I'm struggling to pinpoint where you two are talking past each other. "Length" in the context of a blockchain refers quite specifically to the amount of actual work. So "easy" to make "long" chain, seems like a straightforward contradiction.
The amount of calculation necessary to add a block to the chain is not undoable by someone with modest resources.
What's difficult is doing it as quickly as Bitcoin miners with specialized hardware and big energy footprints.
You can make a new blockchain add got Bitcoin today on EC2 without overwhelming expense. Even 30 of them. No problem! It'd take too long to be practical for mining, but that's not the senario we are discussing
Would you beat other miners in? Absolutely not. But the scenario being discussed here is the historical value of blockchains once there is no mining pool racing on them.
And the answer both the article and I propose is: "It is almost none" compared to the other cryptographic tokens. Especially if the chain isn't receiving constant, small commits.
The competition works by increasing the difficulty. At any given time there's a maximum numerical value of the block hash; the lower it is, the more hashes you have to make before you find a valid one.
The difficulty can be checked after the fact, by simply checking the numerical values of hashes. If a hash is one in a trillion, you know it took about a trillion attempts on average to make it, and you can figure out the economic cost of doing that.
It would be easy to make a chain with lots of blocks of low difficulty. It would also be easy for anyone to see that it didn't cost much to make it.