Elon Musk unveiled Grok, an AI chatbot to take on OpenAI’s ChatGPT, on Saturday. The chatbot is currently accessible to Premium+ subscribers of X, formerly Twitter, and is in its development phase. Musk teased the launch on X briefing about the features.
Grok, now a part of X, is tailored to respond to questions humorously. The company encourages individuals who do not appreciate humour to refrain from using Grok.
Also Read | Apple is getting “serious” about generative AI, CEO Tim Cook hints during earnings call
Elon Musk recently attended the AI Safety Summit held in early November where he also met India’s IT Minister Rajeev Chandrasekhar and the news of one of the billionaire’s middle name is Chandrasekhar surfaced, during which he discussed the potential risks associated with AI.
Announcing Grok!
— xAI (@xai) November 5, 2023
Grok is an AI modeled after the Hitchhiker’s Guide to the Galaxy, so intended to answer almost anything and, far harder, even suggest what questions to ask!
Grok is designed to answer questions with a bit of wit and has a rebellious streak, so please don’t use…
The Independent reported that Elon Musk expressed his view on AI that it is “far smarter than the smartest human.” He stressed the importance of establishing a policy framework to set guidelines for the AI corps. He reportedly also engaged in talks with UK Prime Minister Rishi Sunak about the same.
What is xAI’s Grok?
xAI’s Grok is an artificial intelligence chatbot that is designed to understand and respond to human language in a comprehensive, informative, and humorous way. It is powered by Grok-1, a large language model that has been trained on a massive dataset of text and code.
Grok can be used for various tasks, including question answering, information retrieval, creative writing, and coding assistance. It is currently available to X Premium+ subscribers in the United States which costs about Rs 1,300 when subscribing via web.
Also Read | WhatsApp will stop working on these phones soon, know details
It operates on the Grok-1 LLM, which has undergone a four-month development process. The initial prototype, Grok-0, was trained with 33 billion parameters, demonstrating comparable capabilities to Meta’s LLaMA 2, which supports 70 billion parameters.
How Grok competes with ChatGPT, others
Grok-1 achieved a 62.9% score on the GSM8k benchmark, designed explicitly for middle-class mathematical word problems (Cobbe et al. 2021). This performance surpasses GPT-3.5 and LLaMa 2 but falls short of Palm 2, Claude 2, and GPT-4. Furthermore, Elon Musk also demonstrated a comparison between “typical GPT” and Grok. He claimed that the latter has current information, that others lack.
Example of Grok vs typical GPT, where Grok has current information, but other doesn’t pic.twitter.com/hBRXmQ8KFi
— Elon Musk (@elonmusk) November 5, 2023
Follow FE Tech Bytes on Twitter, Instagram, LinkedIn, Facebook.