DEV Community: Chirag Joshi

November — A Month of Energy and Building | Chirag Joshi

Chirag Joshi — Fri, 06 Dec 2024 10:23:49 +0000

2024 was eager to bid farewell, but my energy and drive were at their peak. The passion for building and learning was so intense that day and night blended together. November was meant to be special, and it truly turned out to be.

The journey began with Mumbai Hacks’24. Reaching the final round of the hackathon felt almost unreal — yet it was happening. Standing among hundreds of tech enthusiasts and presenting my idea was an experience beyond words.

Late nights, countless experiments, and endless energy led to something extraordinary — a tool that transformed educational PDFs into Gujarati podcasts. When I first heard the PDF speaking in Gujarati, my happiness had no bounds.

Then came the most thrilling part of the month — my trip to Delhi. Attending the Google For Startups AI Academy 2024 for three days at IIT Delhi was nothing short of a revelation. Walking through the IIT campus, learning new AI concepts, and diving deep into discussions with fellow tech enthusiasts was an unforgettable experience.

One evening, as I stood by India Gate, gazing at the Rashtrapati Bhavan, I felt as though I was standing at the heart of India — the place from where the entire nation operates. And when I saw the grandeur of the Qutub Minar, I couldn’t help but think: our ancestors laid stone upon stone to build such magnificent structures, and today, we’re building the future, line by line, in code.

This month has been a journey of learning and growth. Now, I’m ready for December, eager to create, learn, and move forward with new things.

How Does AI Work? 🤔🤔

Chirag Joshi — Tue, 19 Nov 2024 05:57:36 +0000

We hear a lot about AI, and when we ask ChatGPT something, it gives amazing answers. But how does it all happen? How does AI work?

I’m sure many people have had these questions in their minds. Today, let's understand the answer in the simplest way possible. 😊

First, it’s important to know that even though ChatGPT was released just 8 months ago and made AI famous, AI has been around us in some form for many years.

For example, when you open YouTube or Instagram and see videos or photos you like, that’s AI working in the background. Although it's much weaker compared to today’s ChatGPT or other AI, it was still AI.

Now, to understand AI easily, we need to know how our brain works and how we’ve learned things from childhood till now.

When a baby is born, they know nothing about the world—how to talk, what people look like, or anything about their surroundings. (Sure, some basic info is already in their DNA, but not much.)

As the baby grows, they hear people talking and learn how to speak, see things, recognize people, and understand their surroundings. During this time, the brain receives all this information through the eyes, ears, nose, or touch in the form of language, visuals, sounds, smells, or touch. And the brain learns and understands from this information. (So basically, we become what we are today just because of the information our brain has collected! 😄)

The AI behind ChatGPT is like a simple copy of how our brain works. These models are called Large Language Models.

They’re called “language models” because they work with language. You give instructions in a language, and the response also comes in a language.

And they’re called “large” because they’re trained on an enormous amount of data—billions of text data across the entire internet. Training means they process this data and learn from it.

AI models are trained on powerful GPUs. Just like how computers have CPUs, GPUs are special high-capacity chips needed for AI. (Also useful for gaming! 🤪)

The AI behind ChatGPT was trained on 10,000 of the most advanced GPUs, which took months. That’s why it has such incredible capabilities! 😇

This is why AI can do many tasks much better and faster than us, although there are still some areas where it's weak.

Now you might wonder—how does the model learn from data?

In reality, AI uses artificial neural networks, which are simple mathematical copies of how our brain works.

As you probably know, our brain and body process information through neurons. Our brain has 86 billion neurons that are all connected in a network and constantly exchange information. Imagine how complex and vast this network is inside our heads! 😊 This is called a neural network.

