Related:
- Causation:
- 1-2g2s8 The unobserved parts of the wider phenomenon have in no way affected what we observe, yet they are essential to its explanation. Causation doesn’t exhaust explanation.
- 10-2g2e1 Our best explanation invokes abstractions including causation and the laws of physics
- 13-5g Do not confuse the effect for the cause—to explain the causality at play, you must explain what must have happened at first and what would have happened in the absence of change
- Correlation:
- 2-1b2b5 Know the correlations between your bets. Holy grail is fifteen or so uncorrelated bets.
- 2-3b ‘Leverage’ - The effort put in and its utility-results doesn’t have to correlate at all. Use this to your advantage.
- 3-1c3c3a Realize that having invested time in something doesn’t make it good. There is no necessary correlation between the time you put in and its usefulness.
- 7-1b3c Some technologies will be structurally overlooked by corporate networks. Barriers to entry doesn’t necessarily correlate to its importance.
- QUE5 - Are they timeless and universal (if yes, things are likely to be uncorrelated)
A public lecture delivered November 1996 as part of the UCLA Faculty Research Lectureship Program
The Art and Science of Cause and Effect
- Inspired memos:
- Correlation does not mean causation—as in the latter doesn’t have to exist, not like the former doesn’t mean the latter (which was how I understood the quote until I read this paper), it’s more radical as it denies the latter completely
- We are rarely concerned with the entirety <> analysis is articulation (digitization) <> win where you can <> why we are fallible (p. 420)
- Inspired by P{y|do(x)} notation:
- The transition from Roman numerals to Indian numerals—How did it happen? How was the concept of zero discovered? What’d be the modern equivalent?