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.

Companies with valuations reliant on perpetual AI progress are at significant risk if the bubble bursts.

Berkshire Hathaway (BRK.A, BRK.B): The company is extremely decentralized - you’ve got the insurance businesses, logistics operations, energy companies, and others. The subsidiary managers have enormous autonomy, which means the entire organization can capture gains from AI efficiency organically, from a bottom-up rather than top-down approach. Moreover, all of the subsidiaries can greatly benefit from the implementation of the “AI Centaur” model (actuaries working for Geico and logistics managers for BNSF can both use AI tools to improve their productivity). They all involve pouring through numerical and textual data, which AI can help accelerate. I rate BRK.A and BRK.B as strong buys.

seekingalpha.com - Agi bust AI boom why ais winners and losers arent who you think