What are pros/cons? Is it snake oil or is it actually useful?
I’ve only seen self organising teams in a flat structure fail, on count that if everyone is responsible for everything there are many things no one is responsible for. I think “bike shedding” is very much a thing. Most people will hitch their contribution to what they can understand leaving many of the hard and but vague things in the wind.
Organizations are highly contextual entities; they are tightly coupled to the external environment and culture. Hence, making organizations better is almost always a study in the socio-economic conditions of the organization. The place and times matter more than anything else in setting up successful policies.
When we study history, we see very few exemplary system builders. One such builder was Lee Kuan Yew, builder of modern Singapore. Another - Admiral Rickover, who built US navy's first nuclear submarine. Both of these people had a great "ways" of handling things at an organizational level, I'd say. Their system was mostly focused on: studying the specific states and ailments of the organizations, using erudition, worldly knowledge, and common sense to propose a way forward, and to commit their lives to improving the situation, and then simply keeping the course.
So coming back to your question: perhaps some ideas under the banner of holocracy could be useful/applicable to some organizations, operating under some specific socio/economic/cultural conditions. That's quite a lot of caveats because, when it comes to organisations, there is rarely a satisfying wholesale answer.
PS: I am an avid student of organizations, and am always in the search for potentially universal (or at least largely applicable) truths about organizations. The result of my attempt is my book "Turnover and Other Organizational Ailments" (Free to read on the web [1]). I have, in the book, some validated (not terribly sensational) conclusions about organizations that are supported by empirical/longitudinal studies.