The fact that some people fail to accomplish a thing doesn't mean others can't. Of course you can engineer addictive games. We know quite a lot about addiction. What do you think casinos are, exactly?
If you look only at examples of successful things and say "it is addictive" but don't look at the same principles as applied to failed examples of the same thing, then you're being intellectually lazy.
In order to establish some facet as "addictive" it needs to work both ways, otherwise you're just using "addictive" as a synonym for "successful" -- Edison's light bulb was "addictive" and people were "addicted" to Pullman cars. But you get to carry all the negative connotations of "addictive"; you're trying to let the informal do all the work while positing nonsensical equivalences.
The rebuttal was: "this is literally what casinos do"
The rebuttal to that was: "well, some casinos fail"
Which isn't really a point. Some casinos fail, that doesn't show that they don't optimize for addiction. It just indicates that optimizing for addiction is necessary but not sufficient for a casino to maintain solvency.
I read it as pointing out that success isn't guaranteed with an addictive game model.
I did not read it as saying that some games don't or cant optimize for addictiveness.