AI Agent Harnesses
topic on 3 shows · 3 statements across 3 episodes
Latent Space 
Lenny's Podcast 
the MAD Podcast
3 statements about AI Agent Harnesses, every show
Schoening: All current AI agent architectures fundamentally resemble coding agents
“All agents are also like, if you look at all the harnesses, whether it's the open source ones or the ones from the model companies ours as well they all resemble a coding agent now.”
Harrison Chase: Modern AI agent harnesses don't strictly enforce execution plans
“And so now what most things do is they just have that plan in the text file, and the main agent can, like, use that to help guide its actions, but there's nothing that says I'm explicitly doing this step, or I'm explicitly doing another step.”
Dax Reed says AI agent harnesses are fundamentally simple tool-execution loops.
“At the end of the day, like it's just a loop where you send the prompt with a bunch of tool descriptions. It tells you what tools to call. You call it tools and the results back. And there's really not too much magic there.”