In this talk, the speaker explains that an AI coding agent consists of a base language model plus a "harness"—a layer of curated evaluations, datasets, plugins, and toolchains that shape the agent's behavior for specific workflows. Current competition among agents like Codex and Cloud Code focuses on benchmark performance (SuiBench, TerminalBench, DeepSui) driven by these harnesses. The speaker highlights the emerging importance of "user harnesses," where individual users customize agents through personal settings (e.g., Claude MD files) and specialized plugins for tasks such as data analytics, sales, or prototype-to‑engineering pipelines. Looking ahead, the real competitive edge will be providing users with portable, adaptable harnesses that work across different agentic platforms, turning harnesses into a commoditized layer while the model itself becomes less of a differentiator.
Chapters
This is really what we mean by harnesses. It could sound really opaque and like, what the fuck is a harness? The harness is the new mode. Oh, or harness as a service. But just real quick clarification for the audience. This is simply it. If you're using a coding agent like Cloud Code or Codex, an agent is simply equal to the model plus a harness. So a model down here in the gray area, an LLM of choice, maybe one of the frontier models, maybe an open source Chinese one. And then the harness on top. The curated evals and datasets that actually power the different user queries, workflows, toolchain calls, all those things. That's this purple area. And then Jonathan and I are thinking like this purple area is really where you see Codex and Cloud Code, they're competing on all these benchmarks, like SuiBench, TerminalBench. The new one is DeepSui for longer running autonomous loops. And then we have this concept of a user harness. So the user steering your own personal Claude MD file, like your global settings. And we're seeing that the Frontier Labs in the latest Codex release, they shipped a bunch of plugins, a bunch of skills as harnesses for different use cases. And those use cases are things like data analytics or sales plugins or how to take a prototype and get it to engineering, like all those different plugins. The real, I think, competition when we zoom out on the timeline in years' time, I think we'll be really at this edge, how do we get the user a harness? And then ideally, and this is where we come in, how do we make that harness portable across different agentic surfaces? I think that's going to be a really cool thing we see. So the take is purple harnesses, that, that's where the competition is now. That's going to be commoditized eventually. We'll get to some level of compute cost quality that feels good for particular workflows. And then above that, how do you use it for your own— the rubber meets the road, this agent meets reality. I think the user harness actually is a huge thing.
