Could using MCP (Model Context Protocol) change Patterns for Building LLM-Based AI Agents and make building apps with LLM easier? As I undestand MCP defined a standard to make connections between Language Natural requests to data or services. Over the past few weeks, thousands of connectors have been developed for databases and services like Google Maps and WhatsApp.

Yes I concur with @Eric Bertrand.

But, it will depend upon widespread adoption and longevity. A standard used by a few people is not useful. MCP has promise but it requires the participants to maintain servers to do the translation from their product or service to the standard (MCP).
Universal adoption of a standard is a challenge. For example, for nearly half a century computer process generated logs but there are still multiple, sometimes incompatible, log formats. “Standard” logs are plaintext and computationally cheap to produce. Spinning up servers to translate data streams to the standard is non-trivial.
AI paradigm shifts such as the utility of running custom smaller models hosted on premises vs bolting applications to monolithic models that change based upon their owner’s direction, will also be a standardization factor.

The short answer is yes. Model Context Protocol (MCP) from Anthropic is an emerging standard for connecting applications to an underlying LLM/Model. It's an abstraction layer (Like an API) that standardizes and simplifies the flow of context and data between the application layer and the LLM. This should simplify the development of LLM enabled apps by removing the need for each application to create its own connections to sources of information and other services needed to feed the LLM. Additionally, by using MCP as an abstraction layer it offers the possibility of building apps that allow for the underlying models (LLMs) to be swapped out without having to re-write the application. All of this should create a much simpler framework for the development of agents and LLM enabled applications.

The other emerging integration framework that compliments MCP is Google’s Agent2Agent (A2A) protocol. This is an orchestration framework for coordinating activities between fleets of agents, which will make it simpler to deploy and manage long-run, complex agentic workflows.