"The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis Always Bet on the Models: How Tasklet Puts the Agency in Agents, with CEO Andrew Lee
172 snips
Oct 22, 2025 
Andrew Lee, founder and CEO of Shortwave and Tasklet.ai, shares his insights on revolutionizing AI automation with an agent-first approach. He discusses how Tasklet enables flexible, conversational agents that replace traditional workflow tools. Lee emphasizes the need for speed in the AI landscape, where rapid model advancement is the only sustainable advantage. He delves into the challenges of managing costs and reliability while scaling integrations and highlights Tasklet's vision of virtual employees as long-lived business teammates.
AI Snips
Chapters
Transcript
Episode notes
Models Should Drive Automation
- Betting on the model means putting the LLM in charge of high-level decisions while code executes tool calls.
- Inverting the stack (LLMs wrap software) lets agents handle unexpected errors and nuance better.
Integration Work Is Core Product
- Building wide integrations, triggers, and team features forms the non-LLM plumbing that makes agents useful.
- Those three categories—connections, triggers, and collaboration—are core product work beyond the core agent loop.
Prioritize Fast Model Rollouts
- Release new model versions quickly even without perfect evals to meet customer expectations.
- Use experiments, internal rollouts, and opt-in features to mitigate risk while moving fast.