AI Agents: Collaboration in Real-Life Applications
summary
Demonstrates concrete agents built on a proprietary platform that use Claude as the LLM and Perplexity for web lookups to do real work: a “Solar Scout” agent runs daily 7am research, saves timestamped research documents, drafts team‑specific emails and posts summaries to Slack (one typo required a fix), and a watchdog blocks unsafe actions until human approval; a multi‑agent “lunch finder” negotiates between three personal agents and books K25 after resolving dietary constraints. Agents work but can loop, hallucinate links less when they fetch real sources, and need careful prompt and safety tuning.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 1:07LLMs and reasoning models
- 7:09Autonomous AI agents
- 9:37Demo: single agent research
- 23:07Human in the loop scaling
- 27:02Multi-agent coordination
- 42:16AI agent adoption
related talks
40mAI Agents in Practice
24mAI Agent Swarms Are Amazing
34mAI Agents: Transforming Autonomous Decision-Making
45mAI agents in practice - beyond the prototype
30mAI Agents Unchained
26mAI agents 101
40mAgents in Action: Advancing Open Source AI Missions
28mThe Put That Sh*t On Agent: Building Multi-Agent AI Systems That Actually Talk to Each Other*