Is Llama 2 going to take over from OpenAI and Google in the LLM market?

While Llama 2 offers transparency and ease of fine-tuning, making it an attractive option for smaller entities and researchers, it's not clear whether it will completely take over from OpenAI and Google in the LLM market. This is due to limitations like licensing restrictions, certain performance benchmarks where it falls behind, and the continuous innovation and advancements from other key players like OpenAI and Google.

I'm positive there will be a LLM market, a significant one.  I am also confident that whether you use a hosted model or a BYO model, you will have security/governance/transparency concerns that are not addressed by the provider ("Please don't let this thing write poems about my CEO").  

If you are creating an AI orchestration layer that sits in front of these models, it shouldn't matter what they are.  In fact, you should probably plan on running multiple LLMs based on what you are trying to do.  I wouldn't put my business in a position where the answer to this question is material.  Check out Microsofts "Copilot Stack".

This is 1989 of the Internet. It existed but almost nobody knew it except universities and research institutions and then the public became aware in early 90’s Many predictions were made in early 90s but 90% were wrong. It will be the same with AI. Someone is making the right predictions but it’s buried with all the wrong predictions. one big difference will be the rate of change will be way faster than the Internet. The winners will likely be clearer in only a few years not decades like the Internet