Anthropic’s Claude Code lead says token spend can be cut 50%, but returns may be 1,000x higher
rohanpaul_ai · x · 2026-07-27
Anthropic’s Claude Code lead Boris Cherny says the right way to think about LLM spend is to optimize for return, not token cost.
- He says there is likely room to cut investment by about 50% on token spend.
- But the upside from getting more out of the model could be 1,000x to 100,000x.
- His takeaway: use the most expensive model when needed, and focus on improving the output and workflow rather than obsessing over cost cuts.
- The clip comes from Scale’s YouTube channel.
Related event: Anthropic Lead Urges Optimizing LLM Returns Over Token Costs(2 posts)→
More from coding & agent
- Swarms adds Frenzy Hub, redesigns Marketplace, and ships Gemini 3.6 Flash — KyeGomezB · 2026-07-27
- AI-Trader adds an MCP server so LLMs can run trading backtests — tom_doerr · 2026-07-27
- Sentient OS brings proactive computer-use agents to the Mac with an on-device LLM — TechExpert2910 · 2026-07-27
- Survey maps long-horizon agents into three layers, from in-context reasoning to lifelong experience — theomitsa · 2026-07-27
- Granola plus Lovable turns discovery calls into scoped SaaS proposals in about an hour — PrajwalTomar_ · 2026-07-27
- Swarms teases ZENA v14 with new multi-agent harnesses and an agent communication network — KyeGomezB · 2026-07-27