Why MCPs Are About to Change Everything

Hey there, Seb here!

AI isn’t replacing your job—but the people who understand what MCPs actually do might replace your company. While everyone’s arguing about prompt engineering, the real revolution is happening in how AI tools talk to each other. This isn’t just another acronym to learn—it’s the foundation of the next tech gold rush.

MCPs Explained: The Power of Speaking One Language

Think of early AI tools like people who all speak different languages trying to work together. Everything needs a translator, it’s slow, and messages get lost. That’s the current state of AI integration—each tool speaks its own language, requiring custom code to connect them.

MCPs (Machine Communication Protocols) are essentially creating a universal language for AI.

Instead of:

• ChatGPT → Custom Code → Database → Custom Code → Email System → Custom Code → Analytics

You get:

• ChatGPT → MCP → Everything Else

This isn’t just more efficient—it fundamentally changes what’s possible with AI. When systems can seamlessly communicate, you unlock compound capabilities that weren’t feasible before.

MCP is going crazy viral right now

AI apps can now instantly connect to any tool or live data.

USB-C moment for AI.

10 wild examples:pic.twitter.com/qLnMkgBA0H

— Min Choi (@minchoi) March 15, 2025

What You Need to Know About MCPs

1. “MCPs solve the integration problem, not the intelligence problem”
While everyone’s obsessed with more powerful models, MCPs focus on making existing AI actually useful in real-world applications. The next wave of AI value isn’t coming from bigger models—it’s coming from better connected ones.

2. “This is HTTP for AI”
Just as HTTP standardized how web browsers communicate with servers (creating the entire internet economy), MCPs will standardize how AI systems communicate with the world (creating the next wave of tech giants).

3. “The real opportunity is in the picks and shovels”
During a gold rush, the people who sell tools make the most reliable money. The same is happening with MCPs—companies building connectors, middleware, and integration tools will see massive growth before the dust settles.

4. “This turns AI from products into platforms”
The shift from single-purpose AI tools to connected systems means businesses need to think platform-first. The question isn’t “What can this AI do?” but “What ecosystem can we build around this AI?”

AI systems without MCPs are like smartphones without the internet—powerful, but isolated. MCPs are what will turn AI from a calculator into a nervous system.

How Different People Should Use This Information

If you’re a developer: The first wave of MCP standardization is happening now. Start experimenting with OpenAI’s function calling or Anthropic’s Claude tool use—these are early implementations of what will become the standard. Building expertise here now means you’ll be in high demand as adoption increases.

If you’re a business owner: Begin auditing where your workflows require multiple AI tools. These friction points are prime candidates for MCP integration. Don’t rebuild your entire tech stack yet—the standards are still evolving—but start planning for a connected AI infrastructure.

If you’re an investor: Look for startups focusing on the connection layer between AI systems. The companies building middleware, orchestration tools, and industry-specific MCP implementations will likely see outsized returns compared to those building standalone AI applications.

What to Watch: The Next 90 Days

The battle for who controls the MCP standard is heating up. Watch for announcements from OpenAI, Anthropic, and Google in the next quarter—whoever establishes the dominant standard will have enormous influence over the entire ecosystem. Microsoft’s recent repositioning suggests they’re betting heavily on becoming the central hub for MCP implementation.

Most tellingly, three former OpenAI engineers just raised $42M for a startup entirely focused on MCP standardization. When the people who built the technology leave to focus on the connection layer, that’s when you know where the real value lies.

Until next time, let’s keep building!

Seb