Nicks co-integration to PlanB´s prediction is based on a violation of the law of large numbers... I mean your reconstruction is rigorous because its not a prediction and has no strong assumptions but the object of your study, the prediction of MCAfee is just bogus?
We all die but while we are alive we should not ignore the local orders around us, unless we want to go off into random loony state such as John Nash did for a while.
PlanB’s model probably fails when the opportunity cost to expand mining is less than the interest being paid in on global, sovereign Bitcoin denominated bonds in the post-dollar, reserve after 2028 or 2032.
Bitcoin can’t be worth 100% of all financial assets in the world. That would require the world stop. Even PlanB acknowledges that. He wonders whether his model will cease when the S/F ratio is the same as for gold. I don’t think so. I think Bitcoin’s marketcap will worth an order-of-magnitude more than gold.
P.S. Thank you to 
RE: McAfee’s Dick Math: illuminating Bitcoin’s ACCELERATING price