This is a general critique of the entire field of machine learning and non-casual analysis, not just GPT-3. I like it - it’s important to keep in mind - but we

This is a general critique of the entire field of machine learning and non-casual analysis, not just GPT-3. I like it - it’s important to keep in mind - but we’re never getting to the heart of “it doesn’t truly understand context” unless we literally start again from scratch: forget NNs and do something new.