Yes, but think about how slow development would be if they weren't so fast. And even this way they are still very slow.

Even if less and less, there are still dealbreaker years-old encryption issues, like that if someone joins an encrypted room, they won't get the keys for the old messages even if history visibility is on the most permissive setting.
Actually this is for security reasons, and as I understand encryption would be worthless the other way, as a server admin could add an account to the room which would get all the keys otherwise. A workaround to this (it was standardised recently) is that if you invite them from an element client (their client also needs support for this I think, as they have to use it at least for the initial join), your client will send them all the keys it knows, but I'm not sure if it will make only your past messages visible or all of those that you can see (at the point of invitation, so maybe it's best to scroll back to the beginning of what you see for it to work best). They are also working on a real solution (they refer to it as the end of the linked proposal), I have seen work on that even this year, but they are working on so much things at the same time that it feels to me they're getting nowhere.

I often feel that even though alt impls can't keep up, the organization is still starved for human resources. They need to do a whole lot of things at the same time, and of course that's not possible, because everyone can work on only so few things at a time.
They were often even bashed for serious mistakes and overlookings resulting from this (like the hackea writup), but it seems entirely plausible to me that in the past because of this lack of resources they just weren't able to do things properly, and they have problems with admitting that with words even to themselves.