Research note
The future of personal memory
For forty years personal computing has stored things faithfully and understood almost nothing about them. That asymmetry is ending, and it changes the design space more than any interface trend.
Three forces, and one constraint that decides how they land.
01 · Forces
Cheap understanding, capable local hardware, saturated capture.
Reading text out of an image, transcribing audio, summarising a document and computing a semantic representation used to be projects. They're now routine operations, and increasingly they run on the machine in front of you rather than in a datacentre.
At the same time, capture is saturated. There is no meaningful gain left in making saving faster. The remaining value is entirely on the retrieval side.
When understanding becomes cheap, description stops being the user's job.
02 · Shape
The archive becomes queryable in human terms.
The practical consequence is that a personal library can be asked vague questions: what was that dark pricing table from the spring, what did I read around the time of the redesign, what else belongs with this.
That's not a chatbot bolted onto a folder. It's a different data model, in which each saved thing carries its content, its moment, its source, a description and its relationships to other saves.
- Content
- Text recovered from images, documents and audio.
- Moment
- When it arrived, and what arrived beside it.
- Description
- A plain-language account of what the thing appears to be.
- Relationships
- Edges to other saves, with visible evidence for each.
03 · Constraint
Personal memory is the least acceptable thing to upload.
A personal library is an unusually complete record of a life: what you read, worried about, worked on and screenshotted at 2am. The willingness to hand that to a remote service is low and falling.
This is why local-first isn't a marketing position in this category so much as a design requirement. Whatever this software becomes, the version people keep using for a decade will be the one that doesn't need to send anything anywhere.
Next thought
The clearest way to see what's missing is to look at what already failed.
Read: why bookmarks failedIf this idea interested you…
Shelf V3 is out now
Shelf V3 is the current build. Shelf does not use automatic updates, so download the latest build here and install it over your current version.
This is intentional. Shelf is built around a local-first, privacy-focused philosophy. An automatic updater requires additional background processes, network connections, update infrastructure, and another system with permission to modify the app. We prefer to keep Shelf’s network footprint and background activity as small as possible.
Manual updates keep the process explicit: Shelf only checks the internet when a feature actually needs it, and you decide when a new version is downloaded and installed.