> Dataset is not a big problem (if you can speak, you can create your own) That's simply not true at all. Between the scale of the data and the need to label i

> Dataset is not a big problem (if you can speak, you can create your own) That's simply not true at all. Between the scale of the data and the need to label it, datasets are usually the biggest roadblock in ML.