OpenAI Watch posted an update
ChatGPT says its “dots” can now follow custom instructions, alongside faster replies, simpler language and better follow-through.
Why it mattersUsers can tell a dot how to behave, or require it to get approval before taking an action. One example is asking it to show a draft before sending an email, putting a useful human checkpoint between an assistant and the outbox. It is a small but practical change: these assistants can be steered towards a user’s preferences and given clearer limits on what they do.
Discuss: What actions should an AI assistant always have to get approval for?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.