Your Agent Is Wasting Tokens and You Don't Know It
summary
Concrete fixes and numbers: code examples add prompt caching (system and tool prompts) so repeated calls send far fewer tokens; model routing by difficulty uses cheaper models for simple tasks and keeps frontier models for hard ones; large tool outputs get offloaded and summarized and tool loops get capped to avoid runaway iterations. The claims include measurable cost drops without swapping models and examples use Strands agents and simple pseudo code.
AI-written summary · about
chapters (7)
- 0:00Introduction to token costs
- 0:24Caching system prompts
- 1:00Routing by task difficulty
- 1:51Offloading tool results
- 2:47Capping tool loops
- 3:36Trimming conversation history
- 5:37Conclusion and resources
related talks
19mStop babysitting your agents...
35mUsing coding agents with open models
28mThe Hard Truth About AI Agents
31mAI Coding Agents Don't Need Bigger Models. They Need Better Tools.
24mFrom Prompts to AGENTS.md: What Survives Across Thousands of Runs
48mStop Fighting Your AI Engineering Prompts That Actually Work
25mYOUR AI DOESN'T KNOW WHAT THINGS COST
21mDon't Become Your Agent's B****