[Link] Beyond the hill: thoughts on ontologies for thinking, essay-completeness and forecasting

by jacobjacob1 min read2nd Feb 20206 comments


Forecasting & Prediction

This is a link-post for: https://www.foretold.io/c/1bea107b-6a7f-4f39-a599-0a2d285ae101/n/5ceba5ae-60fc-4bd3-93aa-eeb333a15464


Epistemic status: gesturing at something that feels very important. Based on a true story. Show, don't tell. Release early.

Why are documents and spreadsheets so successful?

Why does code, which is many times more powerful than spreadsheets, have many times fewer users?

I think it's because code not just forces you to express your ideas in code, but also to think in code. It imposes constraints on your ontology for thinking.

Having spent the last year working on forecasting, I got some experience with how ontologies can significantly constrain technology projects.

I think such constraints have...

  • heavily limited the usefulness of past forecasting efforts
  • resulted in broad misconceptions about what forecasting could be used for
  • hidden a large space of interesting work that can be unlocked if we solved them

So the link-post is an interactive essay where I attempt to show what solving them might look like in practice, using some technology which is currently not supported on LessWrong.

(Note that the link will not work well on mobile.)