Where AI, architecture, and ambition collide

Open AI's DevDay, Agent Builders & Workflows, LEGO's in Parallelarism | Wed, Oct 8th, 2025

The Spotlight

OpenAI’s DevDay 2025: Introducing apps in ChatGPT and the new Apps SDK

OpenAI has just turned ChatGPT into a platform where apps can reside within the chat. Users can interact with a set of limited apps directly within ChatGPT using MCP connectors.

Think Canva, Spotify, Figma, Coursera, Booking, Zillow and Expedia that you can call by name during a conversation, with an Apps SDK for developers to build their own.

Some believe this is a step towards monetising OpenAI's massive free user base, potentially by earning a finder's fee by referring users to certain products and platforms.
Open AI DevDay

The News

Anthropic plans to open India office, eyes tie-up with billionaire Ambani

Anthropic co-founder and CEO Dario Amodei is in India this week, with plans to set up an office in Bengaluru and explore a partnership with Mukesh Ambani’s Reliance Industries and is expected to meet Prime Minister Narendra Modi.
Read more

----------------------------

AMD stock skyrockets 23% as OpenAI looks to take stake in AI chipmaker

OpenAI will deploy 6 gigawatts of AMD Instinct GPUs over multiple years, starting with 1 gigawatt in late 2026. If the deal is fully exercised, OpenAI would acquire 10% of AMD.
Read more

----------------------------

Introducing AgentKit

OpenAI launched Agent Builder, which is a visual drag-and-drop canva for building agents, ChatKit for embedding chat interfaces, evaluation tools including trace grading and automated prompt optimization, and a connector registry for linking internal tools.
know more

The Toolkit

Junie: Make coding productive and enjoyable with an AI agent, explain your task, let Junie collect the context, write code and run tests for you.
Try it out!

Powerdrill Bloom: AI-first data analysis and visualization platform designed to make data exploration and reporting more intuitive.
Explore!

Grain: AI-powered meeting recorder and insights platform for video calls and virtual meetings.
Try it out!

The Topic

Microservices Architecture

Building Like LEGOs, Not Statues

Microservices Architecture

Ever wonder how giants like Amazon and Netflix can update their massive applications constantly without everything crashing down? They aren't editing a single, enormous block of code. Instead, they build with Microservices, an architectural style that constructs a large application as a collection of small, independent services. It's like building a complex model out of LEGOs instead of carving it from a single piece of stone.

Think of a ride-sharing app. In a microservices model, user authentication, GPS tracking, and payment processing are all separate, standalone services. Each service is built, deployed, and scaled on its own, communicating with the others through well-defined APIs. This separation is the key to agility and resilience.

Core Functionality

Independent Scalability

When a holiday sale hits, an e-commerce site can pour resources into scaling just the ‘Cart’ service without touching the ‘Product Recommendation’ engine. This is incredibly efficient and cost-effective.

Improved Fault Isolation

If the ‘Promotional Codes’ service fails, it doesn't crash the entire app. Users can still browse, buy, and get their products, ensuring the core business continues to run smoothly.

Technological Freedom

Teams have the autonomy to choose the best tools for their specific job. The GPS tracking team can use Go for its high performance, while the user profile team might opt for Python/Django for rapid development. This fosters innovation and allows teams to use the right tool for the right problem.

The "Two-Pizza Team" Analogy

This architecture fundamentally changes how teams work. Amazon famously pioneered the "two-pizza team" model, where a team should be small enough to be fed by two pizzas. Each of these small, autonomous teams takes complete ownership of a single microservice—from development and deployment to maintenance.

Imagine a team of chefs, each responsible for one dish in a massive banquet. The pastry chef can perfect the dessert without worrying about how the steak is cooked. They work independently but in concert, each an expert in their domain. This creates a powerful sense of ownership and allows the entire system to evolve faster and more reliably than if a single head chef had to oversee every detail of every dish.

By breaking down complexity, microservices empower organizations to build flexible, scalable, and resilient systems ready for the demands of the modern digital world.

The Quick Bytes

Up to 78% of companies now actively use AI; generative and agentic AI now core business tech. — source
Deloitte is rolling out Anthropic's Claude to more than 470,000 employees across 150 countries, Anthropic's largest deployment yet
ElevenLabs introduced Agent Workflows, a visual tool that routes conversations to specialized Subagents for handling complex scenarios.
A user connected Claude with Blender, and it’s now generating creative 3D scenes automatically.
Microsoft Copilot Studio Wave 2 plans to release agent orchestration & workflow automation by March 2026. — source

The Resources

[Anthropic Research] Building AI for cyber defenders: AI models are now useful for cybersecurity tasks in practice, not just theory. Read more
[Research paper] Beyond the Hype: A Comprehensive Review of Current Trends in Generative AI Research, Teaching Practices, and Tools. Read more
[Framework] ShinkaEvolve: : Evolving New Algorithms with LLMs, Orders of Magnitude More Efficiently Know more

The Concept

System Design Concept: Multithreading

Ever wonder how your computer can stream music, download a file, and let you browse the web all at once without grinding to a halt? The secret is Multithreading, a technique that allows a single program to perform multiple tasks concurrently. Instead of tackling a long to-do list one item at a time, it smartly divides the work among multiple "threads" that run in parallel.

Think of it like a busy restaurant kitchen. Instead of one chef doing everything—chopping vegetables, cooking the main course, and plating the dessert in sequence—the kitchen has multiple chefs working simultaneously. One chops, another cooks, and a third plates. Each chef is a thread, a smaller task within the overall process of making a meal. By working in parallel, they get the food out to the customer exponentially faster. This is exactly how multithreading makes your software feel so responsive and fast.

This "juggling" act is revolutionary because it allows an application to remain responsive. A background thread can handle a heavy, time-consuming task (like saving a large file) while the main thread remains free to respond to your clicks and keystrokes. By breaking a process into smaller, concurrent threads, you create a system that is faster, more efficient, and delivers a seamless user experience.

Thank you for reading The TechX Newsletter!

Disclamer