Intro to AI: What are LLMs, AI Agents & MCPs?
AI, AI, AI — everyone is talking about it these days! And even though it may be overhyped, there are some excellent ways that you, as a...
Karan Datwani
Share:
AI, AI, AI — everyone is talking about it these days! And even though it may be overhyped, there are some excellent ways that you, as a...
AI, AI, AI — everyone is talking about it these days! And even though it may be overhyped, there are some excellent ways that you, as a developer can use AI right now to get your job done faster and better. But first, we need to understand the basics of AI. If you have only heard these terms, but never understood them - this article is for you! Most important concepts in AI in 2025, broken down in simple terms. Let's go!
AI, or Artificial Intelligence, is all about making machines behave like humans. Giving a brain to a computer — that's AI 🧠💻
Think of AI like teaching a child. Just like we teach a child, "if this happens, do that". In the same way, we train machines using data. Based on that training, the machine learns how to respond to situations, builds its own logic and makes decisions.
Okay, but how does it differ from an algorithm?
An algorithm gives the same result every time (deterministic results), but an AI might give a different result every time (non-deterministic).
Algorithms are coded by hand, and you have complete visibility and control in how it works. AI, on the other hand, you "teach", but often have NO FUCKING IDEA how it works most of the time. But hey... as long as it works, right? RIGHT?!?!?
Now, after understanding AI, we know it's like a virtual brain helping us with tasks. But how does it understand our language? How can it chat back like a real person, write code, answer questions, and even crack jokes? 😄
Well, that's where LLMs come in. LLM stands for Large Language Models. It's a type of AI trained on billions of lines of text — books, websites, coding docs, conversations, and more.
So, when we ask an LLM something, like:
"Write a PHP program to sort numbers," It doesn't pull up a saved answer. Instead, it reads the sentence, understands the context, and generates the code, just like a smart coder. 💻🤓
But who tells the LLM what to do? That’s us! We type, ask, and guide — and then the LLM responds. Talking to gen-AI can take a lot of time, constantly giving instructions to reach the goal can get exhausting. So, Can't we have a different AI to keep giving instructions? That’s exactly where AI Agents come in. 💡
An AI Agent is a smart system that continuously communicates with an LLM, thinking, planning, and refining — until it finds the best solution. It's like an assistant that doesn't need you to give instructions over and over again. Think of it as your personal Jarvis from Iron Man, working behind the scenes to accomplish tasks. That's the ideal vision every AI agent is striving for!
But how does it interact with the outside world, like your computer, emails, databases, Photoshop, or other apps? That's where MCP (Model Context Protocol) comes in. 🌉
Think of MCP as a bridge between the AI Agent and the real world. It enables the agent to take action beyond just generating text. With MCP, an AI Agent to do things like:
Without MCP, the AI Agent would be stuck, unable to interact with everything around it. But with MCP, it becomes a fully capable assistant that can handle tasks like a pro!
Want to see it in action? I found several real examples in this X post — check it out to watch MCPs at work. 🚀
As developers, we spend a lot of time writing repetitive code, fixing bugs, searching for syntax, testing small features, debugging errors. All of this takes valuable time. AI can act like our coding partner. We can integrate AI directly into our IDE (code editor), so it can:
These tools are changing the way developers work. Instead of wasting time on repetitive tasks, we can focus on building cool projects. AI can actually:
The result? You move way 10x faster, no doubt! Want to see it in action? Check out the following video from the VS code team sharing demo of it's Agent mode doing things for you:
So, this all sounds great — but what’s actually out there, right now? Here are some AI-powered tools and IDEs already changing the way developers work:
These tools aren't just assistants — they’re becoming your second brain at the keyboard. People are using them daily to ship faster, debug smarter, and focus on what actually matters.
LLMs, AI Agents, MCPs — they’re not just buzzwords. They’re tools that help you get more done, with less grind. From LLMs generating text to AI Agents automating tasks, the possibilities are endless.
As a developer, AI can save you time, reduce repetitive work, and help you write cleaner, better code. Imagine having an AI that's always by your side, suggesting improvements, fixing bugs, and writing documentation — sounds like the ultimate coding partner 😎
Now, the only question is — are you in? Let us know if you want practical advice on how to use AI in coding. We've been experimenting A LOT and are excited to share what we’ve learned!
Subscribe to our "Article Digest". We'll send you a list of the new articles, every week, month or quarter - your choice.
What do you think about this?
Wondering what our community has been up to?
