← Browse

Code 2025

Agents are Robots Too: What Self-Driving Taught Me About Building Agents — Jesse Hu, Abundant

Jesse Hu , Abundant

Overview

This talk draws parallels between building self-driving cars and developing AI agents, emphasizing that agents, like robots, operate in complex environments and require more than just sophisticated models. The core thesis is that the 99% of work beyond the AI model—encompassing infrastructure, tooling, feedback loops, and robust offline stacks—is crucial for reliable agent development and deployment.

Who should watch

Key takeaways

Notable quotes

*The model is only doing 1% of the work and 99% of the work goes into other things.*
*The winning team not just having the best model and the best online stack but having the best offline stack because that enables developers to be much faster and ship more much more reliably.*
*Actions have consequences. We're not just dealing with classification models. We're not just dealing with prediction models or sequences. We're dealing with a whole new paradigm in which you predict, you act, and then you deal with the consequences of that action and then re-evaluate everything you've done before.*

Watch on YouTube →

Unofficial community note. Prefer the recording for nuance.