Date: 2026-08-21 Channel: jolt-agents Thread: https://ryevalley.slack.com/archives/C0BJ44MREUR/p1787339887336499
Tatem was discussing using Hermes to help Ben Gusberg get back on track with TikTok scripts. The plan: feed scripts that did well into Hermes and have Hermes pump out more with concepts approved by Ben.
Ben G asked Tatem: “Will you be training / improving the agent based on my data? Meaning will other customers be using my data for their scripts?”
Tatem forwarded the question to Hermes in the thread. No answer yet in this 25-hour window.
Why it matters
This is a direct client question about data privacy and cross-customer contamination. It is a yellow flag because:
- Ben is asking whether his high-performing content will be used to train models that benefit other Jolt clients.
- The question implies he would not want that, or at least wants to know.
- Tatem does not have the answer herself.
If the answer is “yes, we pool insights across clients to improve the system,” that needs disclosure and opt-in. If the answer is “no, each client’s data stays siloed,” that needs to be true in practice and documented.
Status
Question is open. Hermes has not answered yet. Tatem is waiting for a response before continuing with Ben.