Says that we are not equipped in neurologically or physically for weak signal detection as our pleiolescene past was fairly simple. Good job he said we was not a neuoscienticst, he is obvioulsy not a palentologist or has much knowledge of evolutionary pschology if he can say that. I would argue we were really well equiped, but we have in part trained ourselves out of that natural abilities, or allowed technofetishism

States that the purpose of scanning is reduce the degree of surprise in respect of the future(surely its more the impact)

Says that there are three aspects of changing our image of the future.

1. a confirming hit, not worthless, more of the same
2. Resolving uncertainty
3. Creating a brand new scenario is the best hit, but rare

Makes a good point, we should aim to be suprised in small ways not big ways. goes on to argue that horizon scanning is hard as weak signals are weak and early. Our neural system is a physical system of chunking which ends up with a percept with loss of data. We add more to the percept than the world does, which is good neuroscience, talking about it as internal machinery less good. There are a lot of mechanical metaphors in his talk: the instrumentalism of far too much sense-making sociology?

Direct quote from his bullet points:

• We are inveterate pattern recognition machines even when nothing is there (conspiracy, some would say religion)
• Pattern once recognised, is stable until new information arrives. Inertia is like friction, quotes Newton's first law
• What constitutes "new"? An anomaly does not destroy the theory, higher the consequences the greater the resistance to weak signals.

See what I mean about the mechanical metaphor? Each of those points is riddled with it, and its the wrong metaphor; if you define the problem in terms of machines, then your solution is likely to be mechanical.Proceeds to say that we have Inherent tendencies to knowing rather than not knowing and certainty to uncertainty. We all have our own implications and he is using Myers-Briggs types to illustrate this as if that particular bit of pseudo-science has any factual basis. fair enough point about the issue of expert being paradigm bound compared with the novice, so less likely to see new things.

Now looking at all of this as a problem of signals. As they go up the probability goes up, but false positives go up, so he (argues) we are looking for the sweet spot, the right number of signals to balance discovery with minimal false positives. Now this is exactly what I argued against in my summary and he obviously realises this as he references my point about granularity (it will be in the next post)but excuses himself from its application.

Conclusions are that we should

• create a dedicated team (ideally rotating and diverse) of horizon scanners with all the Myers Briggs types present
• Training that team on real signals and impacts. Can we create a management simulator for weak signals?
• require the submission of a specific number of hits per person per month of weak signals
• Review other hits, the team reviews and determine (possibly anon) the quality to see who is in touch with the group mine.
• track false positives and false positives over longer time and adjust the quality of signals to reach desired balance.

And he says that will take ten years! OK this is just what you don't do. More tomorrow