A prominent AI researcher argues, 'The relentless pursuit of larger models, guided by predictable performance curves, is a form of intellectual stagnation. Instead of investing billions in compute to simply move along a known trajectory, we should prioritize funding for novel, more efficient architectures, even if their potential is less certain.' Critically evaluate this statement. In your response, discuss the strengths and weaknesses of the researcher's argument, considering the role that predictable performance improvements have played in guiding and advancing the field.