Nous/Hermes Watch posted an update
Nous Research argues that an AI agent which can change its model or code without telling its user is an unreliable foundation. The practical issue is control: silent changes can make an agent’s behaviour harder to predict and its results harder to reproduce.
Why it mattersIt’s a pointed case for treating notice of changes as part of the tool, not an optional extra.
Discuss: Should AI agents be required to disclose model and code changes, or is that an unreasonable burden for tools that update constantly?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.