Anthropic’s big data proposal

Author

Matt
November 26, 2024

ANTHROPIC

Anthropic has proposed a new method of linking data to chatbots. Yes, they need even more data.

Anthropic’s Model Context Protocol (MCP) is a new open standard aimed at helping AI assistants access and use data from different systems more effectively.

MCP allows AI models to connect to tools like business software, content platforms, and app development systems, making it easier to integrate data and improve how AI applications perform.

But… what is MCP?

MCP is a protocol that creates two-way connections between AI applications, like chatbots, and external data systems.

Developers can use “MCP servers” to share data and “MCP clients” (such as apps or workflows) to access that data.

Since it’s open source, any developer can use it to build these connections.

Companies such as Block and Apollo already use MCP, and platforms like Replit, Codeium, and Sourcegraph are adding it to their tools.

Anthropic says MCP simplifies integrations by replacing custom connectors with a single, standard protocol, allowing AI to keep context when switching between tools and data sources.

Here’s what you should know:

  • MCP creates a common way to connect AI to different data sources, reducing the need for custom setups.

  • Companies like Block and Apollo are using MCP, and development tools such as Replit are adding support.

  • Anthropic is encouraging developers to contribute to the project and grow the ecosystem.

And how will it get used?

Anthropic has provided pre-built MCP servers for platforms like Google Drive, Slack, and GitHub.

Developers can start creating MCP connectors now, and Claude Enterprise subscribers can use MCP to connect the Claude chatbot to their internal systems.

Toolkits for rolling out MCP across organisations will also be available soon.

MCP’s adoption isn’t guaranteed. Competitors like OpenAI are developing their own solutions, such as ChatGPT’s "Work with Apps" feature, which links the chatbot to coding tools.

Unlike MCP, OpenAI’s approach focuses on exclusive partnerships and proprietary technology.

Anthropic also claims that MCP improves how AI handles tasks, like understanding coding projects, but hasn’t yet provided evidence to prove these benefits.

For MCP to succeed, it will need widespread use and real-world results to back its potential.

MCP could help AI assistants work better by making data connections easier, but its future depends on whether it gains enough industry support.

MCP could be game-changing… or it could be totally ignored.