Elon Musk has launched a chatbot known as Grok, which is expected to pose a significant challenge for ChatGPT. The new AI was built by xAI, a startup he unveiled in July with the mission of “understanding the true nature of the universe.” The company spent months recruiting researchers to work on artificial intelligence that could rival the generative model used in ChatGPT, which launched a tech-industry arms race after its debut nearly a year ago.
Grok is the first product to emerge from xAI, and Musk has been sharing updates on it through his social media platform X, formerly known as Twitter. He has also posted a video of the new AI compared to ChatGPT. The Tesla and SpaceX founder is pushing xAI to become a new source of truth, particularly on real-time news and politics.
Musk claims that the current prototype of Grok is superior to ChatGPT 3.5 in several essential respects. He also says that the bot is designed to understand the meaning behind a given question, as opposed to just responding with information. The company’s blog post about the bot explains that it can recognize sarcasm and other types of humor, as well.
Unlike ChatGPT, which can be prone to providing incorrect or misleading information, Grok is designed to analyze the context of a question and provide an accurate response. The company offers early access to the chatbot to a select group of X subscribers, allowing it to improve its performance based on real-world interactions. Eventually, the company plans to open the chatbot to the general public.
In his latest post, Musk shares a screenshot of a conversation with Grok. In the image, the AI is asked how to make cocaine and responds with a bit of humor. It is a good sign that the chatbot isn’t just going to be an oracle of universal truth but will also be able to provide a little levity and fun along the way.
The chatbot is named Grok because it was modeled after the cult sci-fi book The Hitchhiker’s Guide to the Galaxy. The word Grok, according to a definition from Collins Dictionary, means “to understand thoroughly and intuitively.”
In a separate blog post, xAI said that the Grok-1 model had surpassed GPT-3.5, the model used in the accessible version of ChatGPT, on some benchmarks, such as solving middle school math problems. The company credited its success to its extensive training dataset and high-performance computing resources. The company launched in July with a team stacked with former employees of OpenAI, DeepMind, and more. It is currently hiring for multiple roles.
