A reader who saved a long piece about anticipatory urbanism to Pocket last year to read later opened it recently and found nothing — Mozilla shut Pocket down in 2025, and years of saved reading, across millions of accounts, either had to be exported in time or was simply gone. A tool built entirely around the promise of remembering what you meant to read did not, in the end, remember it for very long.
Slow media makes an implicit promise that fast media does not: that a piece is worth returning to, that its ideas compound rather than expire with the news cycle. Most reading products do not honor that promise structurally, and Pocket's shutdown is a reminder that even a product built specifically around that promise can fail to keep it once the business case for maintaining it disappears.
A reader memory worth building would track something closer to what a good editor keeps in their head about a returning reader, and closer to what Readwise already does deliberately for its users: resurfacing a highlight or an idea months later, at the moment it becomes relevant again, rather than just storing it inertly until an account gets deleted. That is a different kind of personalization than a save-for-later queue — it is oriented toward continuity of understanding, not toward storage.
Building this well means resisting the most obvious failure mode, which is a system that narrows what a reader sees based on past behavior until they are reading a smaller and smaller slice of the publication. Continuity should widen a reader's engagement with a subject over time, surfacing genuine follow-ups and connected threads, not simply repeat what already worked once.
It also raises the exact question Pocket's shutdown made concrete for its users: who controls a reader's memory of their own reading, and what happens to it when the company holding it changes strategy or shuts down. A reader memory requires retaining information about what an individual has engaged with, and getting this right means being unusually clear about what is stored, why, and what happens to it if this project ever stops running.
Pocket's answer to that last question, in practice, was an export button and a shutdown notice. The version worth building treats that as a warning rather than an edge case: a reader's memory of what they cared about is worth building, but only if it is designed from the start to survive longer, and more transparently, than the product that stored it.