Every time I try to explain some complex idea for how things ought to work to people, I am always reminded that the best way to communicate a new complex idea is a demo they can try themselves.
Hyperlink maximalism
I'm a hyperlink maximalist: everything should be a hyperlink, including everything that is hyperlinked by the author, everything that isn't hyperlinked by the author, and the hyperlinks themselves. Words should be hyperlinked, but so should be every interesting phrase, quote, name, proper noun, paragraph, document, and collection of documents I read.
There are two obvious problems with this idea. (1) No author has time to hyperlink infinite permutations of everything they write, and (2) if everything is hyperlinked, nobody will know what links are useful.
But both of these issues are trivially solved if we simply begin with today's lightly hyperlinked documents, and let the reader's computer generate links on-demand. When I'm reading something and don't understand a particular word or want to know more about a quote, when I select it, my computer should search across everything I've read and some small high-quality subset of the Web to bring me 5-10 links about what I've highlighted that are the most relevant to what I'm reading now. Boom. Everything is a hyperlink.
This raises a third issue: How would I, the reader, know which words or ideas are interesting to click on?
That, too, can be solved similarly. The computer can look at every word on the page, every phrase, name, quote, and section of text, and show me a "map" of the words and ideas behind which lay the most interesting ideas I might want to know about. Links are no longer lonesome strands precariously holding together a sparsely connected Web, but a booming choir of ephemeral connections tightly binding together everything I have read and I will read. From explorers walking across unknown terrain guided only by the occasional blue text, we become master cartographers, with every path and trail between our ideas charted out in front of us.
Zuck, in an internal memo about Messenger as a platform in 2013:
We're focused on quality because we think the issue is that the cost of using our platform is too high, when the bigger issue is that the value is too low compared to emerging alternatives. We do have real quality work to do, but we must also acknowledge that no amount of quality fixes will increase the potential value of using our platform,
It seems like knowledge tools are stuck in a similar delusion: we think the basic ideas (bidirectional links, tags, and search in an outliner) are right, but that the cost of using them is too high. So we polish and polish and polish, but really the marginal value they offer compared to pen and paper (or Apple Notes) may be simply too low. They replace problems of recall and speed with problems of information overload and tag fatigue. I don't think they enable fundamentally new ways of working with information. They mostly just feel like ever more beautiful facades in front of a growing mountain of complexity.
Computers are more than fast typewriters, because computers can understand language as more than a sequence of bytes. Computers should enable people to think completely new thoughts and solve previously intractable problems. Simply being more efficient is not enough — we need software that enables creation not possible without it.
A good text-to-image prompt:
fine-detail vibrant New Yorker style illustration of an inventor in his cyberpunk lab surrounded by his tools, ideas, notes, and robots
Evocative food for thought:
We can now hold conversations and write stories alongside AI, but for creative knowledge work there's nothing quite like standing around a huge sheet of paper or whiteboard and sketching/brainstorming with another human by producing a trail of diagrams, words, and illustrations all tangled up in arrows and circles. How can we get closer to whiteboarding with an AI?
Evocative phrases for my work:
- Information wants to be connected
- Documents that think for themselves
- The world as one big document
Just ran some basic analytics on YC Vibe Check from my Nginx server logs using some oak pipe and UNIX pipeline magic. The TL;DR from the first week —
- Around 650 unique users
- 3-4 searches per person on average (2100 total searches)
- Top 3 users did 120 searches
- Top 25 users did 450 searches
- Some notable popular companies include Stripe, Replit, Opensea, Substack
- The most popular topics people searched for, starting with the most popular:
- logistics
- crypto
- crm
- carbon capture
- browser
- supply chain
- manufacturing
- health
- note taking
- API
As AI systems get more and more capable (and comparatively less understandable), there will be more and more leverage placed on the interfaces through which humans work with these capabilities. It seems like a question at least as important to study deeply as the AGI question.
A mathematician without the interface of modern notation is powerless. A human without the right interfaces to superhuman intelligence will be no better. Interfaces and notations form the vocabulary humans and machines use to stay mutually aligned. Interface design, then, is an AI alignment problem.
Chat isn't it. Prompt engineering isn't it. Algorithmic feeds are definitely not it. The answer will be more multimodal and lean on a balance of human abilities in language and other senses.
At the base level, the fundamental question in this space is: what is the right representation for thought? For experience? For questions? What are the right boundary objects through which both AI systems and humans will be able to speak of the same ideas? What are their rules of physics in software interfaces?
What happens if we drag-to-select a thought? Can we pinch-to-zoom on questions? Double-click on answers? Can I drag-and-drop an idea between me and you? In the physical world, humans annotate their language by an elaborate organic dance of gestures, tone, pace, and glances. How, then, do we shrug at a computer or get excited at a chatbot? How might computers give us knowing glances about ideas it's stumbled upon in our work?