I said what I said: Algorithmic discoverability is unproven as a way to actually build audience and there is no proven symmetry between ease of consumption and benefit to the producer. The costs for "free" discoverability are well known and well understood so it's not reasonable to argue that any benefit, no matter how small, is automatically worth it.
Youtube is a great example.on both hands. How often do you see videos where the creator says something like "only 10% of you are subscribed! Hit the notification button!" Popular creators tend not to get popular by algorithmic discovery but by getting boosted by existing, popular creators. You can look at viral videos with millions and millions of views and the account's other videos will have practically nothing in many cases.
On the other hand, attacks in creators are well documented, and it is also demonstrably true that algorithmic boosting contributes; how often do you see YouTubers, Tiktokers, IG posters etc talk about their videos ending up on the "wrong side or the site"?
Based on all this, it is in no way either unreasonable or extreme in any sense to ask the question, "do these mechanisms contribute to building an audience in a sufficiently valuable way to justify their drawbacks?"