That’s an interesting debate, and I’m not sure how clear-cut the answer actually is. Today, there are many resolutions in the UN that pass the general assembly, often with overwhelming support, with the US voting alone or with Israel or Ukraine backing it up. A world government led by the general assembly doesn’t seem inherently worse than what we have today. However, before decolonization, I can see the Soviets’ arguments.

It’s odd that Einstein doesn’t answer at all the question of how to ensure that voting in all constituent countries is legitimate. That question is fundamental to that kind of organization. A body where a dictator or colonial governor can just stuff a ballot box and send it in is not a reasonable foundation for a world government.

On the other hand, it’s worth noting that the Soviets did eventually come around to participating in the UN, in the more limited role of fostering international discourse, which Einstein suggests at the end. They also adopted a general policy of “peaceful coexistence” with the West, which included signing international arms control treaties. With no such treaties, and no precedent to know that MAD would actually work, I think Einstein’s position is understandable.

Also, my impression is that it’s more about the idea in principle, as something to work towards, rather than as a practical suggestion, and it’s more common in the intellectual tradition of the West and of social democrats to focus on agreeing to lofty ideas in principle, with the view that disagreements about how to achieve those views in practice can be reasoned out. Marxist-Leninists, on the other hand, treat ideas as inseparable from their proposed execution. For example, in WWI, social democrats all declared their opposition to the war in principle, but it meant nothing, because they found various ways to justify supporting the war in practice, to protect and advance their careers, for example, saying that they supported “peace without annexations” so they need to keep fighting to reclaim lost territory. This lives on today with social democrats and liberals declaring their opposition to “authoritarianism” on principle, even though they frequently justify actions that contradict that in practice. MLs tend to take a more skeptical view of accepting an ideal “in principle,” without concrete details. Einstein wasn’t an ML, so we shouldn’t really expect him to be above that tendency.