The Economic Trajectory of Artificial Intelligence

An AI model's intelligence correlates directly with the computational resources invested in its training and operation—encompassing training compute, data, and inference capacity. Researchers observe that arbitrarily large investments produce predictable and continuous performance gains, with scaling laws that remain consistent across multiple orders of magnitude. The cost to deploy equivalent AI capabilities drops approximately 90% every 12 months, driving substantially broader adoption. Between early 2023 and mid-2024, the price per token fell roughly 150-fold. This trajectory significantly outpaces Moore's Law, which predicted a 2x improvement every 18 months. The socioeconomic value of increasing intelligence grows exponentially, suggesting sustained and accelerating investment in AI. Intelligent agents are being developed as virtual co-workers with potential to transform multiple domains of intellectual labor. The transformation will unfold gradually, but long-term impacts on society and the economy will be substantial. Agency, determination, and adaptability will become especially valuable, while questions of equitable benefit distribution and public regulation remain critical to prevent power concentration and inequality.