> majority of researchers in the field who have been researching for decades already know deep learning isn't the correct way forward

Do you have a reference for this? It’s not my experience.