When @aronchick mentioned the new Agentic Resource Discovery for service discovery, my reaction was 'I just kind of worked out how MCP and A2A fit together, why do we need *another* protocol?' and then when I started looking into it, that isn't even all the proposed open agent protocols!
Part of this is the pre-Cambrian explosion of experimentation and eventually you get Ediacaran die-off back to what make sense; partly it's that enterprise vendors have both experience in service discovery and an obvious interest in defining how registries & directories of services for agents work.
Those enterprise vendors and the AI labs all want their protocols adopted; that means open governance and that means the AIAF:
@maniksurtani explains to me how the foundation is navigating a list of protocols he expects to get longer before it gets shorter to build a stack for agent work.
MCP is the most mature of the agent protocols: Caitie McCaffrey tells me about MCP 2.0 and the new constructs coming to simplify long-running agents as well as how it fits in with A2A and ARD - because they do have different roles and which you need is clearer once you understand what they each do.
What everyone told me is that this is all still very new and we're making up a lot of this as we go along. But if you want to understand how to give agents connections to the tools and other agents that make them useful, waiting till it's all baked will leave you behind competitors who dive in now.