How LLMs are evolving and the rise of open source LLM capabilities

With the rise in generative AI and large language models (LLMs), we’re getting close to a mass market adoption of AI in a way we haven’t seen in prior years. OpenAI was founded in 2015 and it took them almost 7 years of research and development before they released the first version of ChatGPT. But LLMs have been in the works for some time. Prior to the LLMs we know of today, there were many early language models by other names:

• N-gram models have been able to predict the next word in a sentence since the early 2000s.
• Feed-forward neural networks have worked as probabilistic language models since 2003.
• Two of the models we covered in Chapter 1, recurrent neural networks (RNNs) and long short-term memory (LSTM) models, have been able to handle longer sequences of prompts and responses since the mid-2010s.
• Transformer models came out in 2017.
• Transformers laid much of...