Splitwise: Efficient Generative LLM Inference Using Phase Splitting

Commercial Realization

Startups and Open Source tools heavily associated with the concepts explored in this paper.

• GitHub
  drona23/claude-token-efficient
  One CLAUDE.md file. Keeps Claude responses terse. Reduces output ve...

Associated Media Narrative

• OpenClaw signals shift to AI agents and self-evolving models
  Digitimes • Mar 31, 2026
• Universal Claude.md – cut Claude output tokens by 63%
  Github.com • Mar 31, 2026
• Training a Transformer with 1970s-era Technology
  Hackaday • Mar 30, 2026