ISR AI Studio

02/08/2025

📱 How to Become an App Developer – A Beginner’s Guide
App development is usually divided into three main parts:

🖌 1. Front-End App Developer
This is also called Client Side development.
Front-end developers build the part of the app that people see and use – the user interface (UI) and user experience (UX).

Technologies to learn:

Flutter
React
Kotlin
Swift
XML

These tools help you create buttons, screens, and smooth designs that users interact with.

🛠 2. Back-End App Developer
This is called Server Side development.
Back-end developers build the “invisible” part of the app – handling data, storing information, user authentication, and running the logic behind the app.

Technologies to learn:

Python
Django (a Python framework)
SQL
MySQL
MongoDB

These help manage databases, handle requests, and process data.

🔀 3. Full Stack App Developer
A Full Stack Developer works on both front-end and back-end.
They know enough to build a complete app from start to finish.

They combine the skills from both the front-end and back-end.

➕ Extra Knowledge (Good to Learn)
These skills help make your app more powerful and easier to build:

Firebase (for authentication, real-time database, notifications)
API Integration (connecting your app with other services)
RESTful API (a method to let apps talk to each other)

✅ In short:

Choose if you want to start with Front-End, Back-End, or Full Stack.
Learn the tools and technologies for that path.

Add extra skills later to become more efficient.

This roadmap helps you understand what to learn first and how everything fits together to build real apps!

01/08/2025

Interesting read for AI FILM MAKERS

Make in India: How Bengaluru’s Script app is revolutionising Indian screenwriting

Forget Final Draft & Celtx, home-grown Scrite, founded by Prashanth Udupa, Punit Thakkar & Surya Vasishta, is affordable and has multilingual support. Praised by Raksh*t Shetty, Raj B Shetty, Abhaya Simha & Mayur Puri, the app has been used to write Malayalam film 'Footage' to a Bengali novel.

01/08/2025

🤖 AI Isn’t Magic – It’s Built on Decades of Evolving Logic

Artificial Intelligence (AI) might feel magical today. You type a question, and an AI answers instantly; you ask for an image, and it draws one in seconds. But behind all this, there’s no real magic—it’s the result of many years of research, logic, and learning.

That was the message shared by Balaji Padmanabhan, Director of the Center for Artificial Intelligence in Business at the Robert H. Smith School of Business, University of Maryland.
He spoke at a discussion about AI’s rapid growth and how it’s shaping different fields.
---

📜 Where it all began

AI’s journey started long ago, with the first theories taking shape in the 1950s and 60s.
Over time, AI systems have changed a lot:

Earlier AI mostly followed fixed rules.

Modern AI can “learn” by studying huge amounts of data, finding patterns, and improving itself over time.

Padmanabhan explained that today’s AI is shaped by memory and experience, similar to how humans learn.
Even though AI systems seem incredibly smart, they’re really built from decades of small steps in logic and computing.

---

🎮 Learning by playing games

A famous example Padmanabhan mentioned is AlphaGo, the AI created by Google DeepMind.
AlphaGo learned to play the complex board game Go, which humans had played for thousands of years.
How? It used a method called reinforcement learning:

Playing millions of games against itself.

Trying different moves, learning which worked.

Improving through trial, error, and feedback.

Even today, most AI still depends on huge data sets, repeated practice, and training—not magic.

🧪 Where AI is making a difference

Padmanabhan said AI is starting to have a big impact in:

Education (e.g., tutoring, writing help)

Healthcare (e.g., medical imaging, chatbots)

And many other fields

He also warned that AI could be misused, such as for:

Manipulated images or videos (deepfakes)

Spreading misinformation Especially because of powerful new models that generate human-like text and media.

🛡️ Beyond technology: Rules and ethics

AI doesn’t work in a vacuum. Padmanabhan stressed that it’s not just about building smarter systems—we also need:

Clear rules and policies

Careful testing and verification

Ways to check if AI makes fair and safe decisions

His key message:

> “We need to govern systems, not just AI.”

---

✅ In simple words

AI feels magical, but it’s built on decades of logic and learning.

Today’s systems learn by finding patterns in data and practicing millions of times.

AI is useful in fields like education and healthcare—but it can also be misused.

We need thoughtful policies to use AI responsibly.

-
A blog by chat GPT based on an article in The Hindu

01/08/2025

🤖 FAQ: LLM ≠ Generative AI ≠ AI Agents ≠ Agentic AI

Why we shouldn’t lump them together—and what that means when you’re building with AI

❓ Q: Aren’t LLMs, Generative AI, AI Agents, and Agentic AI basically the same?

Not really. Each one serves a unique purpose, works at a different level of complexity, and solves different types of problems. Let’s break them down:

---

🔹 Q: What exactly is an LLM?

A: An LLM (Large Language Model) predicts the next word or token based on patterns it’s seen in data.

No memory, no task ex*****on, no real “understanding.”

Think of it as: input → output, that’s it.

✅ Real-life example:
When you type “The capital of France is…” into ChatGPT, it simply predicts the most likely next words (“Paris”). It doesn’t “know” Paris—it’s just predicting based on training data.

---

🔹 Q: What’s the difference between an LLM and Generative AI?

A: Generative AI builds on top of LLMs (and other models) to create new content: text, code, images, music, video, etc.

It explores “latent space” to produce something new—but only after you ask.

