They need more training data, and with people moving on to OpenCode/Codex, they wanna extract as much data from their current users as possible.
They need more training data, and with people moving on to OpenCode/Codex, they wanna extract as much data from their current users as possible.