The market is overvaluing AGI potential and undervaluing companies that can drive efficiency by adopting current AI tools. True AGI is unattainable with current computing architectures; LLM progress is likely to plateau, leading to diminishing returns on AI infrastructure investment. I see strong buy opportunities in BRK.A/BRK.B and UNH, and buy opportunities in TRI, SPGI, and MCO, as these firms can leverage AI for productivity gains.