Artificial intelligence and machine learning transform data into actionable business insights, including making predictions about customers and markets based on historical patterns uncovered in data. As David Weinberger, Senior Researcher at the Harvard Berkman Center for Internet & Society, explains in his thought-provoking new book, Everyday Chaos: Technology, Complexity, and How We’re Thriving in a New World of Possibility, the world as revealed in AI and machine learning doesn’t follow the rules we think it does. In fact, Weinberger presents a new, complex reality for middle market companies based on “un-anticipation,” our inability to predict exactly what will happen next.
We spoke with Weinberger about “Everyday Chaos” and what it might mean for middle market companies. What follows is an edited transcript.
How is AI and machine learning impacting the way businesses look at cause and effect, as well as predictability?
Weinberger: We use machine learning in part because it makes better predictions than we do. But machine learning makes its predictions in ways that we sometimes cannot understand. It doesn't think the way we do. It makes its predictions by looking at vast amounts of data, and it uses statistics to figure out very complex relationships among the data, thus generating predictions.
Machine learning systems embrace all the chaos and particularity of the world. Cause and effect are obviously real, but the world is now being revealed to us as being much more complex than we’d ever imagined. Machine learning is revealing to us that the stories we tell ourselves can be very useful, but they're too simple and too orderly. The real story is so much more complicated.
Why should middle market company leaders care about this uncertainty and “everyday chaos”?
Weinberger: We've always known that everything affects everything else, that the world is incredibly chaotic and unpredictable in small ways. But we're now able to take this unpredictability seriously because we're succeeding in a world that's revealed to us as chaotic, using AI. We succeed on the Internet by using new techniques and strategies, while machine learning is succeeding by transforming vast amounts of data and not relying on a handful of general laws and theories. That success is really important to middle market companies and leaders.
Today, disruption is not the exception: it's everywhere. At any moment, something radical in your environment, typically negative, can occur. Now businesses can pivot quickly because the digital world gives them the ability to change plans and know much more about what's going on. We can be far more agile, and the world is a place where agility is necessary.
How does this concept of developing a minimum viable product, or MVP, topple previous ways of developing products, software, and projects?
Weinberger: Minimum viable product is a product that is launched with the minimal set of features required for customers to buy it, and get good use out of it. Compare this to traditional ways of building products, and Henry Ford is a great example. Ford designed a product that he didn’t change for 19 years. And it was fabulous, the Model T, with 15 million sold. He figured out what the market wanted and nailed it, but that was over a century ago.
With an MVP, you launch a product with minimal product specs so you don't have to guess what your users want or how they're going to use the product or what problems they're going to have. Over time, you add more features. This enables you to launch a product without having to fully anticipate every user's needs. You follow the data, you see what users are doing, you talk with them too. The data's incredibly useful but you talk with users to see how they're talking with one another on the Internet, because that's where a lot of product development happens. In this way, you can make a product that’s far better tuned to your customers without having to anticipate.
What are “open platforms,” and how can they benefit middle market companies?
Weinberger: They’re another example of un-anticipation, but different from the MVP. Many companies in the tech space create an open platform with an API by which developers anywhere can start to use some capabilities the product has. And these developers can tweak the products for their own use, for their client's use: they can extend them, they can integrate the products into their own workflows. An open platform allows people to build stuff with and around your product.
Slack, a great messaging site for teams, established an 80 million dollar fund to help fund the development of applications and integrations they didn’t and couldn’t anticipate. Facebook is another example: early on, they opened up their platform to allow people to create new extensions of software and games, integrated into their own apps. Businesses are doing this for good reason. It lets their products address niche capabilities the company would have never anticipated. It allows the product to be integrated intimately into the user's various workloads, so it becomes an integral part of their lives. This sort of un-anticipation multiplies the value of your product.
You’ve used the word “un-anticipation” a few times. How do you define this concept?
Weinberger: When you anticipate, you are narrowing the possibilities. You try to limit the possibilities to the ones you foresee and the ones you want. The price of this approach has been largely hidden to us. You anticipate and you prepare, and inevitably you either under-prepare or you over-prepare or you well-prepare. Over the past 30 years, we've started to find ways to avoid doing this, going all the way back to on-demand manufacturing and on-demand printing. It's hugely efficient. And with the Internet, there are more ways of doing un-anticipation.
Having agility brings an efficiency to business that we couldn’t have imagined before. Business strategy in one example. Our experience of how the world works has been changing, and the future seems less subject to big general rules and laws that we can follow. So you'd better be paying close attention to everything, because anything can turn out to be a black swan or an opportunity. Today, smart businesses pay attention to the small changes and are ready to pivot, seize opportunities, and recognize risks. That goes against the old idea of strategy as predictions based on assumptions.
How then should middle market company leaders approach strategy?
Weinberger: With humility. We need to understand that our ways of thinking about the world aren’t necessarily how the world actually works, and that's pretty sobering. At a deeper level, they need to recognize that we don't have as much human agency as we thought. We can't control the world as we once thought we could. They should also view the overwhelming complexity and chaotic nature of the universe as opportunity.
What I've been talking about here is making more possibilities and opportunities. Making life less predictable. Making it richer. Thriving in a world in which more people can thrive because they can do more with what’s out there. It's really scary to give up the idea of control, but it’s also profoundly thrilling to see what humans can do with all this uncertainty.
Listen to "Interview with author David Weinberger" on Spreaker.