Compute and Energy Win AI: A Bettor Doubles Down on xAI

SydSteyerhart · x · 2026-09-26

One poster argues Grok can no longer catch OpenAI or Anthropic, claiming they likely run meaningful recursive self-improvement internally — models writing code, running experiments, and accelerating research for the next model, so the gap is "falling behind a rocket that's still accelerating." The reblogger doubles down on his earlier call that xAI would dominate: in the limit, all that matters is compute and energy; any special sauce diffuses quickly or can simply be bought. AI is the story of capital reaching escape velocity, not one person's clever idea.

Original post →

More from AGI Musings

AGI Musings channel →