The galaxies of the local group won't recede from us. Telescopes as recently as the early 20th century could not resolve the sky well enough to determine which

The galaxies of the local group won't recede from us. Telescopes as recently as the early 20th century could not resolve the sky well enough to determine which nebulae are in the Milky Way, are satellites of the Milky Way, are in the local group, or are outside it. This lack of observation allowed the Shapeley-Curtis https://en.wikipedia.org/wiki/Great_Debate_(astronomy) https://en.wikipedia.org/wiki/Great_Debate_(astronomy) to last as long as it did -- it ended with clear observations of light curves within M31 and other galaxies in the local group. Those observations will remain available for a lonnnnnnnnnnng time after there are no other galaxy groups in the view of anything in the local group. Moreover, the relic fields (cosmic microwave background and cosmic neutrino background) will be detectable (at least in principle) as they grow ever colder in the far future, so at least the radiation and dark energy sectors can be rediscovered in the distant future when there are no highly redshifted galaxies left to see. We are lucky to have clear views of redshifted, dimmed, smaller-on-the-sky spiral galaxies at all sorts of orientations to us (including face on). Those certainly help us to understand galactic evolution, and also provide us lots of interesting objects (gas and dust clouds, active nuclei, stars, supernovae) "in the past". In the far future, the local group will have fewer spirals, fewer stars, almost no low-metallicity stars, and so on, and recovering those from observation of what's left in the distant future will be more difficult. But around the garganguan black holes left in the local group in the very far future there will still be plenty of dust and gas and cooling stellar remnants; the occasional star will probably form, but not often. Still, the volume will be large enough far into the future that "not often" simply invites wider searches of the sky. Supernovae don't happen often, but thanks to all-sky surveys like http://www.astronomy.ohio-state.edu/~assassin/index.shtml http://www.astronomy.ohio-state.edu/~assassin/index.shtml we see A LOT more of them than could have been hoped for a hundred years ago. Far future astronomers might have a super-hyper-ultra ASAS-SN type programme to look for the earliest signs of star formation in what's left of the local group. Of course as you note, even better would be if the far future astronomers can recover the records of earlier astronomers, rather than having to rediscover everything from scratch using the "records" provided by nature at astronomical scales. The far far future where one cannot see distant galaxies from the local group is very very very very very much earlier than the epoch of fluctuations into Boltzmann brains, vacuum decay, etc. etc., none of which is a really reliable prediction of well-tested theory. (In fact, they are diagnostic of theoretical deficiencies: how do we prevent objects that can happen according to a theory from happening according to that theory, given that we do not see those objects? This recurs a lot at the edges of theoretical physics: in General Relativity it's the basis for developing the https://en.wikipedia.org/wiki/Energy_condition https://en.wikipedia.org/wiki/Energy_condition (s) for instance).