Apr 26, 2017 - Economy

Microsoft's Judson Althoff: Machine learning is everywhere

When most people think of Microsoft, they think Windows software and PCs, but the firm's brass say that its future will be defined by selling services related to artificial intelligence and machine learning. Axios caught up with Judson Althoff, Microsoft's Executive Vice President in charge of worldwide commercial business strategy, to understand how the company is helping to advance these technologies across the global economy.

Why it matters: Microsoft, Amazon, and Google are spending billions of dollars a year researching machine learning and AI in the hopes of becoming the dominant vendor in this space. Whoever wins will become an indispensable part of the global economic infrastructure.

Here's what Althoff had to say about the artificially intelligent economy:

What is machine learning?

Machine learning is about taking the data that anyone might have — whether it's a sports franchise or an industrial manufacturer — and using algorithms to actually reason over the data and to predict outcomes that a businessperson can use to make better decisions.

Take Mearsk, the largest shipping and logistics company on the planet. The shipping world is very paper-based, and the data that they have had, traditionally, about their customers, about the products being shipped, and even the vessels that are shipping them, has been very siloed. By creating a "data estate" and leveraging these algorithms to study the data, that's have allows them to deliver goods more efficiently. For instance, they can now predict much more accurately when a specific product will arrive, better provide visibility to their customers when it's happening, and just streamline their business overall.

What's the difference between really good software and machine learning?

It's the ability to anticipate and provide predictive scenarios based on the data you've collected. It's the forward looking assets that are different.

What was Microsoft's business relationship with a company like Maersk 10 or 15 years ago and what is it today?

Classically our relationship with any enterprise company would have been measured by the total number of PCs and servers they have — the original mission of Microsoft was to put a PC on every desktop. Today it is very different. We now collaborate with a company like Maersk on their future business model, how they better engage with their customers, empower their employees, and how we help them transform their products, like enabling them to provide their customers with information about their orders and offer previously unavailable delivery guarantees.

How will AI and machine learning affect the typical worker in today's economy?

There will be less manual labor and more mid-skill and high-skill jobs across the economy. To stick with the shipping example, there may be less of a need for crane operators, but more need for people who can study these machine learning models and make better decisions, and people who maintain automated systems.

How is this technology going to directly affect consumers?

Even companies as consumer-facing as Hershey's are implementing this technology. You'd think there couldn't be an "internet of Twizzlers," but — George Lenhart, who has been working on their AI strategy — they have put sensors on machines that extrude the licorice, and have saved half a million dollars per batch making candy by being able to fine tune things like temperature and pace of making the licorice.

And there are actual products, like Tesla automobiles, that use machine learning. When you own a Tesla, you're owning a product that learns throughout its lifetime. There are features that are delivered over the wires based on all of the data coming off the car. You'll see that more and more where products learn and provide better service over time.

Go deeper