March 10, 2023
Meta's AI language model, LLaMA, aimed at spurring research and improvements in AI safety, has made headlines following an unexpected leak. Just a week after its announcement, the model found its way onto 4chan, stoking a heated debate about the pros and cons of open access to AI systems. While the LLaMA model requires significant technical skill to operate and hasn't been fine-tuned for conversation like its counterparts, its potential is undeniable. Available in four sizes and outperforming other major models in benchmarks, LLaMA's leak poses questions about the future of AI development and the ongoing battle between open and closed AI systems.
Read more: