How Microsoft became tech’s top dog again

 

Few businesses have gone through the ups and downs that Microsoft has, despite being almost 50 years old. It was instrumental in starting the PC movement in the 1970s and rose to prominence in the 1980s and 1990s, effectively dictating the direction of the tech industry. The business later went into a tailspin and lost its tech lustre during a "lost decade" that started in 2000 as a result of a federal anti-trust lawsuit and CEO Steve Ballmer's poor management.

When Satya Nadella succeeded Steve Ballmer as CEO in 2014, Microsoft started to rise slowly and steadily, putting its wager on the cloud rather than on Windows. However, it didn't resurrect by outperforming rivals with cutting-edge technologies. Instead, Microsoft made progress by making use of already popular platforms like Windows, Office, and the cloud. And until the ChatGPT chatbot and its integration with Bing took the world by surprise, no one turned to Microsoft for ideas on how we might live and work in the future. Microsoft has transformed its business in a matter of months from one that was about as exciting as a utility company to one that is setting the technological agenda and has emerged as the undisputed champion in artificial intelligence. (AI).

Goodbye, Cortana; you're as stupid as a boulder.

The event took years to prepare for, despite appearing to happen quickly. Cortana, the digital assistant introduced in 2014 to contend with Apple's Siri and Amazon's Alexa, was one of Microsoft's biggest and most embarrassing failures.Cortana was never successful. I wrote about the digital assistant's complete collapse in late 2018. At the time, I argued that sales of the single smart speaker with Cortana integrated in were insignificant compared to Amazon's 50 million sales of Alexa-enabled devices.

I also took notice of something, the significance of which even I had not yet realised: Cortana had been transferred from the AI and Research Division to the Experiences & Users team. By the end of the year, Microsoft's vice president of Cortana, Javier Soltero, had left the business, decimating the Cortana team. I got the impression that it was simply Microsoft's way of saying that Cortana would never succeed as a stand-alone assistant.But there was much more to the decision than that. It symbolised Nadella's understanding that intelligent personal helpers like Cortana, Siri, and Alexa weren't the future of AI. Instead, he believed that generative AI, such as ChatGPT as it exists today, would represent the future of AI and perhaps even computing. That's why Microsoft invested $1 billion in OpenAI, the company that created ChatGPT, less than six months after moving Cortana out of its AI business. It recently disclosed a $10 billion additional expenditure.

Why robots are the most effective digital helpers

Understanding the differences between the development and operation of generative AI chatbots and digital assistants helps to explain why Nadella views ChatGPT as the future and digital assistants as the past.

The New York Times has a detailed analysis of their disparities. But here's a brief summary: Command-and-control systems include digital aides like Siri, Alexa, and Google Assistant. They can only comprehend and respond to particular inquiries, such as "What movies are showing nearby?" or "What's the weather going to be like in Rome tomorrow?"

They are significantly more constrained than chatbots like ChatGPT, which can carry out a remarkable variety of jobs like writing marketing copy, summarising articles, creating graphics, writing code, and much more. And there will be more. Microsoft discussed its AI-powered Microsoft 365 Copilot at its "Future of Work" event. This product can generate Office documents on its own, such as putting together a PowerPoint presentation based solely on a Word document and adding styles and animations to the presentation as a whole. Microsoft asserts that it is even capable of creating a business proposal using only conference notes.

Digital aides are significantly less capable than chatbots, and programming new functions for them can be more difficult.

The Times quoted John Burkey, a former Apple engineer who worked on improving Siri, as saying that even a straightforward change, like adding new phrases to Siri's data collection, can take up to six weeks because it necessitates rebuilding the entire underlying database. It can take close to a year to add a more complicated feature, like a new search tool.

Compare that to the startling rate at which new features are introduced to chatbots like ChatGPT, sometimes on a daily basis it seems. This is so because big language model technology is the foundation for chatbots.

And for this reason, Microsoft is now the clear leader in technology. The solution was to say goodbye to Cortana and concentrate on AI chatbots.