TensorFlow

Python is a very accessible language, also it's store cleats. TensorFlow will have an interface that parallels the NumPy or SciPy interface. I think if you want to build a production system in TensorFlow, then you do need to be like a hardcore developer. So could we build a neural net that would pick the model for you? That's already done. But actually what we're launching now is AutoML as a moonshot in brain. And the idea is basically exactly that. Can we have a machine learning model that generates a machinelearning model that does well in the task? Chat, they're doing it wrong. They're putting themselves out of a job.