💾 Archived View for gemini.onedimension.net › thought › hyperfov-dec-28-2021.gmi captured on 2022-01-08 at 13:43:10. Gemini links have been rewritten to link to archived content
-=-=-=-=-=-=-
I was doing some thinking today and some things have started to become clearer.
As you browse the internet, you're remembering little tidbits of context from each site you visit, things like what you were reading when you clicked a link, a particular way the page looked, what else was on the page. The problem is that when you later try to retrieve a specific piece of non-remembered information--the site's url--none of that context is available to you to find it. The only way to search the browser history is through the urls themselves, and there's no reason you'd remember them. But you do remember the context.
So the idea is that Hyperfov preserves this context and gives you the tools to retrieve information that would be otherwise irretrievable.
Why would anyone want this?
In general the problem is that there's a lot of information we encounter day-to-day. We develop systems to sort through and save that information: G Docs, Notion, Confluence, Are.na, Excel, notes, sticky notes, etc etc. But inevitably those systems will at some point fail and we'll find that there's a piece of information that we need but cannot find. This is the system that you use; you plug in all the context you can remember around that information and it gets resurfaced.
If we're able to create an accurate record of the experience of browsing, we're able to add an additional layer of metadata to any information you do save and remember.
Here's a stupid example. Maybe you're looking for images of a tuna fish for a presentation you're making. You search and find a couple good examples, deciding to download one and use it in the presentation. Six months later you need to give the presentation over and go back to polish up the slides. That fish picture isn't quite right though, you want a picture with a bit more danger for emphasis of the point you're making. You remember that the site you found the fish image also had a picture of that fish about to get eaten by a shark, or something like that. Perfect. So you search for 'tuna about to be eaten by shark' to no avail--the pictures are all wrong. So you search again for 'tuna fish' in hopes of finding the original site but lo and behold the results have changed. No problem: you reverse image search. There are 1,500 sites that use this same tuna picture. What to do?
You pop the fish picture in Hyperfov and it surfaces all the times you've encountered it on the web. And sure enough there's the site.
Obviously there's a million other stupid and less stupid use cases here. The retrieval use case is fairly clear. I think one interesting challenge is thinking about possible directions that are less passive. How could a system like this be of use while browsing actively?
Last updated Tue Dec 28 2021 in San Anselmo, CA