I've heard directly from someone at Nikon that they have separate teams working on firmware for each camera. We're not talking about large teams, and they reuse code and have a standard framework with camera specific parameters. But they do separate the teams doing the work, they separately decide what is included and what is not, and there is ongoing development and enhancement that gets added but needs to be prioritized, scheduled, and tested. With a new camera, the firmware and hardware need to be final far in advance. If Ricci was testing the camera for more than four months, the firmware decisions were probably made a year ago. The code is linked to the body in many specific cases because you have different sensor readout speeds, different power available, different file sizes and formats, etc. Someone from Nikon described it as "tuning" - the same code can be set with different parameters.
I've also heard directly from Nikon that there are ongoing enhancements included in every firmware update. Subject detection, stickiness and tracking have ongoing adjustment and improvement. Training subject identification is an ongoing effort - not something at the individual team working on firmware for an individual camera is directly involved with. So the Animal ID we used in earlier cameras does not exist any more - it is updated and reflects all the ongoing training that has taken place to improve subject identification even if Bird ID is not a separate choice.
One of the things we learned with the earlier Z cameras is the image we see through the EVF is the one used for on sensor focus. Nikon has just changed the data being read from the sensor. They are populating an EVF with more data - 5.7 megapixels on the Z6iii compared with 3.7 MP on the Z8. That alone should improve focus, but it's more data and a slower readout, so the refresh rate and processor speed also matter.
