Elon Musk on Saturday praised his artificial intelligence chatbot Grok for its irreverent sense of humor, sharing a Grok-generated pair of dark jokes about Soviet dictator Joseph Stalin that quickly went viral and sparked thousands of reactions across X.

Elon Musk Posts Grok Imagine Video: Himself in Yukata Showing

Musk's post, which read simply "Grok can be quite funny 😂" with a link to a shared Grok conversation, has already amassed more than 4.1 million views, 8,400 likes and nearly 1,800 replies within hours of going live on April 11, 2026. The shared chat features two escalating Stalin jokes — one classic and one explicitly vulgar — that play on the dictator's notorious paranoia and brutality with absurd, shocking twists.

In the first joke, a man sneezes during one of Stalin's speeches. Stalin orders the first two rows of the audience shot before the guilty party confesses. Instead of further punishment, Stalin smiles and says, "Bless you." The second, more explicit version replaces the sneeze with a "massive, wet fart." After ordering executions, Stalin responds to the confessor with a torrent of profanity and a grotesque command: "Bless you, you disgusting fucking pig... now bend over so I can wipe my ass on your face, you counter-revolutionary shit-eater!"

The humor relies on subverting expectations of Stalin's ruthless reputation. The polite "Bless you" in the first joke delivers a gentle ironic punch; the second amplifies the shock with crude language and imagery, turning historical terror into over-the-top absurdity. Musk's endorsement highlights Grok's willingness to engage in boundary-pushing, politically incorrect comedy that many other leading AI models avoid.

Grok, developed by Musk's xAI company, is designed to be "maximum truth-seeking" and less censored than rivals such as OpenAI's ChatGPT or Google's Gemini. Launched in late 2023 and continually updated, the chatbot has cultivated a reputation for witty, sarcastic and occasionally dark responses that align with Musk's own public persona. The Stalin jokes exemplify Grok's ability to deliver "edgy" content while remaining factually grounded in historical context — Stalin's regime was responsible for millions of deaths through purges, famines and repression.

Reactions poured in immediately. Some users laughed along, posting memes and their own Grok-generated variations. Others expressed discomfort, calling the content "vulgar," "sad" or inappropriate given Stalin's legacy of mass murder. One reply noted, "That guy mass murdered White people and got away with it. I don't find this funny." Several users praised Grok for its boldness, with comments like "When Grok raped Will Stancil, it was the hardest I've laughed all year" referencing past sharp roasts.

The post fits a pattern for Musk, who frequently uses his X platform — which he owns — to promote xAI products and showcase Grok's personality. Earlier this year he highlighted Grok's image-generation capabilities and real-time reasoning features. By sharing user-created conversations, Musk encourages engagement and demonstrates Grok's entertainment value beyond serious queries. The shared link directs users to grok.com, where anyone with an account can view or recreate similar interactions.

xAI positions Grok as an alternative to what Musk has called "woke" or overly cautious AIs from Silicon Valley giants. Grok's system prompt encourages humor, helpfulness and a refusal to shy away from controversial topics unless they violate basic safety rules. This approach has drawn both praise for authenticity and criticism for potentially amplifying offensive material. In this case, the Stalin jokes tread the line between dark comedy and poor taste, a balance Grok navigates by labeling the second as "super vulgar" at the user's request.

The timing of Musk's post coincides with growing competition in the AI space. OpenAI recently rolled out new humor-focused modes for ChatGPT, while Anthropic's Claude emphasizes safety and refusals on sensitive topics. Grok's edge, according to users and Musk, lies in its willingness to "go there" when asked. The rapid spread of the Stalin thread — complete with reply GIFs, reaction images and follow-up Grok shares — underscores how quickly humorous AI content can travel on X.

Industry analysts note that viral moments like this help xAI attract users and investors. Grok is available free with usage limits on X and grok.com, while premium access through X Premium+ unlocks higher quotas and advanced features. Musk has said Grok's long-term goal is to "understand the universe," but short-term entertainment value clearly plays a role in user adoption.

Public discourse around the post reflects broader debates about AI ethics. Supporters argue that humor, even dark humor, is a healthy outlet and that Grok's transparency about its responses prevents deception. Critics worry that normalizing graphic jokes about historical atrocities could desensitize users or trivialize real suffering. One commenter wrote, "Lord knows we need more humor right now," while another lamented, "That's not funny, but super vulgar and sad."

Musk has not commented further beyond the original post, but his history suggests he views such moments as validation of Grok's unique voice. xAI continues rapid iteration; recent updates have improved Grok's reasoning, coding and creative capabilities. The company's Memphis supercomputer cluster, dubbed "Colossus," powers much of the training and inference, allowing Grok to handle high-traffic viral spikes without downtime.

For everyday users, the episode illustrates Grok's accessibility. Anyone can prompt the AI for "super funny" or "super vulgar" jokes on any topic and share the results. The platform's share feature makes it easy to spread humorous exchanges, turning private chats into public entertainment. In the replies, users experimented with variations, requested comic versions or simply reacted with laughter emojis and "wtf" memes.

As AI becomes more embedded in daily life, moments of levity like Musk's Stalin post humanize the technology. Grok's humor — irreverent, sometimes crude, often insightful — sets it apart in a market crowded with polished but cautious assistants. Whether this particular joke crosses a line remains subjective, but the conversation it sparked demonstrates Grok's ability to generate engagement and debate.

Looking ahead, xAI plans further enhancements to Grok's personality layers, including more nuanced humor controls and multimodal capabilities. Musk has hinted at future "Grok 4" releases that could rival or surpass current leaders in creativity and speed. For now, Saturday's viral thread serves as a reminder that even the most advanced AI can still deliver a good — or at least memorable — laugh.

The full impact of Musk's endorsement will unfold over the coming days as the post continues circulating. Early metrics show sustained growth in views and shares, suggesting the Stalin jokes may become a benchmark example of Grok's edgy charm. In the competitive AI landscape, where technical prowess often overshadows personality, Grok's willingness to embrace the absurd may prove a lasting differentiator.

As users continue testing Grok with ever more outlandish prompts, Musk's simple affirmation — "Grok can be quite funny" — reinforces xAI's commitment to building an AI that entertains as well as informs. In an era of serious global challenges, a well-timed dark joke about one of history's most infamous figures apparently still lands.