He categorizes mental types that will probably thrive in the future (I try to compile and select, motivated by a possible future of quant finance).
Freestylers - they use technologies, but overrule them if they have better approaches
Synthesizers - extract knowledge from patterns in massive data
Humanizers - cover machine intelligence with a nontechnical look and feel
Conceptual engineers - use technology for explorative and constructive learning
Moralizers - measure performance of human-machine collaboration - not humans
Motivators - managers who appreciate human efforts based on the joint human-machine outputI see quants and quant developers as freestylers and conceptual engineers - they enjoy the freestyle of quant work, staying strong and agile by using technologies intensively without becoming their slaves. Their innovative spiral is characterized by building higher level tools atop lower level ones.
IMO, this is related to First We Build The Tools - Then They Build Us.
But there are still different paradigms to do that: a fractal model is different from a builder approach. It is not only more efficient, it supports synthesizers, moralizers and motivators. They support a wider coverage, general purpose fitness, automation and understandability.
When we decided for the UnRisk Financial Language we did not fully understand what difference it makes. We explored a lot by using it ourselves - it is symbolic, parallel, declarative and domain specific for multi programming paradigms. Each new construct enriches the programmable and computational knowledge base.