This comes after a morning of massive backlash against the bots «politically incorrect» settings, and «serving up unfiltered truths» — the entire mainstream media piled on to the bots inflammatory and upsetting replies.
It seemed particularly active in the Turkish language, with a particular scathing reply about Erdogan.
Serious damage control
X.com had initially banned hate speech in tweets to the bot, but seems now to have gone further — and shut off the bots replies entirely.
That means it can no longer reply to questions or respond to comments to @grok on the platform.
It can, however, still generate images, and does so creatively, still seemingly agreeing now to being censored by jews.
Emptied account page
This is just three days after teknotum reported on the seemingly rightward turn and embrace of conspiracy theories, and it has grown out of control since.
The main account page for @Grok now sits empty, posts have been taken down, and the only thing there is a pinned comment from the administrators.
Grok 4’s release date is still today at 8pm Pacific Time, and it will be interesting to see how they play the weights and priorities of that newer model.
x.ai and grok.com seems to be regular, filtered and politically palatable still, though.
News: X.com CEO Linda Yaccarino has resigned, saying the future for the platform lies in its partnership with x.ai.
EDIT: Here are the current system prompts for Grok. According to Engadget, the line instructing to “not shy away from making claims which are politically incorrect” was removed sometime on Wednesday, July 9.
Read more: Express Tribune on the shutdown along with Reuters, media reactions from NBC News, CBS News and The Guardian.