Ideas Worth Implementing as Research Org

Category: Organizing Research

Read the original document

<!-- gdoc-inlined -->

1. Team of Explainers

1. All knowledge of the research frontier should go through you

2. State of the art tracking

   1. Your standard is the gold standard for what the true state of the art methods / techniques are and how well they work and for code that replicates it

3. Understand what science is and make machine learning more like that

   1. Create a list of ‘things that great science does’ and run it on each axis of the research frontier

4. Contrarian truth generator:

   1. Be the first to criticize bad research practices - it’s the thing nobody wants to talk about

5. Replicability

   1. If anything we publish can’t be replicated, we pay money (or something equivalent)

6. Polymath Project

7. Systematizing Creativity

8. Research Superforceasting 1.

Source: Original Google Doc