AI token costs dropping orders of magnitude yearly; LLMs to integrate into all computing infrastructure within 1-2 years as GPU efficiency doubles every two years, models get smarter and cheaper, making quality—not quantity

AI token costs dropping orders of magnitude yearly; LLMs to integrate into all computing infrastructure within 1-2 years as GPU efficiency doubles every two years, models get smarter and cheaper, making quality—not quantity—the limiting factor by 2026-2030.

tokens too cheap to meter

The price of using machine learning intelligence is decreasing by several orders of magnitude a year and shows no signs of slowing. We are likely to see LLMs integrated into every part of computing as infrastructure, not just as a product, in the next year or two. We are likely to see LLMs running locally at current frontier-quality on commodity hardware in the next 3-6 years. Starting very soon, we are likely to see quality and access become the limiting factor to AI 1 use, not sheer number of ...