Once intelligence is too cheap to meter, the world will be unrecognizable (defensesindepth.bio)

The concept of "intelligence too cheap to meter" suggests a future where advanced AI systems can perform tasks at a fraction of the cost of human labor. Current AI models can solve complex coding and mathematics problems cost-effectively, but they still fall short in replicating the full range of human remote work due to reliability issues. As chip production grows exponentially, it's projected that by the mid-2030s, billions of AIs could enter the workforce. This shift could lead to a world where human labor is significantly outpaced by AI output, drastically altering societal structures. For the AI/ML community, this scenario raises critical questions about the implications of widespread AI deployment. If artificial general intelligence (AGI) becomes a reality, the pricing of compute would shift based on the value of human work being automated. During an interim phase, AI labor may compete directly with human wages, keeping AI costly until a sufficient number of models flood the market. The potential consequences range from job displacement to the risk of power being concentrated among elites controlling the AI, underscoring the need for proactive discussions around governance, equity, and the ethical use of AI technology as we edge closer to this transformative reality.