In AI, we have “artificial neural networks,” and we train them with data. At their core, these networks work by predicting the next word in a sentence. So, the answers you get from ChatGPT are just predictions of what the next word might be based on the data it’s trained on. 😊

This artificial neural network isn’t made of plastic or human tissue 😄. It’s purely mathematical. Yes, math is at the core of all the hype around AI. (I never realized how important math was until I saw its power in AI—otherwise, I just wanted to score higher! 😄😄)

Now, you probably have an idea of how this “AI” thing works and why people fear it can become incredibly powerful! 😊😊

AI: What Will the Future Look Like!

Chirag Joshi — Tue, 19 Nov 2024 05:55:58 +0000

I've often said that just within the span of around a hundred thousand years, we've gone from roaming the African jungles like apes to now creating artificial intelligence even stronger than ourselves. This could be seen as the most incredible revolution in Earth’s 4.5 billion-year history. (So even with our flaws, we humans are no less extraordinary. 😄)

Friend, you may have heard a little about AI, maybe even used ChatGPT, but it might still not feel like this AI will bring any major changes to your life, as some people claim. (You might think, "Oh come on, stop going on about AI, all this AI talk has started to stress us out for no reason. 😅😅")

To understand why AI will greatly impact every life and humanity as a whole, we need to know how it’s developed so far and where it's heading.

It wasn’t just that ChatGPT came out eight months ago and suddenly AI was here. We've actually been surrounded by AI in various forms for about a decade now.

When you open YouTube, you’re immediately shown videos that match your interests. On Instagram or Facebook, posts from people you're most interested in show up more. When you open Netflix, it suggests movies you’re likely to enjoy.

All of this is AI in some form, but it’s very basic AI. It simply identifies patterns in our data and gives us suggestions based on that, but it doesn’t actually "think" for itself.

Then, eight months ago, ChatGPT arrived. You could ask it anything, and it gave interesting answers—sometimes it even seemed like it was thinking like a human. It was so impressive that within just two months, it had reached 100 million users. For the first time, we saw an AI that could understand, think, and respond like a person.

ChatGPT is a language-understanding AI model, but there are many other models designed to work in different areas with varying strengths. For instance, Claude is another model similar to ChatGPT, known for its accuracy. And beyond language, there are also models that work with images and videos. A model named MidJourney, for example, has become popular for creating high-quality images.

Even with all this power, today's AI is still considered weak because these models aren’t capable of thinking completely like us across a range of tasks. For example, they often struggle with complex math problems, and sometimes they even generate incorrect facts or stories.

Despite these weaknesses, today’s AI is quite useful and powerful. Don’t think it’s just good for chatbots or making pictures; these models can be adapted to our needs, which is known in technical terms as "fine-tuning." After fine-tuning, AI can:

Keep track of all accounting tasks,
Respond to messages and emails,
Generate work reports,
Look through thousands of reports and summarize the essential information,
Write new content,
Design social media posts and even publish them.

So, the AI tools you see around you are powered by such strong models, customized for specific tasks.

As AI grows more capable, it will eventually evolve into what’s known as AGI (Artificial General Intelligence).

AGI, as the name suggests, will be like us in terms of capability. It will be able to handle general tasks just as we do or at least help us greatly in doing them. It will be able to do math, teach, and even fill out forms in seconds. 🤪

Now, think about all the tasks you do daily that require thinking but not much physical effort—accounting, diagnosing and advising patients, teaching students, managing tasks, supporting customers in call centers, coding, designing, researching, and so on. Probably most people’s jobs fall into this list.

Now, think about how much time, salary, and energy it takes for us to do these tasks (energy here is a good term for the headache of managing people 😄😄).

