EP02 - Saturday Nights with Technology : Saturday, September 19, 2026

Hello and welcome to Saturday Night with Technology. I am Zaref Anwar. Today we will quickly go through the major technology stories announced during the week and then discuss two of them in a little more detail. After that, I will have a quick review of Jane King's book, The Unicorn Project, and finally, I will recommend an open-source GitHub repository from Microsoft called Data Science for Beginners.

Let's start with this week in technology. On September 14, Apple released its latest operating system update, including iOS 17, with the new generation of Siri AI. The important change is that Siri is becoming more conversational and context-aware, with deeper access to information and action across applications. The biggest story here is Apple's attempt to turn Siri from a command-based assistant into something closer to an AI agent.

On September 15, Google introduced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking. These models are designed around natural real-time voice conversations, visual understanding, and multi-state reasoning. This is important because AI interfaces are gradually moving beyond typing prompts into continuous conversations. We will return to this story shortly.

Also on September 15, Salesforce and Nvidia announced Koa, a reasoning model designed specifically for CRM and enterprise workflows. This reflects an interesting trend. Instead of almost every company relying entirely on general-purpose AI, we may increasingly see specialized models designed for particular industries and business processes.

On September 16, Google, Nvidia, and Emerald AI announced the AI Energy Management Alliance. The objective is to make AI data centers more flexible. Electricity consumers' AI workload could potentially move or adjust according to electricity availability and grid conditions. AI is increasingly not just a software and computing story, it is also becoming an energy story.

On September 17, Applied Materials announced plans to invest $5 billion in India over 10 years, focusing on semiconductor research, suppliers, and workflow development. This is particularly important because Applied Materials provides technology and equipment required for semiconductor manufacturing. So this is about developing the ecosystem around chip production, not simply assembling electronics.

On the same day, semiconductor company Xperia announced a partnership with Tata Electronics. Both of these companies plan to work together on chip manufacturing, assembly, and testing in India. Combined with the Applied Materials announcement, this makes September 17 a significant day for India's semiconductor ambitions.

On September 18, AI infrastructure company Nscale filed for a U.S. IPO. Its financial numbers illustrate just how expensive the AI infrastructure race has become. The company reportedly reported rapid revenue growth but also enormous losses. Millions of the revenue and billions of the losses. Behind AI models are data centers, GPUs, cooling systems, electricity, and billions of dollars of infrastructure investment.

Also on September 18, Anthropic and Accenture announced a major partnership around testing, evaluating, and red-teaming frontier AI models. It is another indication that AI evaluation and safety testing are becoming businesses in their own right.

Now let's return to Gemini 3.8 Live. For most of the generative AI era, our interaction has been fairly simple. We type something, AI answers, we type again. But live AI systems are moving towards a different model now. You speak, the AI listens, you interrupt, it sees what you are seeing, it remembers the context, it reasons, and eventually it takes actions while the conversation continues. Voice itself isn't new. We have had Siri, Alexa, Google Assistant for years. What is changing is the combination of voice, context, reasoning, and action.

Imagine telling AI, I need to attend a conference next Tuesday, check my calendar, compare the flights, consider when I need to reach the venue, and tell me which option makes the most sense. Then you simply continue. No, that flight is too early. What about traveling the previous evening? That starts feeling less like search and more like collaboration.

And Apple is heading in a similar direction with Siri. Apple has one major advantage. It has deep access to the operating system. Google has another because they also have access to their Android ecosystem. It has extremely capable AI models and cloud infrastructure. I'm talking about Google here. But both are heading towards the same destination. AI becomes the interface between you and your applications.

Today we think open Word, open Excel, open a browser, open a travel app. But perhaps eventually we may simply say, prepare last month's sales report, create three charts, and send it to my team. Applications may still exist underneath. We may simply stop thinking about them. So my question is this, five years from now, will we still open applications or will we mostly tell AI what we want done?

