OpenAI: A business that scales with the value of intelligence

The 'scaling with intelligence' argument often ignores the 'scaling of verification costs'. As models get smarter and cheaper, the cost of generating complex code or artifacts approaches zero. However, the cost of verifying that output (especially in high-stakes enterprise environments) remains non-zero. We might see a value shift where the premium isn't on the 'Intelligence' that generates the work, but on the 'Reliability' systems that validate it. In a world of infinite cheap tokens, trust becomes the scarce asset.