Bias and assumption are needed features of non-deterministic thought. We live in an infinite universe so thinking and exploring until you have truly 'got it rig

Bias and assumption are needed features of non-deterministic thought. We live in an infinite universe so thinking and exploring until you have truly 'got it right' would lead to death. I have built non-deterministic NNs (which are exceptionally effective by the way) and the key is the loss function and training for learning when to predict that the loop should exit. I really like this article because it shows someone who keeps pushing their 'when to exit loop' training and the positive and negative implications of that.