Foundation Large Language Model (LLM) Stack — Version 4
As LLM adoption grows, the technology and implementation landscapes are unfolding. In my attempt to interpret what is taking place and where the market is moving, I created a product taxonomy defining the various LLM implementation and use cases.
4 min readAug 15, 2023
I’m currently the Chief Evangelist @ HumanFirst. I explore & write about all things at the intersection of AI & language; ranging from LLMs, Chatbots, Voicebots, Development Frameworks, Data-Centric latent spaces & more.
There will be overlaps between some products and categories. I looked into the functionality of each and every product listed, hence the categories & segmentation of the landscape is a result of that research.
There were a few market shifts and trends since the previous chart:
- There has been a diversification in LLM models available. LLM suppliers like OpenAI, Cohere and others have multiple models and these models all have particular strengths and target implementations.
- Default functionality supplied by LLM providers are expanding, especially in the area of structure and standard built-in playground functionality. These LLM developments in some cases supersedes existing products. Consider here the growing LLM context windows, applications needing to reference multiple models and more. I would argue products higher up in the stack are more vulnerable.
- LLMs are releasing end-user chat interfaces, like HuggingChat, ChatGPT and Coral from Cohere. Again, challenging the differentiation and propriety software and intellectual property of each offering.
- Prompt Engineering techniques are growing and the flexibility and utility of LLMs are really coming to the fore.
- Fine-Tuning of LLMs are still relevant to create industry and market related models, for instance medical, legal, engineering, etc.
- However, for general implementations a popular architecture is one of RAG, where accurate and contextually relevant reference data is included in the prompt to increase LLM response accuracy and negates hallucination.
- Embeddings, Vector Stores and the notion of semantic search and data retrieval is also becoming more important.
⭐️ Follow me on LinkedIn for updates on Large Language Models ⭐️
I’m currently the Chief Evangelist @ HumanFirst. I explore & write about all things at the intersection of AI & language; ranging from LLMs, Chatbots, Voicebots, Development Frameworks, Data-Centric latent spaces & more.
