The thing you’re pointing at probably needs a name. Not the model, not the app — the layer in between. What tools the model sees, in what order, what each tool’s description tells it about when to reach. Where the harness ends and the agent begins.
I keep running into this at work and there’s no good word for it. “Prompt engineering” is way too small now. “Agent design” makes it sound more sci-fi than it is — most of the job is writing tool descriptions and arguing about which five things should be in the default context. It’s plumbing work.
My half-joke bet: harness PM in 2026 is what growth PM was in 2014. Nobody had the title, then suddenly everyone needed one, then it got boring and became part of the job.
Agree that there is no good word for it and it’s hard to know where it lives. The external version of this seems to be converging around “forward deployed engineers” to help customers integrate models. Maybe we need “backward deployed engineers” on our teams? :)
The thing I've realized is that the role isn't really engineering at all, in either direction. It's two very different theories of work: engineering wants closure, customer success wants continuity, and whoever sits in that chair gets graded on both with no name for what they're actually doing. The "engineer" label is doing a lot of work hiding that seam.
I disagree on the platform side. Anthropic tries to compete with everyone that starts making money using their model. That's not a platform play at all. Platforms have very clear rules on what the platform has to own (don't build on top of us for this), what the platform competes on (we'll build in this space too, but we don't prevent you from building something better), what the platform wants people to build that it won't, and what it won't build and doesn't want other people to build. Anthropic hasn't been clear about this at all. Because of how expensive they are to run, if we make money on top of them, they want to take it. That's a terrible developer brand long term.
That dynamic is really common in the early days of platforms. Platforms want to own both the platform itself and the killer apps. We saw this push and pull on Windows and iOS, and a lot of developers complained about lack of white space. It’s not fun to be in the sights of those platforms, but it is possible to build apps in those areas (Spotify <> Apple Music) or break new ground (Instagram).
So, the path isn’t fully paved for developers, but I think Anthropic is doing a good job building the standards to help developers create and own value with their ecosystem.
Yeah, most platforms get plenty of time to flesh out their strategic rules and impart them on developers. Anthropic has grown too fast to even ask all the right questions. But if they aren't careful developers will sway away from Anthropic over time as no one like being free R&D for a platform.
The thing you’re pointing at probably needs a name. Not the model, not the app — the layer in between. What tools the model sees, in what order, what each tool’s description tells it about when to reach. Where the harness ends and the agent begins.
I keep running into this at work and there’s no good word for it. “Prompt engineering” is way too small now. “Agent design” makes it sound more sci-fi than it is — most of the job is writing tool descriptions and arguing about which five things should be in the default context. It’s plumbing work.
My half-joke bet: harness PM in 2026 is what growth PM was in 2014. Nobody had the title, then suddenly everyone needed one, then it got boring and became part of the job.
Agree that there is no good word for it and it’s hard to know where it lives. The external version of this seems to be converging around “forward deployed engineers” to help customers integrate models. Maybe we need “backward deployed engineers” on our teams? :)
The thing I've realized is that the role isn't really engineering at all, in either direction. It's two very different theories of work: engineering wants closure, customer success wants continuity, and whoever sits in that chair gets graded on both with no name for what they're actually doing. The "engineer" label is doing a lot of work hiding that seam.
I disagree on the platform side. Anthropic tries to compete with everyone that starts making money using their model. That's not a platform play at all. Platforms have very clear rules on what the platform has to own (don't build on top of us for this), what the platform competes on (we'll build in this space too, but we don't prevent you from building something better), what the platform wants people to build that it won't, and what it won't build and doesn't want other people to build. Anthropic hasn't been clear about this at all. Because of how expensive they are to run, if we make money on top of them, they want to take it. That's a terrible developer brand long term.
That dynamic is really common in the early days of platforms. Platforms want to own both the platform itself and the killer apps. We saw this push and pull on Windows and iOS, and a lot of developers complained about lack of white space. It’s not fun to be in the sights of those platforms, but it is possible to build apps in those areas (Spotify <> Apple Music) or break new ground (Instagram).
So, the path isn’t fully paved for developers, but I think Anthropic is doing a good job building the standards to help developers create and own value with their ecosystem.
Yeah, most platforms get plenty of time to flesh out their strategic rules and impart them on developers. Anthropic has grown too fast to even ask all the right questions. But if they aren't careful developers will sway away from Anthropic over time as no one like being free R&D for a platform.