People often attempt to predict when LLMs will function as 'drop in remote workers.' Before Jan 1 2028, will a leading AI lab pay remote workers to gather training data in order to facilitate this?

This does not mean paying humans to verify the outputs of LLM remote workers, but actually recording and saving a human worker's actions while they solve tasks, for the specific purpose of training on the data.

Resolves on credible news story that they have spent $100k+, or if they announce a plan to spend $100k+ and it seems pretty likely they did before resolution date.

'Leading AI lab' means OpenAI, Google, Anthropic, Meta, or xAI.