Since March 2024, we (Matt, Miki) have been working on Subset, a new way to save, share and search. We're prolific curators ourselves and we've long been dissatisfied with the time, energy and expense required to find and store interesting things, share them to relevant people and destinations, and search amongst found things once the original context has disintegrated. So we've taken up the mission: make saving simple, sharing effortless, and search joyous, and do it in a way that scales.

Here are some of the many established and emerging themes driving this pursuit of a "new way":

The shifting tides of read-later apps, tools for thought and modern collaboration suites
The widespread assumption that the only way to "be online" is to consume or create
The emerging class of machine agents and the novel abstractions they unlock
The growing sophistication of local-first applications and tailored experiences
The proliferation of generative tooling and sloppy information
The approach of a sufficiently decentralised, consumer web
The calls for human-led taste-making and culture
The death of platforms via enshittification
The retreat to the cozy web
The rise of protocols

Uniting all of the above is the realisation that, to borrow a phrase from the Bitcoin maxis, it's still so early. The internet as-is and the web as we experience it today is but a few decades old. There's still so much to play for, in both finite and infinite game terms, and we're determined to bring some of the many possible alternatives to life (and deny some of the more dystopian alternative futures in the process).

Right now, we have a 0.2 product. It's a Google Sheet that generates email digests. Matt uses it to run his Mag7 newsletter; Miki uses it to run, Create Intelligently, his own digest; we both use it to share the interesting things we find related to saving, sharing and search, along with Subset posts like this one. It's a process with three simple steps.

First, you set a digest production mode:

Manual: select things and schedule digests by hand
Automatic: time-based (e.g. send a daily digest with all things marked as "ready")
Automatic: volume-based (e.g. send a digest every time I have five things "ready")
Automatic: time x volume (e.g. send out seven "ready" things every seven days)

Second, you save the interesting things you find via a form.

Third, you relax as digests are auto-generated and sent to contacts and the things you send out are published to your interactive, online repository.

This makes it really easy to produce digests sustainably over time and accrue the upside that comes from deliberate saving, consistent sharing and the accumulation of a deep, highly relevant archive that yourself and others can reference and explore. It unlocks a simple way to nurture relationships, foster spontaneous interactions, and regularly share relevant information from across the web.

If that sounds like something you'd be interested in using and benefitting from then get in contact. We're looking for prolific savers, sharers and searchers in a personal and/or professional context to use the 0.2 product and help us shape its emergence.

Additionally, if the new way to save, share and search we're focused on is something you'd be interested in building then we want to hear from you, too. Especially if you have expertise in:

Local-first mobile applications
Device/OS-level agents
Many-to-many platform / channel / protocol integrations
Search and retrieval involving federated learning, differential privacy and P2P networks
UXs that make high information density and powerful abstractions accessible to end users

As we push forward we'll be exploring the themes above and charting our progress right here, on Paragraph. You can also sign up for our email digest, follow along on our Warpcast channel, catch us on LinkedIn and X, or just send us a good ol' email.

Here's to the new way...