This week on The AI Report, Liam Lawson is joined by Eric Sui—indie hacker, AI builder, and creator of Agents Playbook—to talk about designing with agents instead of apps.

Eric breaks down how he moved from casual GPT-3 experimentation to building structured AI workflows that solve real problems. They dig into agent UX, why many AI tools fall short, and how indie builders can actually move faster by thinking in systems.

Also in this episode:

• Why agents are more than just automations

• How no-code stacks can launch powerful workflows

• Lessons from “building in public” on Twitter

• How to find product ideas in your own frustration

• The future of indie AI projects and agent-led design

Whether you’re trying to build smarter tools or just want to understand how agents are reshaping what software can do, this episode is packed with insight.

Subscribe to The AI Report:

https://theaireport.beehiiv.com/subscribe

Join the community:

https://www.skool.com/the-ai-report-community/about

Chapters:

(00:00) Intro to Eric Sui

(01:12) Hacking on GPT-3 Before It Was Popular

(03:05) Where AI Products Go Wrong

(04:56) User Experience for Agents

(06:29) Thinking in Systems, Not Features

(08:15) How LLMs Shift the Tool Paradigm

(09:53) Launching With No-Code + GPT

(11:27) Product Validation as Self-Therapy

(13:10) Getting Better Feedback Faster

(14:52) Public Building as a Strategy

(16:47) Twitter Feedback Loops

(19:01) Starting From Zero Audience

(20:45) Workflow Design and Mental Models

(22:39) How to Think With Agents

(24:26) Indiehacking, Investment, and Staying Lean

(26:11) Staying Close to Your Problem Space

(27:53) Teaching Builders Through Agents Playbook

(29:45) What’s Next for Eric

(31:22) Where to Connect

Ecology of undernutrition and infection