How Microsoft and Meta will change AI economics with Llama 2

Microsoft and Meta are about to transform the business of AI with Llama 2 with more focus on responsible, ethical, and safe AI practices.

Llama 2 has upgraded open-source code, that has been “red-teamed” for safety and equipped with a responsible use guide.

Llama 2 is probably comparable to GPT 3.5 in terms of performance: According to Meta, its Llama 2 pre-trained models are trained on 2 trillion tokens and have a context window of 4,096 tokens. Meta also says that the Llama 2 fine-tuned models, developed for chat applications similar to ChatGPT, have been trained on “over 1 million human annotations.”

Microsoft is making it available through the Azure AI catalog. AWS is expected to follow suit.

+ With the Llama 2 release on Azure, it is targeted at businesses, not just academics and researchers.

+ Its open-source nature enables checks for biases, inaccuracies, and other flaws.

+ The acceptable use policy prevents potential misuse like criminal activity and spam.

+ Microsoft likely sees Llama 2 as a tool to outmaneuver AI competitors like AWS and GCP; but it might impact OpenAI usage on Azure.

p.s. With the advent of responsible AI like Llama 2, we’re witnessing an era where responsible, ethical, and safe AI practices are becoming the norm, not an exception.

What are your thoughts on Llama 2 release?

#generativeai #llama #azure #aws

Data: Ars Technica