Context always matters. And estimation is always present in some form, but I think how we approach it matters too.
When you create a new piece of software there is often two ways of doing it. You can start from nothing and gradually add on only what you need. Or you can start by taking something done before (or scaffolded through a tool) and remove the pieces you don’t need. The second is often quicker, and may work perfectly well, but inevitably you end up with something that includes things that aren’t needed, or is over complex for the context. Finding those can be difficult, so generally they stay, and those that come after to look at the code will not be easily able to tell why certain things are there to determine the key pieces.
Estimation can be like that. Anyone who has worked in software for a while recalls the large spreadsheets filled with days and hours, planning out months in advance. We know that isn’t right, but often we start from there are and remove pieces to get to a better process.
What if we didn’t. What if we started from the premise that we didn’t need any estimates, and then added in the pieces we found we really did need as we went along (there are most certainly things we do). We would be much less likely to end up with a process still holding onto extra pieces or complexity we don’t need. We would help everyone shift to a mindset that these really are guesses, as you say.
Breaking our immediate work down into stories can be reasonably accurate. Planning out way in advance cannot (and honestly a lot of the time there is a fixed date regardless of what estimates will say). Prioritisation is hard, but how long something will take (or cost) can often end up being weighed much more heavily than it should in that equation. Ask how much it will cost not to do before how much it will cost to do. Sometimes that is known exactly, sometimes not – but if you’re going to have a guesstimate that is a better one, as it will speak to the immediacy of need to address that item, and guide prioritisation much better.
The only person who can truly understand your context is you, so while it’s good to listen to others, in the end your judgement must be considered more than anyone else’s. Good luck in your efforts to fix what needs to be fixed :)