Meta New AI Model: Throughout this year, discussions surrounding artificial intelligence have been on the rise. As we enter the final stretch of 2023, an increasing number of companies are actively embracing AI and harnessing its potential for their strategic advantage. While OpenAI, Google, and Microsoft appear to be at the forefront of the AI race, Apple and Meta have remained somewhat discreet on the matter. However, recent reports have shed light on Apple’s substantial daily investments in AI and their pursuit of multiple AI initiatives. In a parallel development, emerging reports indicate that Meta is also vigorously developing an AI model with the intent of challenging OpenAI’s formidable ChatGPT-4.
As per a report initially credited to The Wall Street Journal and later featured in The Verge, Meta’s upcoming AI system appears to be potentially surpassing OpenAI’s most advanced language model, the GPT-4, in terms of sheer power. Furthermore, Meta is actively implementing strategies to reduce its dependence on Microsoft for the training of its AI system.
Table of Contents
What is Meta New AI Model?
The report indicates that Meta is gearing up to commence the training of its forthcoming AI model in early 2024. With this new large language model (LLM) on the horizon, Meta aims to introduce an AI system that surpasses even the capabilities of OpenAI’s GPT-4. Currently, GPT-4 powers ChatGPT’s premium version, serving as the most advanced generative AI chatbot capable of coding, crafting poetry, tackling intricate problems, and much more.
Furthermore, the report unveils Meta’s strategic moves, including increased procurement of Nvidia’s H100 AI-training chips and infrastructure enhancements designed to reduce reliance on Microsoft’s Azure cloud platform for training the upcoming chatbot. Earlier this year, Meta assembled a dedicated team to accelerate the development of AI tools capable of emulating human expressions.
The intriguing question remains: What exactly will Meta’s new AI chatbot accomplish, and how will it measure up to ChatGPT 4 in terms of power and capabilities? Only time will unveil the answers to this compelling inquiry.
What is ChatGPT?
Regarding ChatGPT 4, it made its debut in March 2023, characterized as “a large multimodal model capable of processing both image and text inputs, generating text outputs. While it may fall short of human capabilities in certain real-world scenarios, it demonstrates human-level performance across various professional and academic benchmarks.”
At its launch, GPT-4, the successor to GPT 3.5, garnered attention for its enhanced capacity to handle more substantial and intricate queries compared to its predecessor. OpenAI even provided a detailed list of exams that GPT-4 had successfully completed upon its introduction, along with the corresponding scores, which proved impressive. For instance, the language model achieved an 88th percentile score in LSAT, an 89th percentile score in SAT Math, an 80th percentile score in GRE Quantitative, and exceptional scores of 99th percentile in GRE Verbal and 54th percentile in GRE Writing.