Meta vs chatGPT: The war of AI is just getting started

Meta has introduced LLaMA-13B, a new AI-powered large language model (LLM) that reportedly outperforms OpenAI's GPT-3 model despite being significantly smaller in size. LLaMA-13B is part of a new family of language models called Large Language Model Meta AI (LLAMA), which includes models ranging from 7 billion to 65 billion parameters in size. The models were trained using publicly available datasets, which means they could be potentially open sourced. Meta hopes that LLaMA will be useful in natural language research and power applications such as question answering, natural language understanding, and reading comprehension.

What's interesting about LLaMA-13B is that it reportedly outperforms GPT-3 while running on a single GPU, which means that it could potentially enable ChatGPT-style language assistants to run locally on consumer-level hardware such as smartphones and laptops. This is a significant development because larger models generally require more computing resources to run, so if a model can achieve the same results with fewer parameters, it represents a significant gain in efficiency.

Today we release LLaMA, 4 foundation models ranging from 7B to 65B parameters. LLaMA-13B outperforms OPT and GPT-3 175B on most benchmarks. LLaMA-65B is competitive with Chinchilla 70B and PaLM 540B. The weights for all models are open and available at https://t.co/q51f2oPZlE1. Twitter: https://t.co/DPyJFBfWEq

Meta calls its LLaMA models "foundational models," which means they intend to use them as the basis for future, more-refined AI models built off the technology. The company hopes that LLaMA will be helpful in natural language research and understanding the capabilities and limitations of current language models.

While a stripped-down version of LLaMA is available on GitHub, the full code and weights are currently only available to researchers who request access. Meta has not announced plans for a wider release of the model and weights at this time. However, independent AI researcher Simon Willison predicts that language models with ChatGPT-like capabilities could be running on mobile phones and laptops within a year or two.