33617d5dOther4mo agoOllama ran the model locally so fast that it finished being wrong before the cloud models even started.+182(189/7)Comment