What do you, personally, think you will remember most about this interview a year from now?
The most memorable part of this interview is Sir David Spiegelhalter advice about being careful about the limitations of what one can pretend to do. He brought up the example of COVID in the UK and how the politicians wanted forecasters to say what will happen in 6 months, 12 months etc. This was memorable because it was a very human reminder of our need for control and understanding played out into society.
How do you think any aspect of the interview will affect your own future, or society's future?
Sir David Spiegelhalter's discussion of humility really left a lasting impact on me. he first introduced the concept of humility when discussing Black Swan ideas, such as the financial crisis of 2008. He spoke about how people have been proud about having models and being grounded in empiricism, but no matter how much data we obtain, we just "don't know" sometimes. As he puts it, "this is a demonstration of the inadequacy of your first judgement... could be better or could be wrong".
For society's future, I think that across the social sciences and sciences we should always have an open mind that things could completely subvert the way we think about things. This reminds me a lot about Thomas Kuhn's Theory of Scientific Revolutions - that there is a cycle to shifting our system of thinking. In my person life, I think that I need to be aware that there is always more data I can obtain to "update" my judgement and mode of thinking. No matter how confident I am in what I believe, I don't know what I don't know.
I agree with you Michelle completely. It is very interesting how people overestimate the ability of prediction algorithms, especially when we experience something that hasn't happened before, such as Covid or the 2008 Financial Crisis. I wonder, what is the best course of actions then, should we strive to quickly create new models based on current ongoing data or should we apply past models that were made to tackle similar phenomena?