video1.0<iframe src="https://www.loom.com/embed/78efd5be4e884f58aae9af0ff0359962" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/78efd5be4e884f58aae9af0ff0359962-9294dd9f803509a3.gif45.699This Loom demonstrates a workflow for training a model on Discord chat history using a bot. The speaker takes all Discord chat files from team members, runs four different training iterations on an open weight model, and then asks questions that reflect what was discussed in the chats. They compare results from using a memory-enhanced approach versus relying on the open model and also against a lower-trained model that provides less context. The main focus is how these different training setups affect the model’s ability to answer based on the team’s prior conversations.