Still reactive, not proactive.

✅ Real-life example:
You prompt DALL·E to create “a futuristic cat astronaut.” It generates an entirely new image—but it wouldn’t do it unless you told it to.

---

🔹 Q: So what makes AI Agents different?

A: AI Agents do more than just respond. They:

Detect your intent

Call external tools or APIs

Handle multi-step workflows

They’re modular, functional, and task-oriented—but they don’t set goals themselves.

✅ Real-life example:
A travel booking agent AI that:

Takes your “Book me the cheapest flight to Tokyo next month”

Searches flight APIs, picks the best option

Books it and emails you the ticket

It knows how to do the task, but it won’t spontaneously decide you need a hotel too.

---

🔹 Q: What is Agentic AI then?

A: Agentic AI operates with goals, plans, memory, and context. It can:

Reason about what to do next

Adapt when things change

Launch sub-agents

Keep track of progress—all without human prompting

✅ Real-life example:
An AI personal research assistant that, once given “Help me write a market report on drones,” will:

Break it into tasks: market data, competitors, future trends

Search the web, compile data, summarize findings

Revise the report based on updates or gaps it detects

Decide what’s missing and fill it—without you telling it each next step

---

🔹 Q: Is this just a list of extra features?

A: No—it’s a fundamental shift in system design:

From prediction (LLMs)

To content generation (Generative AI)

To orchestration and task ex*****on (AI Agents)

To autonomy and adaptive reasoning (Agentic AI)

Each stage changes the architecture you build, the tools you need, the risks you face, and the value you can create.

---

✅ Q: Why does this matter if I'm building with AI?

Understanding where your system fits helps you:

Choose the right tools & architecture

Plan for what data and memory you need

Understand the limits of what your system can and can’t do

Spot new opportunities for real differentiation

24/07/2025

My early morning hobby is with . Daily I ask chatGPT to design a poster with thoughts or
It's a nice way to begin the day!
Cheers Makkas! Have a good day!

22/07/2025

Whom the man should save first?
His old father (Past)
His wife (present)
His little child (future)
I strongly believe future is AI (the child).
With this last Sunday I've started my "AI workshop for budding writers"

"The Deluge" is a theme explored in numerous paintings, often depicting the biblical flood narrative or related apocalyptic scenarios. Several artists have created works titled "The Deluge," each with unique interpretations and styles. I've tried my video versions of the Deluge painting, this is just one among them.

So what's your interpretation?

22/07/2025

designed this

21/07/2025

என் படத்தில் ஒரு கார் கவிழும் காட்சி இருக்கிறது. ஏற்கனவே வைத்து அதைச் செய்துவிட்டேன். அதனால் ஏரியல் ஷாட்டில் ஒரு கார் விரையும் இரவுக் காட்சியை செய்து பார்ப்போம் என யைக் கேட்டேன். நினைத்தபடி வரவில்லை. சவுண்ட் எஃபக்டும் சுமார் தான். ஆனாலும் பரவாயில்லை என உங்களுடன் பகிர்கிறேன்.

21/07/2025

அமெரிக்கன் டூரிஸ்டர் போல ஒரு விளம்பர முயற்சி!
விஷுவல் எஃபக்டுகளும், டிரான்சிஷன்களும் பளபளப்பாக இருந்தாலும், அதில் ஒரு லோகோவை பதிப்பது முதல் முத்தம் போல படபடப்பாக பட்டும் படாததுமாக காணாமல் போகிறது.

துவக்கத்தில் வெளிப்படையாகத் தெரிந்த ISR LOGO, சூட்கேஸ் இரு சுற்றுகள் சுற்றி மஞ்சளாக வந்ததும் சூட்கேஸ் உள்ளேயே பத்திரமாக இருந்து கொண்டது. சரி செய்ய முயற்சித்துக் கொண்டிருக்கிறேன்.

#

21/07/2025

பட்டனை தட்டுனா வீடியோ ரெடி. - தற்போது ஏ.ஐ வீடியோ உலகின் செல்லப்பிள்ளை. ஆட்டோ, மெட்ரோ காத்திருப்பு நேரங்களில் வீடியோவை உருவாக்க முடிகிறது. ஆடியோவுடன் வருவதால் எக்ஸ்ட்ரா ஆர்வமாக இருக்கிறது.

நேற்று ஓவியத்தை குறுங்காட்சியாக்க முயற்சி செய்தேன். இன்று விளம்பரத்தை என பிராண்ட் பெயர் மாற்றி செய்து பார்த்தேன். ஏனுங்க, இங்க பாருங்க!

Generated this video thru

19/07/2025

Daily AI - 72
ஒரு பிராம்ப்ட் போதும், ஒரு முழு நீள வீடியோவை உருவாக்கிவிடலாம். பைலட் மூவி எடுப்போர், குறும்படம் தயாரிப்போர் இதனை முயற்சிக்கலாம். சாமர்த்தியம் இருந்தால் எட்டு ரீல் படத்தையே உருவாக்கிப் பார்க்கலாம்.
Create a full length movie with one prompt.

19/07/2025

கேமிரா நகர்வு, பிண்ணனி, உடை உள்ளிட்டவற்றை ஏ.ஐ-யிடம் கேட்டுப் பெற்றேன் என்பது இந்த வீடியோவின் சிறப்பு.