As we've answered earlier today the description of what happens when 90% of the hashpower is lost is approximately accurate, although somewhat exaggerated:
"A 90% loss will create 2 hour block intervals and take a year to resolve."
Actually, it would be 100 minute blocks and would take less than half a year even assuming that no new hashrate were added and the event were directly after a difficulty reset.
Now, luckily, no single government can intervene with 90% of the hashpower. I've seen estimates that about 60% of the hashpower is in China, which if taken out at once would be a major blow, but would only delay the network's rebalance for three weeks (five instead of two weeks for the difficulty period).
Besides, if really 90% of the hashpower were taken out at once with no chance of recovery, a lot would speak for an emergency hardfork to a new hashing algorithm, both to resolve the difficulty hike as well as evading a potential majority attack from the government(s) that took control of the 90% hashpower. While the article alludes that
"forking is impossible without an effective voting / governance system for Bitcoin"
It seems to me that in the face of such an external attack a fork with the single purpose of changing the hashing algorithm would be achievable.
Finally, the article states:
"Only Delegated Proof of Stake (DPOS) has the ability to remain independent of direct government control due to the much lower resource requirements for operating a node."
Which feels like an oversimplification to me. PoS is widely considered as being less secure than PoW among experts, but even so one can say certainly that it makes different security trade-offs. E.g. it is susceptible to one or a few parties holding a majority of the system's tokens, and since POS deployment often is combined with a large pre-mine, it is not hard to be concerned about the initial distribution. One must also consider that the author Daniel "dantheman" Larimer is the founder of Steemit, so it would make sense to get another opinion on the trade-offs between PoW and DPOS…