
Unlocking Creativity with Azure OpenAI
By :

In recent years, there has been a huge buzz around LLMs such as ChatGPT sweeping across the world. LLMs are a subset of a broader category of models known as foundation models. Interestingly, the term “foundation models” was initially introduced by a team from Stanford. They observed a shift in the AI landscape, leading to the emergence of a new paradigm.
In the past, AI applications were constructed by training individual AI models, each tailored to a specific task using specialized data. This approach often involved assembling a library of various AI models in a mostly supervised training manner. Its foundational capability, known as a foundation model (see Figure 1.3), would become the driving force behind various applications and use cases. Essentially, this single model could cater to the very same applications that were once powered by distinct AI models in the traditional approach. This meant that a single model could fuel an...