This year 2024 I had Firefly bonus on Adobe in September, dataset earnings on Pond5 in April and November, and contributor funds on Shutterstock in January, March and May. Now the agencies have effectively 'trained' (copied) the assets for their programs they no longer need to keep paying contributors. One small $ step for contributor, one giant $$$$$$$$$$ leap for microstock agencies.   

ETA apologies just realised the original post was specifically regarding Deposit Photos.

@Stoker2014:
I had a dataset earning at Deposit as "kundenspezifischer Verkauf" (sorry for the german text here) on August, 2nd 2024.

Maybe you want to check your earnings stats too, it should be there...

kundenspezifischer Verkauf - these are individual sales ?

This year 2024 I had Firefly bonus on Adobe in September, dataset earnings on Pond5 in April and November, and contributor funds on Shutterstock in January, March and May. Now the agencies have effectively 'trained' (copied) the assets for their programs they no longer need to keep paying contributors. One small $ step for contributor, one giant $$$$$$$$$$ leap for microstock agencies.   

Well you need to retrain each time you upgrade your model. If you change algorithms you need to do machine learning again, again and again to get better results. I did not tried yet but i think there is something like time-frame window to use this "dataset content" (content and metadata).

Also Shuttertock/pond5 uses Databricks Marketplace since June 2024 to sell for machine learning.
Databricks Marketplace is an open marketplace that enables you to share and exchange data assets such as tabular datasets, volumes, notebooks, and AI models across clouds, regions, and platforms.  (read: https://www.databricks.com/blog/shutterstocks-content-datasets-now-databricks-marketplace)

Anyway my guess is SSTK/P5 uses probably something like Databricks Clean Rooms and after a while you need to pay extra. These so called "clean rooms are a  privacy-safe environment for sharing and joining data between organizations, challenges and solutions for safeguarding data, key partnership integrations, industry use cases, and best practices.