#thecompaniesapi train instruct model, it reaches better loss than the thinking one! ship lots of refactoring on our bot preparing a potential performance overhaul of our extraction flow; fix a customer support request on `v1` of our API
#thecompaniesapi optimize our extraction flow code; prepare textsynth replacement by ollama by converting our existing model to it; start fine tuning new model to support thinking
#thecompaniesapi finish making one example workflow for each operation to close publish tasks for zapier integration; need to fix create operations return