"In complex environments, fluid teams can increase agility" and "Fluid teams have more control over what they are working on than stable teams"

What evidence do you have to actually support these claims? It seems to me that such claims have significant ramifications on those who would follow it. So should we be making such claims if that evidence isn't there (or at least not presented here)?

I know there's a lot of talk about fluid teams in the Agile sphere. But the scientific evidence we have about team dynamics suggests the opposite; namely that stable teams are more performant - especially in volatile environments. It is for this reason that first responder teams train together and remain together. Same goes for specialized army units. Both types operates in highly volatile environments.

There is a substantial body of work that explains why this is so (e.g. DeChurch & Mesmer-Magnus, 2010 provide an overview). When people work closely together over time, they develop shared mental models about who does what in a team and how their individual actions should be synchronized in order to achieve a shared goal. In teams where these shared models don't exist yet - e.g. new teams - a lot more explicit communication is required, and hence productivity suffers. Similarly, we know that it takes time in teams to build psychological safety and trust (e.g. Edmondson, 1999).

I am fully for empowering teams to change their composition as needed. At the same time, and given the scientific evidence I am aware of, I would also recommend teams to stick together as much as possible.

I also think that if fluid teams are indeed shown to be effective for Scrum teams, we need to think more about how to support those teams. Simply putting people together at will won't help. If we take the research on team mental models into account, these groups of people should be thoroughly trained together also.

I would happily engage with you and others to investigate existing research that may shed a light on this question. Until then, I would be very careful with these kinds of recommendations. They may cause harm to teams and their organizations.