the scaling laws / bitter lesson would disagree I have to note that taking the "bitter lesson" position as a claim that more data will result in better LLMs is

the scaling laws / bitter lesson would disagree I have to note that taking the "bitter lesson" position as a claim that more data will result in better LLMs is a wild misinterpretation (or perhaps a "telephone version) of the original bitter lesson article, which say only that general, scalable algorithms do better than knowledge-carrying, problem-specific algorithms. And the last I heard it was the "scaling hypothesis" that hardly had consensus among those in the field.