
The eternal complement
An OpenAI-hosted essay argues AI should complement scarce execution capacity, not just multiply genius.
7 stories, newest first · one card per story

An OpenAI-hosted essay argues AI should complement scarce execution capacity, not just multiply genius.

Few consumers pay for AI despite model gains, pushing major labs toward enterprise revenue.

Bain says AI must earn about $6 trillion a year by 2031 to justify data-center spending.
ARGUS, a language-model pipeline, audits identification assumptions in climate-policy difference-in-differences studies and abstains when evidence is missing.
Analysis argues LLM inference cost per task is falling by orders of magnitude yearly through GPU, model, and inference-engine efficiency gains.
Real-options model shows firms should pilot AI early and commit late because rapid frontier progress raises experimentation value over irreversible deployment.
OpenAI CFO Sarah Friar describes how compounding gains across chips, compute, models, and products cut intelligence costs.
Today 1.6K readers · 2.5K views · 778 reads · all time 189K views