Unlike other projects I have on here, Seamus is not yet available to use. This post is more of a pre-mortem and problem statement.
Seamus is a curator of links. Almost every person I know curates lists of something, or at the very least, wants to curate them. Lists are a way of manufacturing identity, of telling the world - "hey! this is what interests me". Unfortunately, list curation is no easy job. Apart from dedicated apps for things like books (Goodreads) or in-app curation mechanisms (bookmarks), there is no easy way to collect all your links about a single subject, easily. The typical flow for most people I talk to is:
- Create a note in Apple notes
- Be extremely intentional about your content consumption, and when you see something that fits in a list dump it there
- Go back and clean up the list periodically
Some people replace list specific notes with a journal dump, but this only makes the problem worse. You no longer have a way to curate your content, no way to go back and find things that were interesting to you, no way to show the world what you like.
Seamus solves this problem by giving you a dead simple way to dump links - no work necessary. Seamus will then hydrate the link with useful information like a title, description and an image. It'll then place it into an appropriate list (one, many, or none), where you can add additional notes and thoughts about the content you linked.
Crucially, Seamus will never "suggest" content to you. It does not seek to think for you, it only seeks to make it easy for you to think by tracking information you wanted to track.
After this basic integration is done, Seamus should be able to go into your content, and extract list specific metadata for you. For example, you might want to get the full recipe transcript for any food links you paste in. You can add a "field" to your list describing what you want Seamus to do, for example, extracting recipe transcripts. This makes it incredibly powerful for all sorts of workflows and curation.
Another piece I want to setup is publishing your lists. In the past, when the internet was small and you still used dial-up, people wrote these magical things called blogs on the internet. As the walled gardens of big tech have been built, the joy of maintaining a personal, public space that you can share with your friends has dwindled. Seamus seeks to make this easy again - with no walls. Put in any link, from anywhere, and we'll do our best to parse it so you can start thinking about it without doing the busy work of tracking it.
One risk factor I'm acutely aware of for Seamus is that it might lead you to curate too much. Perhaps the act of forgetting is as important as the act of remembering, and tracking lists of everything you see is not a good mental state to be in. Still, two things keep me optimistic:
- We don't track enough today. We don't give ourselves chances to build identity by revisiting interesting instagram posts, or twitter posts later. A chance for us to relive the joy we had when we first saw it, or to say "why the hell did I ever track this?". I think that reflection is powerful to reinforce thoughts that you want to stick with you.
- Seamus has the potential to make media consumption less brain-dead. I've been trying out an experiment where I write a review, even a couple of lines, for almost every piece of media I consume (excluding tweets and short-form video content). This has forced me to engage with media (especially Youtube) better, and made me ignore pieces of content that are fundamentally uninteresting. That's not to say that I track every single thing I watch, but for the specific class of Youtube videos that are entertainment disguised as learning, it forces me to actually learn from them.
Backstory
Seamus started off as (and currently exists as) a personal content management system. It is a way for me to track lists, but also tasks, and many other random pieces of my life in a way that is deeply integrated with an LLM that can categorize and hydrate all these pieces of micro-content I create. While I still believe in this longer term vision of Seamus, the explanation and implementation became incoherent and too hard to understand. This was abundantly clear to me after discussing the idea with a friend, who encouraged me to hone in on a core use case. Together, we came up with curation as the first fundamental problem that Seamus should solve. Eventually I want to build a system that can solve much more complex problems like this:
This system (very roughly) is the combination of an unstructured note taking app with a structured database. I am currently using it, but Seamus as a curator will be a better first step while the vision and UX of this much more complicated app is ironed out.