Hey! Ankur here, and this is the 18th edition of Lazy AI — 5 mins of reading to help you stay ahead of the AI curve.
You use AI at work. But somewhere you’ve accepted that you still have to do all the copy-pasting yourself. Today’s issue is about why — and what’s already changing it.
The invisible tax nobody talks about
Every time you use AI at work, there’s a ritual you’ve probably stopped noticing.
Open a chat. Copy something from your email. Paste it in. Give a prompt. Get a response. Go back. Copy something from your notes. Paste that in too. Repeat until you have what you need.
The AI is helpful. But you’re doing a large amount of work just feeding it. You’ve become the bridge between the AI and every other tool you use.
And that’s not an intelligence problem. It’s a connectivity problem.
There are two reasons this happens.
AI models have a training cutoff — a fixed date beyond which they don’t know what’s happened. They usually can’t pull live information; they only know what they were trained on.
Even for information that exists right now — your inbox, your calendar, your company’s systems — the AI has no way to reach it. Nobody built the bridge.
So YOU became the bridge.
But MCP is what finally builds it.
What MCP actually is
MCP stands for Model Context Protocol. Anthropic created it in November 2024 and immediately open-sourced it. So any company could adopt it for free. And the uptake has been striking.
MCP is a shared language that lets AI models connect securely to external tools and live data — your inbox, your calendar, your CRM, your company’s internal systems — in a standardised way. Instead of you ferrying information in and out, the AI reaches out, gets what it needs, and returns something actually useful.
Think of it like Swiggy. Before it existed, every restaurant had its own ordering system — call this one, use that app, visit this website. Every combination was different work.
Swiggy built one standard layer. Restaurants plug in once. Customers order from anywhere, the same way, every time.
MCP is that standard layer — but for AI and software tools. Every tool plugs in once. Every AI that speaks MCP can use it. No custom wiring for every combination.
Every MCP connection has two sides.
The MCP Server is built by the tool that wants to connect — Gmail, Google Calendar, your company’s CRM, whatever. It announces what it can do and controls exactly what the AI is allowed to access.
The MCP Client is the AI itself. It connects to the server, understands what’s available, and uses it in real time — based on what you need in that specific conversation. The AI doesn’t get a master key. It gets a specific, deliberate set of permissions — defined by the tool, not the AI.
You don’t need to be an engineer to talk AI. Subscribe to this newsletter — every other day, you’ll understand AI well enough to be the smartest person in the room at work.
Each issue is just 5 minutes — less than the time you spend doomscrolling before bed. Except, this actually moves your career forward. Join 8,000+ subscribers now!
Who’s actually building on it
Here’s where it gets interesting — because it’s not a niche experiment anymore.
In March 2025, OpenAI officially adopted MCP across its products. Sam Altman said: “People love MCP and we are excited to add support across our products.” Google DeepMind confirmed MCP support in Gemini models in April 2025.
It’s not just the big AI labs. Early enterprise adopters include Block, Apollo, Replit, Codeium, and Sourcegraph — using MCP to let their AI agents access live data and execute real functions.
So what does this mean for you, right now?
If you’re not a developer, you won’t go set up MCP yourself, don’t worry. But here’s the one thing I’d want you to take from this: When you’re evaluating or recommending AI tools at work, ask whether the tool connects to your existing systems — or whether it expects you to keep feeding it manually.
That question — Does this AI work with my tools, or do I work for it? — is exactly what MCP is designed to answer. The tools that support it will feel categorically different from the ones that don’t. Not because they’re smarter. Because they’re finally connected.
The copy-pasting won’t last much longer. The bridge is being built.
If this got you thinking, please share it with a colleague who still spends their mornings copy-pasting between tools and an AI window. There’s a better way coming — and they should know about it.
And if you liked this, Subscribe.
See you next time..
Cheers,
Ankur