The discussion around Grok 3 highlights contrasting approaches to AI development, emphasizing the debate between scaling compute versus investing in talent. Critics argue that while Grok 3 from xAI utilizes a significant amount of compute (100 times more than competitors like Deepseek), this may not yield proportionate returns compared to Deepseek’s approach of prioritizing talent over sheer computational power. The article raises questions about the sustainability of such vast investments in AI compute and the US's potential position as the leading force in AI technology. Many comments critique the notion that increased compute alone will lead to better AI, suggesting that the so-called ‘bitter lesson’ should also include considerations of efficiency and the skill of researchers as vital factors in progress. The tension between computational investment and scalable talent development is apparent, indicating a need for a balanced strategy moving forward.