No matter how much we work, our capacity has a limit. But AGI would be able to do all this in minutes or help you do it. Imagine how much more we could accomplish if we could do so much more work in less time. (Think of teachers not having to worry about tasks like creating and checking papers, uploading grades, taking attendance, and doing all sorts of administrative work—if AI could handle all of that, wouldn't that be amazing? 😃😃)

AGI is thus the biggest promise for improving everyone’s life. It’s estimated that by the time we reach 2030, our GDP could grow by 30–50% each year. 🥳

Then comes the question—will AGI, which learns and improves itself, just stop at some point? We learn slowly, but AGI will learn millions of times faster. At this rate, by 2030, AGI could become a common sight. (If you’re studying to become a doctor now, think about what it’ll be like in 8 years when you graduate... 🤪)

As AGI advances rapidly, it will eventually become so much more powerful than us that we’ll have what’s called ASI—Artificial Super Intelligence. It’s believed that ASI could be more powerful than all human brains combined.

Beyond this, we can’t say what might happen.

(No need to be scared—we are brave, aren’t we? 😄 But also, don’t assume that something we create can’t become stronger than us. 😉)

Yes, powerful AI has the potential to solve our most complex problems today, such as climate change, low-quality education, diseases like cancer, universal healthcare, poverty, and more. For 8 billion people to lead excellent lives, AI might be the only path forward.

(Now, read that first line I wrote again. It’s true—we humans are no less amazing. 😅😅)

AI - The Beginning of a New World (2)

Chirag Joshi — Tue, 19 Nov 2024 05:53:26 +0000

For a while now, you’ve probably heard the term "AI" being mentioned everywhere, but it's not always clear what it actually is, right? And why is there so much buzz about it?

AI (Artificial Intelligence), as the name suggests, simply means artificial intelligence. Yes, it’s the intelligence that we humans have had control over for thousands of years. 😄 (Though, moving forward, that might not be the case.)

AI refers to a machine that can think, learn, and understand like humans. This means it can perform tasks that typically require human brainpower. (Here, machine refers to computers—not factory or household machines. 😊)

Up until about 70–80 years ago, the machines we built were primarily designed to assist us with physical tasks, whether in factories or cars.

But the invention of computers and calculators was the first time we developed a machine that could assist with mental tasks. Early computers were programmed to handle basic calculations, and they followed simple instructions to complete these tasks. Gradually, we began programming computers to handle more complex calculations and tasks using more sophisticated instructions.

This pattern continued: programming computers with more advanced instructions to carry out increasingly complex and challenging tasks. Most mobile apps and software today still follow this approach, such as WhatsApp, accounting software, phone cameras, Zoom, alarm clocks, YouTube, etc. (Although platforms like YouTube, Facebook, and Google also incorporate a bit of AI, they mostly rely on predefined instructions.)

But ultimately, computers were limited to following the instructions we gave them. They couldn’t think for themselves, and only followed what was programmed.

(When I say "computer," I’m not just referring to PCs. This applies to all digital devices around us, especially mobile phones. 😊)

Now, imagine if a computer or phone could function without being programmed or given specific instructions. What if we could just tell it what to do, and it would understand and act on its own?

That would be pretty close to how a human operates, right? (Though, let’s be honest—some humans struggle to work without instructions too! 🤪)

Well, guess what? We’ve built machines that can think, reason, understand, analyze, and learn just like the human brain. This is what we call AI.

AI is essentially software or computer programs that can think like a smart person. They can respond to questions, write content, create photos and videos, have discussions, and handle any task that requires thinking or creativity.

And that’s why I keep saying AI is not just some new tool or technology—it’s the most powerful technology in human history, with the potential to change every aspect of our lives.

In AI language, these programs are called models. They’re much more powerful than any software we’ve had before.

Yes, today’s AI models have limitations. In some areas, they’re weaker than us humans. But in other areas, they’re faster, more efficient, and way more creative than we are. Just think of current AI as the starting point. (You can compare it to smartphones from 12 years ago.) If AI is already this capable, imagine how powerful and useful it will become as it continues to evolve!!! 😇

Now, you might wonder—how can a machine think? Our brain is such a complex mystery that even scientists haven't fully figured it out yet, so how are machines thinking?

I’ll answer this in the next post. It’s the result of 70 years of hard work, research, and investment. 😊

AI - The Beginning of a New World

Chirag Joshi — Tue, 19 Nov 2024 05:50:41 +0000

AI: The Beginning of a New World!!

