For the first time ever, parents going through IVF can use whole genome sequencing to screen their embryos for hundreds of conditions. Harness the power of genetics to keep your family safe, with Orchid. Check them out at orchidhealth.com. On this episode of Unsupervised Learning, Razib talks to Cody Moser, co-author of a recent paper, Innovation-facilitating networks create inequality. Moser is an evolutionary psychologist and cultural evolutionist at UC Merced, where he is completing his doctorate. A previous guest on the podcast, Moser immediately digs deep into the abstruse and technical model that shows that more is not automatically better when it comes to innovation and discovery. First, he contrasts his results with the Tasmanian cultural evolution model outlined by Joe Henrich nearly 20 years ago. In short, Henrich showed that very small populations tend to lose cultural traits and skills over time. Going through a population bottleneck has a memetic as well as genetic effect. The converse scenario is one where a large population is able to retain and even accumulate more cultural traits and skills. Moser’s main finding is that some fragmentation of these large populations may in fact foster innovation. On the evolutionary psychological scale, massive groups may tend toward conformity, and disrupting information flows may foster independence of thought. A significant immediate implication is that scholarly thought might benefit from separating into competing schools and departments where distinct groups can develop solutions collectively but retain enough independence to resist being drawn into broader irrational herd behavior. Moser’s results have broader implications for how businesses and corporations should operate, and perhaps quantify why nimble startups often outpace and defeat massive organizations despite the latter having almost infinite resources. Groupthink is powerful. Though small populations will be hit by skill loss with the death of keystone individuals, large populations may ossify, “locking in” regnant ideologies.  Razib also probes Moser about the rise of agent-based modeling and simulations in social science over the last 20 years, and how they have allowed scholars to circumvent the limitations of relying purely on college students to act as experiment subjects.