Now let's look at the two India related semiconductor announcements. Applied Materials plans to invest $5 billion. Nexperia is partnering with Tata Electronics. Why does this matter? Because semiconductor manufacturing isn't about building one factory. It requires an ecosystem. You need fabrication facilities, specialized equipment, chemicals, packaging, testing, reliable electricity, water, engineers, suppliers, and customers.

This is what makes the Applied Materials announcement interesting. The company operates close to the actual semiconductor manufacturing processes. Nexperia adds another piece by bringing real semiconductor production and packaging experience into the ecosystem.

But there is also an important reality check. There is a big difference between announcing investments and becoming a major semiconductor manufacturing nation. India is still building these capabilities, and semiconductor development has to be judged over decades, not months.

AI makes all of this even more important. AI needs compute, and compute needs chips. Chips need labs and labs need equipment. Data centers need electricity. So something that looks like a software revolution is actually driving huge investment into physical infrastructure.

The interesting question for India is this. Can India eventually build in semiconductors the kind of global positioning it created in software services? The opportunity is certainly here, but semiconductor manufacturing requires much more capital infrastructure and supply chain depth.

Now let's move to the book of the week. This week's book is The Unicorn Project. Those who are with me on video can see that it is in my hand. If you have read The Phoenix Project, this book returns to the fictional company Parts Unlimited. So The Unicorn Project is the second part of The Phoenix Project. This book returns to the same company, Parts Unlimited, but looks at the technology transformation much more from a developer's perspective. In the earlier book, The Phoenix Project, it was DevOps, the processes. In this book, programming is at the central stage.

The main character is Maxine, a senior developer and architect. She wants to solve problems. Instead she keeps running into approvals, dependencies, broken development environments, missing documentation, and organizational barriers. And this is really what the book is about. Not simply DevOps. It is about creating an environment where developers can actually be productive.

Jan Kim described five specific ideas. The first is locality and simplicity. Teams should be able to make changes without coordinating with half the organization. The second is focus, flow, and joy. Developers need uninterrupted time to actually solve problems. The third is improvement of the daily work. Don't keep fighting the same problems. Improve the system creating those problems. The fourth is psychological safety. People must be able to admit mistakes and raise concerns. And finally, customer focus. Technology should ultimately create value for the customers rather than simply satisfying internal processes.

What I like about the book is its storytelling. Instead of telling you that dependencies reduce productivity, it lets you experience Maxine struggling with those dependencies. Some situations are deliberately exaggerated in my view because it is a business novel. But technology professionals, or many other technology professionals, will recognize the problem immediately when they read the book.

My biggest takeaway from this book is simple. Developer productivity is rarely just about the developer. Architecture, processes, culture, and organizational design can either multiply someone's productivity or destroy it.

And finally, let's look at this week's open source repository. It is Microsoft Data Science for Beginners. It is essentially a free introductory data science curriculum available on GitHub. The course contains 20 lessons designed across 10 weeks. It covers fundamentals such as statistics, probability, and data ethics before moving into areas including Python, Pandas, SQL, NoSQL, data cleaning, and visualization.

What I particularly like is its structure. One of the biggest problems beginners face isn't lack of information. It is too much information. One person says learn Python first. Another says mathematics. Another recommends 10 different online courses. This repository gives you a sequence. Start at lesson one, finish it, move to lesson two. It also contains quizzes, assignments, and practical exercises. And importantly, it doesn't treat data science as only a coding problem. It also asks questions such as where did the data come from? Is it biased? What does it actually represent? Are the conclusions justified?

For someone starting data science, or a developer moving towards AI and machine learning, this is a very good resource to bookmark.

And that's it for this week's Saturday Night with Technology. This week we saw AI becoming more conversational, enterprise models becoming more specialized, data centers becoming connected to energy strategy, and India's semiconductor ecosystem attracting significant investment. We also looked at The Unicorn Project and the importance of creating a better environment for developers. And finally, Microsoft Data Science for Beginners shows how much high-quality learning material is now available openly on GitHub.

I am Zareef Ahmed. Thank you for listening to Saturday Night with Technology, and I will see you next Saturday. Thank you.