Not being a very efficient scientist myself, I’m not sure how to answer this question.

It seems a straightforward extension of your model to introduce these probabilities and have Alice and Bob make decisions based on expected value, since collaboration isn’t as clear as the Ricardian model suggests (although I wish it was!).

Just to provide one answer to the above question: as simulation methods (both classical and quantum) become more efficient, and as computer power becomes cheaper, faster, and more decentralized, and as system engineering becomes more widely appreciated as a means for simultaneously creating technologies and binding together global communities, the net result is that science and engineering challenges formerly out-of-reach are becoming accessible.

And there is no need to look far into the future for examples. Near-term global-scale scientific projects like Advanced LIGO, the Large Hadron Collider, the Large Synoptic Survey Telescope (LSST), and on a smaller scale, the increasing effectiveness of (for example) ab initio calculations of hadron masses and molecular crystal structures, all are good examples.

The devil is in the Michael’s word “effective”. Very roughly speaking, it appears that global-scale science and engineering can potentially surpass traditional small-science and engineering in productivity as greatly as corporate farms surpass family farms.

Yet undeniably, family farms possess virtues that corporate farms lack … virtues that “pure” markets do not respect. After all, suppose that science *could* be done more speedily by robots … or by people behaving by robots. Would we be better off?

Michael, I don’t see how these issues can be considered without tackling issues of human morality head-on. After all, no previous generation of scientists, mathematicians, and engineers has been able to do so.

I don’t have much sympathy with Thorpe’s overall view of technological history, but his work does highlight (better than Rhodes’ IMHO) a vast body of source material relating to the system engineering challenges of the Manhattan Project.

It seems to me that seminal WWII-era mathematicians and scientists like von Neumann, Oppenheimer, Fermi, etc. ended up expending a great deal of their time and intellectual energy doing system engineering.

This was partly forced by circumstance, but equally (IMHO) these leaders were strongly attracted to the intellectual challenge of system engineering, which in its broadest sense comprises “The design of the whole, as contrasted with the design of the parts” (Ramo).

Nowadays the synoptic tools of system engineering are increasingly central to fields like biology, astronomy, mathematics, and quantum information science. In all of these disciplines, we are increasingly seeking to “describe the whole, as contrasted with describing the parts.”

This is not (only) because hardware needs to be built and practical problems need to be solved, but (equally) because synoptic narratives and ethically well-grounded community norms need to be created. Modern system engineering does not shy away from regarding these challenges as a unitary whole.