"The most powerful tool we have as developers is abstraction."
Gino Ferrand, writing today from Lima, Peru 🇵🇪
This month, something quiet but profound happened:
Microsoft and Google both announced support for a new standard called MCP... the Model Context Protocol.
If you haven’t heard of it yet, you’re not alone. But you’re about to.
MCP is being called the "USB-C of AI applications"... a universal interface that lets AI agents connect to tools, databases, and services just like apps connect via APIs.
But instead of just fetching data, these AI agents can now operate like intelligent intermediaries: checking your logs, querying Jira, calling functions, testing outputs, and stitching it all together into an actual workflow.
So what is MCP, really?
Originally proposed by Anthropic and now backed by Microsoft, Google, and open-source communities, MCP standardizes how AI models interact with external systems.
In non-marketing terms: It gives AIs a common language to call tools and access context.
Think of it like giving every AI a backpack full of adapters. Suddenly, they can:
Auto-fetch ticket context from Jira
Scan docs or codebases for relevant info
Call APIs to act on your behalf
Chain tools together like a mini-agentic workflow
And soon, as more tools support MCP, your agent won’t just "answer prompts"... it’ll operate like a junior developer who actually knows where to look, what to fetch, and how to act.
AI-Enabled Nearshore Engineers: The Ultimate Competitive Edge
The future of software engineering isn’t just AI... it’s AI-powered teams. By combining AI-driven productivity with top-tier remote nearshore engineers, companies unlock exponential efficiency at a 40-60% lower cost, all while collaborating in the same time zone.
✅ AI supercharges senior engineers—faster development, fewer hires needed
✅ Nearshore talent = same time zones—real-time collaboration, no delays
✅ Elite engineering at significant savings—scale smarter, faster, better
Why it matters for dev teams
Right now, AI assistants are mostly limited by the sandbox they live in. Your IDE assistant knows your code file. Your chat assistant might guess your intent. But neither talks to your build system, CI logs, analytics, or docs.
MCP fixes that.
It makes it trivially possible for an AI agent to:
Investigate a flaky test by checking GitHub + CI logs
Update documentation after reading your API diff
Help a PM draft a feature spec using context from product analytics
With MCP, we don’t just get "smarter models."
We get context-rich agents that can operate across the software lifecycle.
What comes next
Microsoft is baking MCP directly into Windows 11. Google is rolling it into Vertex AI. Devs are already talking about writing "MCP skills" that work across vendors. Some are calling for standardizing internal tools (Jira, CI, docs) to be MCP-accessible.
The long-term implications?
Your agent can become a first-class developer tool.
DevOps, QA, and PM work can be partially automated by shared agents.
AI features become plug-and-play across your stack.
And if you’re building software tools? Supporting MCP might become table stakes.
Bottom line: MCP isn’t just another protocol. It’s an abstraction layer for the AI-powered SDLC.
We’ll be tracking this one closely.
More to come…
Recommended Reads
✔️ Figma will let your AI access its design servers (The Verge)
✔️ MCP Explained : The New Framework Transforming AI Capabilities (Geeky Gadgets)
– Gino Ferrand, Founder @ TECLA