"The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis Living Lindy: a No-BS Conversation on AI Agents with Flo Crivello
566 snips
Jun 21, 2025 
Flo Crivello, CEO of the AI agent platform Lindy, shares his deep insights on AI agents, stripping away the hype to highlight what's actually effective. He discusses the importance of model selection, fine-tuning, and the role of human oversight in developing AI agents. Flo also sheds light on the evolution of AI capabilities, expresses skepticism about excessive optimism in AI trends, and emphasizes the need for structured workflows. The conversation is a treasure trove of practical advice for AI builders navigating the evolving landscape of intelligent systems.
AI Snips
Chapters
Transcript
Episode notes
Defining AI Agents as LLM Control
- An AI agent is defined as software where part of the control flow is managed by a large language model (LLM).
- Agency varies on a spectrum, depending on how much control the LLM has over decisions within the system.
Balancing Open-Ended and Structured Agents
- The Lindy platform allows users to set guardrails for agents, balancing open-ended behavior and deterministic workflows.
- Some agents operate freely within defined rules, while others follow strict, stepwise processes to ensure reliability.
Use Deterministic Tools with Agents
- Keep tools used by agents deterministic, avoiding agentic tools to simplify system reasoning and operation.
- Using sub-agents as tools complicates interactions and reduces overall reliability.