Is anyone using self hosted LLM day to day and training it like a new employee

The author is contemplating the idea of using a self-hosted Language Learning Model (LLM) in their daily routine, training it on their personal emails, notes, and chats. They envision the LLM drafting replies which they can then edit, allowing the model to learn and improve from these interactions. They have already downloaded all the necessary open-source tools and have been experimenting with interactive chat and semantic search functionalities.

The author is optimistic about the progress and maturity of these technologies and expects a clear reference stack to emerge in a few months. They are particularly interested in the assistant stack and are looking to maximize the return on investment from the resources they have put into self-hosting and feeding their data into the system. They are seeking feedback from others who may be using a similar setup.

Key takeaways:

The author is considering using LLM daily and training it on their personal data like emails, notes, and chats.
They plan to have the LLM draft replies which they will edit as needed, allowing the LLM to learn from this process.
They have downloaded all the necessary open source tools and have been experimenting with interactive chat and semantic search.
They are looking forward to a clear reference stack emerging in a few months and are interested in maximizing the ROI of their self-hosted assistant stack.

Is anyone using self hosted LLM day to day and training it like a new employee

Key takeaways:

Comments (0)

Newsletter