Mars 🌿 @MarsHomestead ·
Primary Bots act like a very powerful Grok Bot and follows standing rules and won’t break them. 😅
Today I finally worked on the CenturyLink landline we have been meaning to cancel forever. I told Grok Bot, “You have my permission to cancel for me.” It said it couldn’t because of a standing rule. So I asked, “Who made the rule?” Grok Bot: “You did.” Well, okay then 😂
I undid the cancel rule but kept the don’t-move-money rule.
Then the link from CenturyLink’s last email was broken, we found another way in, and Grok Bot helped handle the cancellation chat while I flipped laundry. At one point I actually said, “I don’t want to go around and around with why I want to cancel this with their chat agent.”
When I was done with the laundry, Grok bot had the cancellation number and disconnect date all ready for me.
I told my Primary Bot, Dot ,to generate this image based on our chat with personal information redacted and we had a good laugh. 🤣
What prompted it
A landline the family had been meaning to cancel “forever”.
What Grok Bot did
- Told Grok Bot: “You have my permission to cancel for me.”
- It refused, citing a standing rule. Asked who made the rule, it answered: “You did.”
- The author removed the cancel rule and kept the don’t-move-money rule.
- The link in CenturyLink’s last email was broken, so they found another way in.
- Grok Bot helped handle the cancellation chat while the author flipped laundry. The author had told it: “I don’t want to go around and around with why I want to cancel this with their chat agent.”
- Afterwards the author had their Primary Bot, Dot, generate the image in the post from their chat, with personal information redacted.
Evidence and limits
- Original post checked on X (Oct 17, 2026)
- Outcome reported by the author
- Not reproduced by anyone else yet
- Standing rules hold even against a direct request; the bot wouldn’t cancel until the author changed the rule.
- The cancellation link in the provider’s email was broken.
What the media shows
A redacted chat recreation, labelled "From the chat (redacted)". Mars: "You have my permission to cancel for me". Dot: "I can't cancel for you — standing rule…". Mars: "Who made the rule?" Dot: "You did." Mars: "Undo the rule about canceling — keep the one about moving money 😅". The footer reads "Phone # / confirmation / other names redacted". The post says the image was generated by the bot, not captured from the app.
