I am not sure the case you argue generalises terribly well. Two issues strike me as missing in the argument.
Fail fast models are all based on a presumption that 'early' judgements can be made about whether success is likely. For some things, this is clearly possible. But achieving many outcomes (public policy is a classic example) involve long gestations with complicated endogenous pathways. The risk of your approach is that you create an endless cycle of change without really learning what the ultimate impact of those change are.
The real purpose of elevating decisions in a hierarchy is to unlock judgement and experience. I would generally argue that this is a good approach for high consequence / cost actions. Yes, it may slow things down. But I would argue that, when done well, it allows for far more robust 'theory building'. The real problem is that it is often not done well, and is often applied to low consequence / cost actions.
Behind both of these points lies, I think, a common issue. Implicit in your model is that uncertainty is best addressed within a fast thinking model. I agree that this is sometimes true, but it is not always true. Sometimes in applying imperfect human knowledge imagination and consideration are the key to better decision making.
I also think that in pursuing outcomes, we need to have a clear (if imperfect) idea of how long an experiment needs to run before meaningful results emerge. Many effective cancer treatments make patients sicker in the short term before making them well in the long term. Measuring success at the right point in time is key to achieving the best outcome.
Thanks. And welcome back! There isn't space to reply properly here as you've picked up a lot. I need to write a post on fast/slow thinking and these ideas as I don't think they map onto the same things. Lots of slow decisions are built on fast thinking processes and vice versa.
To pick up on your last point, I think the problem is that we need a much better calibrated sense of how long things should take and when/how to measure results. Decisions tend to either get rushed or endlessly delayed (or both in turn) - and I think we need a better understanding of how to act in the face of uncertainty to change that.
I agree that elevating decisions to unlock experience is often valuable and when done well it matters a lot. My worry is that it is typically not done well because we elevate at the wrong time and for the wrong reasons.
I am not sure the case you argue generalises terribly well. Two issues strike me as missing in the argument.
Fail fast models are all based on a presumption that 'early' judgements can be made about whether success is likely. For some things, this is clearly possible. But achieving many outcomes (public policy is a classic example) involve long gestations with complicated endogenous pathways. The risk of your approach is that you create an endless cycle of change without really learning what the ultimate impact of those change are.
The real purpose of elevating decisions in a hierarchy is to unlock judgement and experience. I would generally argue that this is a good approach for high consequence / cost actions. Yes, it may slow things down. But I would argue that, when done well, it allows for far more robust 'theory building'. The real problem is that it is often not done well, and is often applied to low consequence / cost actions.
Behind both of these points lies, I think, a common issue. Implicit in your model is that uncertainty is best addressed within a fast thinking model. I agree that this is sometimes true, but it is not always true. Sometimes in applying imperfect human knowledge imagination and consideration are the key to better decision making.
I also think that in pursuing outcomes, we need to have a clear (if imperfect) idea of how long an experiment needs to run before meaningful results emerge. Many effective cancer treatments make patients sicker in the short term before making them well in the long term. Measuring success at the right point in time is key to achieving the best outcome.
Thanks. And welcome back! There isn't space to reply properly here as you've picked up a lot. I need to write a post on fast/slow thinking and these ideas as I don't think they map onto the same things. Lots of slow decisions are built on fast thinking processes and vice versa.
To pick up on your last point, I think the problem is that we need a much better calibrated sense of how long things should take and when/how to measure results. Decisions tend to either get rushed or endlessly delayed (or both in turn) - and I think we need a better understanding of how to act in the face of uncertainty to change that.
I agree that elevating decisions to unlock experience is often valuable and when done well it matters a lot. My worry is that it is typically not done well because we elevate at the wrong time and for the wrong reasons.