AITechTalks6 months ago
LLM tool-use evolves from API calls to agents
1 min read
Tool-use for large language models has progressed from supervised fine-tuning on static data to native tool-use training and standardized protocols like MCP, A2A, and Claude Skills. Early models like Toolformer, Gorilla, and ToolLLM taught LLMs to call APIs. By 2025, agentic applications and specialized frameworks dominate.
Level
Hype check
Tap to vote and see what everyone thinks.
