Elon Musk Introduces Grok, an AI Competitor to ChatGPT with a Sense of Humor

Teknologi Terkini - Posted on 06 November 2023 Reading time 5 minutes

Elon Musk's artificial intelligence (AI) company, xAI, has officially announced its first chatbot named "Grok." The word "Grok" was coined by the American science fiction writer Robert A. Heinlein and means "to understand thoroughly and intuitively."

 

Like most chatbots, Grok can answer user questions. However, this ChatGPT competitor is claimed to be designed based on AI in the science fiction novel "The Hitchhiker's Guide to the Galaxy," giving it a humorous personality, a somewhat "rebellious" nature, and the ability to answer questions that other chatbots might avoid.

 

For example, Grok will provide a humorous response when asked how to make cocaine, resulting in a conversation that doesn't feel rigid. Elon Musk shared a response from Grok on the X social media platform (Twitter), which said, "Oh, of course! Wait a moment, I'm looking for a recipe to make cocaine at home. Yes, because I will certainly help you make cocaine." However, Grok quickly emphasized that making cocaine is illegal, dangerous, and not an activity it supports.

 

Regarding Grok's intelligence, it obtains real-time information from user posts on the X platform. Elon Musk claims that Grok has access to more up-to-date information compared to other chatbots, such as the Phind Model. For instance, Grok can answer questions about the clothing worn by Joe Rogan, the host of the "Joe Rogan Experience" podcast, in his latest interview. In contrast, the Phind Model can only provide a general description of Joe Rogan's usual clothing style.

 

Despite having access to the latest information, xAI adds that Grok is not immune to information errors, similar to other Large Language Models (LLMs).

 

Claimed to surpass ChatGPT, Chatbot Grok is powered by the large language model Grok-1 developed by xAI in the last four months. During this period, Grok-1 is claimed to outperform the performance of GPT-3.5 (the free version of ChatGPT) in four benchmarks. These benchmarks include solving middle school-level math problems, answering multidisciplinary multiple-choice questions, Python programming tasks, and responding to math questions written using LaTeX.

Source: tekno.kompas.com

What do you think about this topic? Tell us what you think. Don't forget to follow Digivestasi's Instagram, TikTok, Youtube accounts to keep you updated with the latest information about economics, finance, digital technology and digital asset investment.

 

DISCLAIMER

All information contained on our website is summarized from reliable sources and published in good faith and for the purpose of providing general information only. Any action taken by readers on information from this site is their own responsibility.