alt Hacker Newshave worked in analytics for a while. Engagement is a very broad category. Not all engagement metrics are toxic. For example, daily active users is healthy for a business to measure and is engagement. Commenting on Hackernews is engagement. Meta also has a wide range of engagement metrics - not all of which are problematic
I've never seen a NAXALT fallacy used for the concept of engagement metrics before.
It's not relevant to this discussion that in certain contexts, useful engagement metrics exist that supposedly aren't toxic, because it doesnt apply to the current context. We're clearly talking about the larger systemic issue of companies maximizing all engagement at all costs, and the consequences of such a policy.