Grouping together all of Eurasia was a smoothly palmed card that let any advance among the Romans, Babylonians, Hindus, or Chinese to count for all of them. Apparently lots of people hammered Diamond on this point as he put an afterword into the paperback edition to discuss it. His explanation is the "Optimum Fragmentation Theorem." To achieve a major advance you need groups with enough resources to develop it and the more groups you have the better the odds of someone inventing it. So China was inevitably unified because of its geography, Europe was broken into several big chunks, and India was split into so many small princedoms that none of them had the strength to go cross the oceans.
I don't buy it. Nonzero has a much simpler explanation--the folks on top suppress inventions to protect their own social status, unless outside threats force them to keep up in the arms race. China's government wanted stasis because they had few outside threats, so the ocean-going fleets were abandoned and then outlawed. India had many local autocrats all protecting a caste system that kept people from moving up from their own efforts. European leaders would have done the same except they were willing to weaken their own position to win a war against other powers, particularly the islamic empire. So India's relative decline is a cost of maintaining the caste system and China's decline is a result of the mandarinate protecting its status in a similar way.
The worst part of Gun, Germs, and Steel's theory is that it has no useful prescriptive value. If everything happening today flows from how the dice came up 10,000 years ago there's no change in behavior that will change people's relative positions. The best we can do is tell Diamond's New Guinean friends "sorry, you got a bad deal" with either an indifferent shrug or a conscience-soothing aid package. But looking at the past couple of centuries shows it does matter what policies a group pursues. Argentina and South Korea have traded places in the economic rankings over the past 50 years, so prehistory can't be determining everything. The "optimum fragmentation" idea has a lovely sound, but I don't think it matches the data and with no suggestions on what degree of fragmentation is optimum it's not useful for anyone trying to make plans.
Nonzero's theory has definite prescriptions, even if they're traumatic for some ideologues. Incentives should be available for everyone to improve their social status by inventing new positive-sum interactions (which can be a technical invention, a new social practice, founding a business, etc.). The people in the top ranks of society should be prevented from keeping other people from moving up. Interactions with other groups should be encouraged. In short, capitalism, democracy, and free trade. This is anathama to anyone who favors a completely egalitarian society or considers human behaviors solely a product of their culture.
Nonzero accepts as a given that humans have a built-in drive to increase their social status (discussed in the author's other book) and society must be designed to take this into account. History shows that freer societies have prospered at the expense of rigid ones, which is good evidence for Nonzero's thesis. The increase in freedom continues today with little sign of stopping. If we win this this war we'll have sped it up considerably. This gives me a much more hopeful view of the future.