1 clip
The Twenty Minute VC (20VC)
OpenAI's Head of Codex discusses how GPT 5.3 Codex has become significantly faster through model improvements and infrastructure optimizations, achieving 40% speed gains in API and 25% in Codex. The clip explores a provocative thesis that inference costs are replacing traditional sales and marketing spend, creating a new form of product-led growth where AI handles user onboarding and value demonstration.