Code refactoring of hivemind has been done for a while. A week or so ago we refactored some code in hived to make it easier to updates to the API without requiring as much boilerplate code changes. But in general, code refactoring is never done, you often find ways to improve the code when you work in it.

There hasn't been a lot of API restructuring. We've optimized the performance of the API calls and also brought some of the responses from the various API into greater conformity with each other (mainly in the face of error cases). But restructuring the API itself is a big task, because there's many 3rd party dapps that depend on the API, so any major changes in the API would require coordination changes to all the dapps. So we've limited our work of that type.

Now as part of our modular hivemind work, we're likely to start introducing new orthogonal sets of API calls depending on the configuration of the hivemind selected by the API node operator. For example, a simple "bare-bones" modular hivemind would only have API support for blockchain operation data and account information. And then we'll likely introduce new APIs for expanded versions of a module hivemind that have access to additional data like posts.

But that will probably have lower priority over the smart contract platform that will be built on top of the base module hivemind layer. Our priority for now is 1) get out the base layer modular hivemind and then 2) create the smart contract layer on top of that.

I'll check back on the dev count later, I need to go for dinner now. I think it's still roughly the same at the moment, around 10-12 devs.