Hidden behind all of the buzz around ChatGPT is the dependence on the nature and quality of the data that built the models that run it. Peter Licursi and Chris Tanner join host Eric Hanselman to discuss how data drives the model’s capabilities, but can also create unwanted bias and scaling issues driven by the large compute requirements. Is generative AI the single most compelling piece of computational technology that humanity has seen? So much of what it can do comes down to the data.