If you're doing something like selling your startup, you want to set the anchor price, yes. There is such a huge range of outcomes, it's best for you to shoot high (as long as you're comfortable with the risk of scaring the buyer off) as the seller. (Or to start and shoot low as the buyer; with the walkaway risk reversed.)

For something like salary negotiation where results are more well-bounded and there is a lot more commoditization (ie replacements are relatively easy to find, or are perceived to be relatively easy to find), you generally don't want to anchor. An anchor that's way out of whack is easy to shoot down (show comp data). Anchoring slightly high/low doesn't buy you much as there is no real anchor... you're just in the established/known range... and so it's more valuable to see where the counterparty starts off an dplay from there.

There are unique situations / exceptions of course. For example, you might have uniquely valuable skills for a given company (or be able to convince them you do) and therefore move into a less-bounded zone of negotiation, in which making the ask/establishing the anchor would again be a good idea. But for a normal, well-established position, there is literally no benefit in "anchoring" as the bounded range is already fairly well established and that is the anchor.