AI labs’ ability to capture the value they create

A user argues that private passenger rail carries huge upfront costs while nearby landowners, employers and cities benefit from value the railroad creates. The post draws a conditional parallel with AI: if models become commodity infrastructure and apps capture most of the value, labs could face a similar problem after spending billions on models and computing resources. The author makes the case for labs building coding tools, personal agents, enterprise software and consumer apps—likening that approach to Japanese rail companies capturing the land value they create.