Over the past 12 months, the entire world has been buzzing about AI, but why has this happened? And what are the real reasons behind it?

Many people might wonder why all this buzz? Why does everyone need to understand or adopt AI? 🤔

Before we dive into what AI is and how it works, let’s talk about why it’s crucial for everyone to understand AI. 😇 In one sentence: “AI is the most powerful, influential, and dangerous technology/discovery in the entire history of humanity so far.” 😎 Finally, we have created artificial intelligence!

(When we were young, we used to hear elders say that no matter how hard scientists try, it’s impossible to create a brain and blood — that’s something only God can do. 😊) The idea of machines thinking for themselves has been part of our stories and imagination for many years. But creating artificial intelligence was really challenging and uncertain. After 70 years of continuous research, we have now reached this point.

Yes, today’s AI has many flaws and limitations, but this is just the beginning. There are two important points to understand here:

AI can process information millions of times faster than our brains, especially in learning at this speed and improving itself. So, the knowledge and capabilities our brain has gained through millions of years of evolution can be achieved by AI in much less time. Yes, it’s still weak in many tasks, but it’s learning at an extraordinary pace.

Secondly, while it took billions of years to create 8 billion human brains, it takes almost no time to create billions of copies of our artificial intelligence. Now think carefully about how many tasks around us involve active or passive use of our brains? Almost all of them! And that’s why it’s certain that AI’s impact will be widespread in all aspects of our lives. I will write a simple explanation that anyone can understand about AI in my upcoming articles. 😊

(The sole purpose of starting to write articles on AI is to increase general understanding of AI and explore the possibilities that arise from its proper use.) Moreover, it’s essential to steer AI in the right direction in every field. We are actively working on developing AI models in the education sector and steering them in the right direction. We strongly believe that more and more people need to contribute their thoughts and opinions for this purpose. 🙏🙏

*This article was originally written in Gujarati and has been translated into English to reach a wider audience. You can read the original Gujarati version here:
Gujarati Blog

🚀 Introducing README GPT 🤖 : for Effortless README Generation! 🚀

Chirag Joshi — Tue, 14 May 2024 16:44:22 +0000

Are you tired of spending countless hours crafting README.md files for your projects? Say goodbye to manual documentation struggles with README GPT 🤖 - a revolutionary solution that automates the process for you!

In this article, we'll explore how README GPT 🤖 leverages the power of generative AI to streamline README.md file creation, saving you time and effort while ensuring comprehensive documentation for your projects.

The Problem:

Documenting software projects is often a tedious and time-consuming task. It requires a lot of effort to ensure that the README files are well-structured, informative, and engaging. This is where README Generator comes in.

README GPT 🤖 : A Solution:

README Generator leverages the power of generative AI to automate the creation of README files. It ensures that every project has a well-structured, informative, and engaging README.md file, adhering to best practices and enhancing the project's visibility and appeal.

Key Features

AI-Powered Content: Generates high-quality, relevant content.
Customizable Configuration: Tailor generation settings to your preferences with configurable options such as temperature, top p, top k, and safety settings.
Dynamic File Input: Users can add multiple files with custom names and code snippets, enabling comprehensive documentation of project components.
User-friendly Interface: README GPT 🤖 features an easy-to-use interface that allows users to input project details with ease.
Emoji Support: Adds a touch of personality and engagement to the README files.

Why README GPT 🤖?

README Generator is not just a tool; it's a step towards making documentation more efficient, enjoyable, and impactful. Whether you're a seasoned developer or just starting out, this project aims to make README creation a breeze.

Getting Started

I'm inviting you all to join me in making README files the cornerstone of every software project's documentation strategy. Your feedback, ideas, and contributions are invaluable!

🔗 Check out the project repository here: README Generator on GitHub
🌐 Visit our web application here:README Generator Web App

Happy Journey 😊 !!