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.
Related Insights
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.
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.
Unlike traditional APIs, LLMs are hard to abstract away. Users develop a preference for a specific model's 'personality' and performance (e.g., GPT-4 vs. 3.5), making it difficult for applications to swap out the underlying model without user notice and pushback.
Despite user requests, Supercut is holding back on building a traditional video editor. They believe it would become an "excuse" for their AI-powered "auto edit" to be mediocre. This strategic constraint forces them to perfect their core differentiator before adding table-stakes features.
When products offer too many configurations, it often signals that leaders lack the conviction to make a decision. This fear of being wrong creates a confusing user experience. It's better to ship a simple, opinionated product, learn from being wrong, and then adjust, rather than shipping a convoluted experience.
Saying yes to numerous individual client features creates a 'complexity tax'. This hidden cost manifests as a bloated codebase, increased bugs, and high maintenance overhead, consuming engineering capacity and crippling the ability to innovate on the core product.
The developer abstraction layer is moving up from the model API to the agent. A generic interface for switching models is insufficient because it creates a 'lowest common denominator' product. Real power comes from tightly binding a specific model to an agentic loop with compute and file system access.
Instead of debating individual features, establish a clear "perspective" for your product. Artist's perspective as a "push-based product for quick insights" makes it easy to reject requests that don't align, like building an in-house video hosting tool. This aligns the entire organization and simplifies the roadmap.
Instead of offering a model selector, creating a proprietary, branded model allows a company to chain different specialized models for various sub-tasks (e.g., search, generation). This not only improves overall performance but also provides business independence from the pricing and launch cycles of a single frontier model lab.