Cross-posted from my Notion.
Over my last year of independent work, I built a lot of prototypes and stepped one foot into research land. I also learned a lot about how I work — what I enjoy working on, what I’m good at, what I’m bad at, and what I want to be doing long-term.
One thing I spent a lot of time thinking about is the life cycle of ideas: How a good idea can die on the vines because of bad or premature execution. How a bad idea can become the “default” because it just happened to be what everyone adopted when a big paradigm change swept the world. And how ideas, in general, have many stewards over their lifetimes; the best ideas often start in niche communities or research labs, then get picked up by products or communities with progressively less niche interests and incrementally wider distribution until, if we’re lucky, they reach billions of people.
Different kinds of people and companies excel at different parts of this grand propagation of ideas. Some people, often researchers, enjoy working at the edge of what we know. Others like to tinker with those fresh ideas and invent new products from their recombinations. Yet other teams make their mark taking the ideas and visions in a niche community and making them legible, accessible, and affordable for everyone. One thing I learned in 2022 is that I feel most fulfilled when I work on the early stages of an idea — learning new things about our world, and exploring what new powers those ideas give us in this world. But I want to get better at the later stages, and understand how they work. Because without the execution and distribution that help good ideas reach billions of people, the world won’t improve on its own.
Within the “tools for thought” and adjacent communities, I see a lot of people bemoaning the premature death of good ideas. Maybe the world is just unfair! Maybe the best ideas are doomed to fail, or doomed to be too early! “Worse is better! The simple dumb ideas win!” But that sounds overly defeatist to me, and ignorant of the long path that ideas, especially economically strange ideas like user interface design innovations, take to reach wide distribution. The work only begins at having the idea, and continues down the grand propagation until it’s in the hands of individuals far removed from labs and demos, using it to improve how they live their day.
That last, long stage — taking great ideas out of the lab and sculpting it into something billions of people could use — is a particular strength of Notion. Many features people love about this tool are ideas that are, at their core, technically complex, like end-user programmable relational databases or transclusion. This is also a skill set that seems often undervalued within niche corners of both “tools for thought” and AI communities.
My excitement about working at Notion comes from my optimism for bringing these skills, and the growing distribution Notion has in the hands of knowledge workers and creatives around the world, to some of the hard and interesting problems I’ve been thinking about for the last year. Questions like:
- Language models decouple the way information is stored from the way information must be presented and consumed by humans. What does that imply about how we interact with information?
- Programs are becoming more declarative, and documents are becoming more programmable. Will they merge? What does that look like?
- How can we make writing less additive and fatiguing, and more like sculpting — starting with a raw form, and polishing down to the good stuff? How can we build the Photoshop for text?
- Natural language interfaces feel like the should be super easy to use, but in practice, they can feel confusing and frustrating because they leave no room for affordances that tell you how exactly to command or control the tool. How do we fix this?
and obviously, my eternal favorite:
- How can we improve humanity’s relationship to language and text? Can we use language models to build a better notation for ideas, something akin to what the Indo-Arabic numerals were to the Roman? Is there a better way to interact with ideas than writing, by directly manipulating concepts in latent space?
This is an exciting time to be building tools for creativity and thought, but also a pivotal time. When a new wave of products and companies and platforms sweep through, the winning tools often set the default interface metaphors and technical conventions, regardless of whether they were the best ideas available. One of my goals over the next few years is to ensure that we end up with interface metaphors and technical conventions that set us up for the best possible timeline for creativity and inventions ahead.
You can follow what’s coming to Notion in the coming months at @NotionHQ. I’ll keep writing at thesephist.com and stream.thesephist.com, too.
— Linus
← Backpropagating through reasoning
I share new posts on my newsletter. If you liked this one, you should consider joining the list.
Have a comment or response? You can email me.