Curated by

M

Mike Boscia

stacklist.com/michael-boscia-871

More in AI inbox

See all 60 →

More from Mike Boscia

See all stacks →

hoeem on X: Efficient Token Use in AI Models

A tweet by Hooeem sharing two cost-effective prompts for planning AI tasks before using expensive models like Claude Code or Codex. The first prompt (the Karpathy prompt) instructs an AI to act as a world-class expert with detailed, accurate responses while avoiding disclaimers and political correctness.

View card
Built for AI agentsACO · 1053 tokens

Summary

A tweet by Hooeem sharing two cost-effective prompts for planning AI tasks before using expensive models like Claude Code or Codex. The first prompt (the Karpathy prompt) instructs an AI to act as a world-class expert with detailed, accurate responses while avoiding disclaimers and political correctness.

Tags

prompt-engineering · ai-optimization · token-efficiency · claude · cost-reduction · llm-strategy

Key entities

Hooeem (person, 0.95) · Karpathy (person, 0.85) · Claude Code (technology, 0.9) · Codex (technology, 0.9) · X (technology, 0.85) · prompt-engineering (concept, 0.92) · token-optimization (concept, 0.88)

Classification

snippet · language en · status final

Provenance

claude-haiku-4-5 via @stacklist/be@0.1.0, confidence 0.85, 11 Jun 2026