alt Hacker NewsThat is an interesting thought.
They could run some sort of analysis to find high value input, such as proprietary technology, algorithms, or strategy.
Then they could group them together for one specific topic, and produce a report that analyzes if the information is plausible.
If so, they can have it send to staff for review, who could then create a test set that rewards the model for going into the direction of the proprietary solutions known to work.
I'm no expert, but at least something like that sounds plausible to me. I still very much doubt they are doing this.
It's actually simpler: your code is not the valuable part, it's the telemetry/metadata/conversation surrounding your session that's valuable. Every time you press escape, every time tell Claude/Codex that it's being an idiot, your back-and-forth conversations, etc. "when/why did we fail and how can we improve?" is what they want to know.
They can use LLMs to launder confidential customer sessions into trainable data. Then they can claim that they don't train on "your data" without it being technically incorrect.