Building features like custom commands and sub-agents can look like reliable, deterministic workflows. However, because they are built on non-deterministic LLMs, they fail unpredictably. This misleads users into trusting a fragile abstraction and ultimately results in a poor experience.
Simply offering the latest model is no longer a competitive advantage. True value is created in the system built around the model—the system prompts, tools, and overall scaffolding. This 'harness' is what optimizes a model's performance for specific tasks and delivers a superior user experience.
With AI agents automating raw code generation, an engineer's role is evolving beyond pure implementation. To stay valuable, engineers must now cultivate a deep understanding of business context and product taste to know *what* to build and *why*, not just *how*.
Historically, developer tools adapted to a company's codebase. The productivity gains from AI agents are so significant that the dynamic has flipped: for the first time, companies are proactively changing their code, logging, and tooling to be more 'agent-friendly,' rather than the other way around.
Instead of trying to convert skeptics, AMP focuses exclusively on users already at the frontier of AI adoption. They believe that building for someone who doesn't know how to prompt well forces them to build simplistic features and fall behind the pace of innovation.
To avoid disrupting existing enterprise customers and being disrupted themselves, Sourcegraph launched a new brand, AMP. This freed them from Kodi's contracts, customer expectations, and release cycles, enabling a much faster, more radical development pace for their new coding agent.
The AMP team believes that accommodating popular user requests like model choice or 'bring your own key' would slow their ability to innovate. They argue that their target users ultimately prefer a superior, opinionated product over peripheral features, even if they ask for them.
The AMP team is small and agile, bypassing traditional processes like code reviews to ship 15 times a day. They leverage Sourcegraph's customer trust, revenue, and platform teams (e.g., security), allowing the core team to focus purely on product velocity and radical innovation.
The AMP team considered killing its VS Code extension to reduce complexity but realized the ideal interaction model for AI agents is still unknown. Future use cases like async agents managed via web or phone make a GUI necessary, so they maintain both interfaces to avoid prematurely optimizing.