I think many ancient cultures seem empty because they wrote on things like paper that doesnt last that long. We know a fair bit about sumeria and mesopotamia thanks to cuneiform on tablets that can lie in a ground for millenia.
And since we use electrons and stuff like that we too, will seem an empty age.
Even if no Rosetta Project discs survive, silicon chips can easily survive many millennia even if they are only encased in epoxy rather than ceramic. Data in Flash chips will leak away after only decades to centuries, but the Flash itself will survive. Silicon is an extremely stable substance; although as with its companion aluminum, it does oxidize in air, forming a very hard, adherent, poorly permeable, highly chemical-resistant oxide layer, which protects the rest of the IC.
For diffusion in silicon (as opposed to silicon dioxide) the empirical diffusion coefficient of oxygen is given as 0.13 exp(-2.53 eV/kT)cm²/s by Binns, Londos, et al., 1996. This works out to 3.9 × 10⁻⁴⁵ cm²/s (3.9 × 10⁻⁴⁹ m²/s in SI units) at 293 K (20°) and 8.5 × 10⁻³⁶ cm²/s at 373 K (100°). If I'm calculating this right, the room-temperature number is some 10³² times slower than at the usual silicon oxide layer growth temperatures used in chip manufacturing, which can grow submicron-thickness oxide layers in a time on the order of an hour https://www.iue.tuwien.ac.at/phd/filipovic/node29.htmlhttps://cnx.org/contents/lJGecnVz@89.1:lYPudNFT@2/Chapter-6-... although sometimes using steam rather than just air. So, if the diffusion coefficient through amorphous SiO₂ is similar, the chip surface will rust away by a similar amount at room temperature in on the order of 10³² hours, which is 11 octillion years (1.1 × 10²⁸ years), long after star formation ceases and the last red dwarf goes out. A much more urgent consideration than silicon rusting is that, unless urgent conservation measures are taken soon to prevent it, the Sun will engulf and vaporize the Earth in under 8 billion years.
(It's possible there might be other weathering effects that become dominant at lower temperatures, for example, diffusion of corrosive species with lower activation energies than the 2.53 eV measured above; it seems plausible that even 10³² hours of cosmic rays might be enough to blast the chips apart into nothing. We'll know for sure in a few octillion years.)
(Laura Nuccio's dissertation http://www1.unipa.it/lamp/NuccioPhDThesis.pdf gives 2.9 × 10⁻⁴ exp(-1.17 eV/kT)cm²/s on p. 30, which works out to 2.2 × 10⁻²⁴ cm²/s (2.2 × 10⁻²⁸ m²/s) at 293 K (20°). This gives a time of only 10¹⁰ hours, only about a million years, for a similar amount of oxide formation to take place at room temperature to the oxide layers commonly used as insulators in chips. So perhaps the aluminum traces on the surface of silicon will be converted to insulating aluminum oxide within the next several millennia.)
None of this means that the chips will work, though; much tinier amounts of dopants diffusing much shorter distances through the chip are sufficient for the chip to stop working. But it seems likely that post-human archaeologists millions of years in the future could reverse-engineer a 6502. Perhaps a future Chris of Clickspring will document his meticulous reconstruction of the "Commodore Mechanism" on the Centaurian equivalent of YouTube.
Aside from silicon, though, the humans' civilization has produced many million-year-lifetime objects like the screen-printed ceramic floor tiles I'm standing on (which have text stamped into their back), bricks, pottery, and glass, which bear witness to a complex material culture with widespread literacy. Somewhat less long-lived, but still capable of lasting millennia, are the numerous objects made of brass, bronze, aluminum, and thick stable plastics. (Polyethylene, polypropylene, and PET will last millennia, no problem, unless they're only 10 microns thick like this shopping bag. Acetate will eat itself, and the paper around it, in decades through vinegar syndrome; similarly for most elastomers, although silicone is pretty stable. Celluloid dissolves into brown goo when it doesn't spontaneously detonate. Things like polystyrene and ABS are of intermediate stability. I have no idea about nylon, epoxy, and phenolic.)
Also, though, people do lithography today too, including to reproduce writing. The best stones tend to be palimpsests, ground down by the artists to expose a fresh flat surface to print new art with, but many lousier stones are buried in sealed, dry 20th-century municipal middens. Intaglio printing has often used gypsum plaster, as has metal casting. These pieces could easily survive millennia if kept dry. So too with typemetal plates used for book printing.
Steel, of course, the 20th century's favorite drug, rusts away to slag in a few short decades; but stainless steel, like the spoon I am eating with, forms a surface oxide film similar to that of aluminum, but made of chromium oxide rather than amorphous sapphire. Like silicon and aluminum, it too should last for millennia or longer.
It should be easy to do much better than this with a little effort.
In conclusion, if the current global civilization collapses, enormous volumes of its writings will survive for thousands to millions of years in the form of durable manufactured objects that contain writing more or less incidentally.
And since we use electrons and stuff like that we too, will seem an empty age.