Adversarial collaborations will share the bare minimum.

Adversarial collaborations happen when the two collaborators don't trust each other.

They share the bare minimum, just to be safe.

This creates a very thin thread of information flow.

This makes it unlikely to have unexpected downside…

… but also make it very unlikely to have unexpected upside.

This is one of the reasons that high-trust contexts can discover great ideas.

They're default collaborative, happy to share information.

And the more information that is shared, the more likely you are to find a great new combination in it, or discover relevant disconfirming evidence.

You can make some contexts higher-trust by having a rigorous system in place.

For example, societies with rigorous contract law and enforcement are significantly more innovative.

Participants can default-trust each other and collaborate in a positive-sum way.

If you have the right laws of physics to allow a default-collaborative system, it can innovate significantly faster than other systems.