Figma co-founder Dylan Field thinks MCP is not a perfect protocol

Figma's decision to support and expand its partnership network for MCP, an open standard for AI agent data access, is accompanied by notable reservations from co-founder Dylan Field. Field has pointed out that Anthropic, a co-creator of MCP, does not fully expose its own cloud resources via the protocol, raising questions about the standard's consistent application across its developers.

Field's misgivings stem from what he identifies as "asymmetric context sharing" and a belief that MCP is designed in a way that advantages "frontier labs." This perspective suggests a potential imbalance in how the protocol functions for different types of users or developers. The CEO further elaborated that the company's caution about full openness through MCP is driven by the increased risk of data abuse that comes with exposing more company data.

This scenario highlights a common challenge in the development and adoption of open AI standards: balancing the desire for interoperability with the necessity of data security and fair design. Figma appears to be navigating this by embracing the standard to some extent, as evidenced by the new partnerships, while simultaneously articulating its limitations and the company's own cautious approach to data exposure. This indicates a measured engagement rather than unreserved endorsement.