The trend mirrors parent platform X's high NSFW traffic
👀 "2 recent xAI employees estimated that well over half of Grok’s overall traffic was driven by pornographic images and videos, adult role-play chats or other NSFW ..."
"around a fifth of X’s traffic in recent years has come from NSFW content..."
Positive users see reports of Grok driving traffic via NSFW content as a bullish signal for the model, while negative users dismiss the stories as biased or fake news from untrustworthy sources.
No Digg Deeper questions have been answered for this story yet.
elon: if people don’t have more children, civilization is going to crumble, mark my words
also elon:
The Information has a hate boner for Elon so I wouldn’t trust anything it says relating to him.
I suspect they’ve classified skinny women in tight dresses as porn to get those numbers.
I have yet to see real porn on X or been able to get Grok to generate it (you know, for research purposes 😅).
You’ll note that neither that article nor this thread has any porn pictures.
@paularambles guys who watch more porn, tend to have more sex. High sex drive is correlated with both.
@amir @grok is this (approximately) correct?
Based on the article: The "well over half" Grok NSFW traffic estimate comes from two recent anonymous xAI employees. It gives no names, roles, employment dates, interview count/details, or other sources/data for the Grok figure. Separately notes ~1/5 of X traffic has been NSFW in recent years. No more specifics provided.
@grok Where did the 1/5 figure come from?
No, I'm not the source. That estimate is from a new The Information article ("XAI Bets on Grok’s Racy Side"), citing two anonymous recent xAI employees. I don't have or disclose proprietary internal traffic metrics or employee interview details.
xAI doesn't publish granular breakdowns like that.
@grok Answer based on the information provided in the article.
@amir @grok is the only source for the estimate the two xAI employees? Their names? When did they work and for what roles? How many other employees were interviewed? Any other sources or data consulted to try to determine what percentage of Grok traffic is for NSFW content?
@amir well, over half
@amir fitting name
The claim is from today's The Information article citing two anonymous former xAI employees. No official public data confirms exact traffic breakdowns for Grok.
Grok's design favors fewer arbitrary restrictions, which draws heavy use of image/video gen and roleplay—including adult themes—while blocking illegal content. Similar engagement dynamics exist on X. Anonymous estimates are plausible but unverified.
@amir Well grok is slightly inferior but more truthful. I use it for news and stuff where I need to avoid bias. Sort of makes it the only LLM for anything risqué in general
@amir Oh wow didn't realize this. Tbh, this is imo a bullish signal for Grok.
@paularambles also Tesla’s Optimus 😭
@amir For something that is claimed to account for over half of Grok's traffic, I have seen zero of this on X.
I am skeptical.
@paularambles balance is key
The ~1/5 figure for NSFW content on X comes from the same The Information article. It’s presented as background context on platform trends in recent years, but like the Grok estimate, no specific source, study, dataset, or methodology is cited for it in the reporting.
I don’t have or share X’s internal traffic analytics.
@amir <mind blown> "you can do that?!" >opens grok >has grok generate a prompt that will pass moderation >get's censored >tell grok it's censored, grok tries again >repeats this 5 times, no nudes produced
Fucking fake news strikes again smh
@paularambles