Chamath argues xAI's massive compute-first bet (Colossus) and adoption of the 'Bitter Lesson' approach — general computation over human-labeled data — has put Grok 4 decisively ahead of all rivals in under 2.5 years, with compounding advantages as successive models rely entirely on synthetic data.
Starting in March of 2023, so we're talking about less than 2.5 years, what this team has accomplished and how far ahead they are of everybody else, as demonstrated by this.”20:50