AI Has Entered The Chat: Humans, AI, and the Art of Collaboration
summary
Describes a working Slack prototype called frankbott built in about a week that connects GPT-3/3.5/4 style LLMs to a local vector DB (switched from Pinecone to Chroma DB) for cosine similarity retrieval and a simple weighted memory store scoring recency importance and relevance. Uses LangChain/AutoGPT/BabyAGI patterns, a memory + reflection agent and separate privileged/quarantined LLMs; demos run live and reveal multi-agent coordination limits and latency tradeoffs.
AI-written summary · about
chapters (12)
- 0:00AI collaboration intro
- 1:47Why talk about AI agents now
- 5:37Centaur chess and teamwork
- 8:24AI advances and hype
- 9:29Working with LLMs
- 12:40Stanford’s generative agents
- 17:04Vector databases for memory
- 19:32AI ethics and trolley problem
- 23:24Key takeaways for AI systems
- 27:20Q&A with Frank and FrankBot
- 37:43Building FrankBot and future
- 42:23Closing and interaction
related talks
55mAI in Java? Meet LangChain4j and Build Intelligent Services
24mAI as a Teammate: Redefining Agile for Human–AI Collaboration
01h 12mBuild Your Own AI Chatbot with OpenAI & LangChain
38mSponsored Session: Jupyter AI v3: Human + AI Collaboration
33mCollaborative Intelligence: The Human & AI Partnership
25mLets Build A Node.js ChatGPT-Like Bot With AI APIs
01h 23mBuild your own AI Agent
46mAI Integration for Devs: ChatGPT and Copilot in Your Projects