x.comthreadDhairya Karekar (@dkare1009) on X
An X thread summarizing a 1-hour Google course on agentic engineering, covering building a first AI agent, agent memory (short-term, persistent, long-term), agentic loops for long-running agents, MCP vs API for tool use, and multi-agent systems. The thread's value is framing these as one integrated system rather than separate tricks, which the poster argues is where most agent-building guidance falls short.
TL;DR X thread pointing to a Google course teaching agent memory, loops, and MCP together as one system.
Organisations named here
- Dhairya Karekar
Technologies named here
- MCP
- Model Context Protocol
Techniques named here
- Agent memory
- Agentic loops
- MCP vs API
- Multi-agentic systems
Related in agents
