Meet Grok: Elon Musk's AI chatbot coming to X with attitude and a sense of humor
In July, Elon Musk shared that his new company, xAI, will be developing a "maximum truth-seeking AI" that tries to understand the nature of the universe. Well, now the AI chatbot is official. Its name is Grok, and it is modeled after the Hitchhiker’s Guide to the Galaxy and as the company says, it is “intended to answer almost anything and, far harder, even suggest what questions to ask.”
xAI's official statement describes Grok as possessing "a bit of wit" and "a rebellious streak," making it adept at addressing those "spicy questions" that other AI chatbots might avoid.
However, Musk hinted that Grok would refrain from answering specific sensitive queries and leading up to the release, he posted an example on X, showcasing Grok responding humorously to a request for a step-by-step cocaine recipe.
Looking at the screenshot, Grok's response to this specific question appears to carry more humor compared to ChatGPT or Bard. However, it remains unclear if it is a canned response or a genuine display of the system's humor, as Musk claims.
xAI and Elon Musk emphasize Grok's unique advantage—real-time knowledge of the world through the X platform. Grok has "real-time access" to posts on X, as well as capabilities already seen in ChatGPT or Google’s Bard, like internet browsing to fetch updated information on specific topics.
Right now, in its early beta stage with just over 2 months of training, Grok is available for a limited number of users in the United States to test the prototype and provide feedback for further enhancements before a broader release. Those interested can join the Grok waitlist.
Musk also announced that subscribers to X’s Premium Plus plan, costing $16 per month for ad-free access to X, will gain access to Grok once it exits the early beta phase.
The post explains that the driving force behind Grok is the large language model, Grok-1, developed over several months on a cluster of "tens of thousands" of GPUs (Graphics Processing Units). Training data is sourced from the web, and feedback from human assistants referred to as "AI tutors."
xAI claims that Grok is “surpassing all other models in its compute class,” probably referring to OpenAI’s ChatGPT and Google’s Bard. In the blog post, xAI hints at Grok’s future evolution, stating that to better assist users, Grok will be equipped with other senses, such as vision and audio, for broader applications and real-time interactions.
What is funny (or at least in line with Musk’s sense of humor), Grok was introduced just after Elon Musk took part in the first AI Safety Summit, hosted at Bletchley Park, England, where he warned that AI was “one of the biggest threats to humanity.”
xAI, under Musk's leadership and with experts from various fields, including veterans from DeepMind, OpenAI, Google Research, Microsoft Research, Tesla, and the University of Toronto, collaborates with X and other companies to drive the development of Grok.
However, Musk hinted that Grok would refrain from answering specific sensitive queries and leading up to the release, he posted an example on X, showcasing Grok responding humorously to a request for a step-by-step cocaine recipe.
xAI’s Grok system is designed to have a little humor in its responses pic.twitter.com/WqXxlwI6ef
— Elon Musk (@elonmusk) November 4, 2023
xAI and Elon Musk emphasize Grok's unique advantage—real-time knowledge of the world through the X platform. Grok has "real-time access" to posts on X, as well as capabilities already seen in ChatGPT or Google’s Bard, like internet browsing to fetch updated information on specific topics.
Right now, in its early beta stage with just over 2 months of training, Grok is available for a limited number of users in the United States to test the prototype and provide feedback for further enhancements before a broader release. Those interested can join the Grok waitlist.
Musk also announced that subscribers to X’s Premium Plus plan, costing $16 per month for ad-free access to X, will gain access to Grok once it exits the early beta phase.
So, how did Grok come to be?
The post explains that the driving force behind Grok is the large language model, Grok-1, developed over several months on a cluster of "tens of thousands" of GPUs (Graphics Processing Units). Training data is sourced from the web, and feedback from human assistants referred to as "AI tutors."
What is funny (or at least in line with Musk’s sense of humor), Grok was introduced just after Elon Musk took part in the first AI Safety Summit, hosted at Bletchley Park, England, where he warned that AI was “one of the biggest threats to humanity.”
Grok is just a part of Musk's plans to turn X into an everything app. Recently, he spilled the beans on how X will become a financial powerhouse in a year. Elon Musk took the reins over Twitter a year ago, and in this time, big changes have swept through the social media platform, now rebranded as X.
Things that are NOT allowed: