Skip to main content

Beyond Digital Doom: Exploring the Future of Innovation (Part 2)

Roland Alston, Appian
March 7, 2019

Greg Satell, Author, Speaker, Innovation Adviser

This is the final installment of our two-part series on innovation, featuringGreg Satell, author ofMapping Innovation, andCascades: How to Create a Movement that Drives Transformational Change, which will be published by McGraw-Hill in April, 2019. Learn more about Greg on his website,GregSatell.comand follow him on Twitter@DigitalTonto. Read Part 1 here.)

In the age of ever-changing customer expectations, the wrong approach to innovation can be lethal. We're talking the conventional wisdom of listening to customers, investing in continuous improvement and watching the bottom line.The danger is that if market conditions shift, you end up getting continuously better at things customers no longer want.

That's when it's time to innovate your business model not your product.

So says Greg Satell.

"The truth, says Satell, "is many organizations get stuck because they end up locking themselves into a single strategy. They find something that works and say, "this is how we innovate" and end up trying to apply essentially the same solution no matter what the problem is. Eventually, that ends badly."

"That's why organizations that were once called great innovators fall behind. They get snagged on the mindset of being a square-peg company in a round-hole world. And they lose relevance."

Much better to match solutions to problems, not the other way around.

Which is a good segue to the final episode of our 2-part interview with innovation adviser Greg Satell.

Hope you enjoy the conversation.

Appian: I want to shift gears and talk about the argument you make that it's time to think less about hackathons and more about tackling big challenges. What did you mean by that?

Satell: Yeah, I was talking the new era of innovation. We've come to associate innovation with digital technology and agility. And that's largely because we understand digital technology really, really, well. And we've gotten used to this idea that every two years, we're going to get a new chip that's twice as powerful and better than the last one.

Appian: It was like free innovation.

Satell: And it's also like the value has shifted to the front end, to things like the user interface and design. Think about the iPod, and Steve Job's idea of putting a thousand songs in your pocket.

So, the manufacturer of the technology (behind the iPod) made money, but not nearly as much as Apple did with the iPod. So, over the last 20 or 30 years, a lot of that value has shifted to the front end. But now, that's all ending. Moore's Law is ending. So, the question is: "What do we do when that happens?"

Post-digital computing: On the cusp of change

Appian: So, what comes next? And how do we take advantage of it?

Satell: We need to come up with new fundamental technologies. And there's a bunch of candidate technologies, primarily quantum computing and neuromorphic chips. But they don't work anything like the old microchips. And that's going to cause some problems.

Appian: But isn't quantum computing more futuristic? Or is it something that we need to worry about right now?

Satell:People are using quantum computers right now, and we know that they're potentially extremely valuable for things like simulating physical systems. It won't pay off for at least five years.

But in five years, the companies that aren't preparing for the age of quantum computing are going to have a real problem, because they won't know how to take advantage of the technology.

Forget move fast and break things

Appian: Which gets back to your argument that it's time to prepare for life after digital. What does that mean for business leaders?


It means that we've spent the last couple of decades learning how to go fast. We're going to have to spend the next couple of decades learning how to go slow again.

You can't rapidly iterate a quantum computer. You can't rapidly iterate a revolutionary new material for solar panels.And you can't rapidly iterate a cure for cancer.

One of the organizations I talked to quite a bit is the Joint Center for Energy Storage Research at Argonne National Laboratory (JCESR). This is a consortium of five national laboratories, research universities and a network of over 100 private sector companies.

Appian: So, what did you talk to JCESR about?

Satell: They're on a mission to identify the battery chemistry of the future. They were supposed to come up with a prototype for the grid and another for devices and cars. Because when you think about it, it's kind of crazy that we use lithium ion for both.

By the way lithium ion has a similar problem to Moore's Law. It's also reaching its theoretical limits.

Not quite as fast or imminent as Moore's Law we're talking perhaps five to 10 years out.

So, JCESR not only identified one chemistry for each. They identified two chemistries for each. Now we know the chemistries that are viable for next generation batteries. And we know what it takes to make them work. The only problem is that the materials don't exist yet.

Materials science: Sparking a revolution

Appian:That sounds like a problem without a solution.

Satell: But it's a much smaller problem than it used to be.

I think that materials science is the most important technology over the next decade or so.

So, what we've started doing is building up these materials genomes, and we're using high performance supercomputers to basically do it faster than ever before.

Appian: So, what benefits do we get out of speeding up the traditional way we do materials science?

Satell: Traditionally, you start with the properties you want and then work backward with structures and processes to get those properties. Consider the 787 Dreamliner, the new plane by Boeing. It works pretty much like it's predecessor the 777,but it's 20% lighter and uses 20% less fuel.

That kind of innovation has huge economic and environmental benefits when you leverage it across the entire global market for aviation.

Appian: You mentioned that we're using supercomputers to create materials genomes. What's the significance of that?

Satell: The problem with traditional materials science is that you can spend tons of time going through all the material possibilities and testing them until you find something that's better than what you already have.

But with computer simulations, you can take thousands of possibilities and narrow it down to just 40 or 50. You can eliminate over 90% of the possibilities and solve the material science problem at astonishing speed.

Appian: So, we're getting a glimpse of the materials science revolution. But how long will it be before we see this kind of innovation happen?

Satell: The materials experts that I've been talking to tell me that within the next decade we'll have a 10x improvement in development of materials. Which would change the economics of lots of things.

Companies spend hundreds of millions of dollars on materials research. If you can speed up the (discovery) process, you're talking about creating an entirely new economic model.

The astonishing evolution of synthetic biology

Appian:You've also written about genomics and the rise of synthetic biology.

Satell:The discovery of CRISPR in 2012 (a gene editing tool) represents a quantum leap in our ability to do genetic engineering.

I've never seen investment happen anywhere as fast as it has happened with CRISPR never. It was discovered maybe five years ago. And it's already millions of times more accurate than what (gene editing) we were doing before.

Investors can't throw money at it (CRISPR) fast enough. And we're already seeing drugs in trial.

Appian:Switching gears, what about the mainstreaming of AI. It's everywhere now. Where does the evolution of AI fit into your category of revolutionary trends

Satell: AI is another technology to watch. But AI is not a digital technology necessarily. Neuromorphic chips are very well suited to AI. Quantum computers will play a big role in running and training the algorithms for AI, because of quantum computing's ability to handle complexity.

Automation: Net positive for human labor

Appian: When you think about all of the hype around AI, what do you make of the debate about whether AI will be a net positive or net negative for society?


You mean are the robots going to take our jobs? Nobody who's actually looked at the problem thinks that. Nobody who has seriously studied this problem thinks that.

Appian: But there are plenty of dark, dystopian predictions about the AI apocalypse. It's all over the internet.

Satell: First, just look at the numbers. We've had AI applications and robotics now since about 2011. And during that time, we've gone from an unemployment problem to a massive labor shortage. Talk to anybody running a business, especially a manufacturing business which is the most easily automated and you'll learn that that's where the labor shortage is worse.

So, nothing in the data says that the rise of AI is going to be a net negative for human labor.First of all, automation doesn't replace jobs, it replaces tasks. And every time you automate something, to a certain extent you also commoditize it. So, value shifts to another part of the organization.

When you automate something, it sort of sucks the value out of it. But it also creates a need to create value somewhere else.

Appian:Intelligent automation is a hot trend right now. KPMG recently released a study saying that nearly half of all corporations intend to use some form of intelligent automation at scale over the next three years. Is that how companies are going to compete in the future? If that's the case, which companies do you think are getting automation right?

Satell: So, if you want to understand the value of automation, just go to an Apple store, which is one of the most automated retail experiences you can imagine. When you walk into an Apple store, you see a sea of blue shirts. That's because the function of a retail store is no longer to drive transactions. It's to do everything that you can't do online. Get advice, up sell, get service, get training, whatever it is that you can't get online.

Beware the AI ethics conundrum

Appian: In wrapping up, let's talk about another hot topic the ethical implications of AI and intelligent automation.

Satell: Yeah, so that's a big problem. That is a major, major problem. I wrote about that in Harvard Business Review a couple of years ago. I was at an academic conference up at IBM, and everybody was saying that we've got a serious, serious problem.

The thing is, there's not one ethical problem, there are a number of them.

Appian: So, how would you describe these ethical problems?

Satell: You've got the basic (science fiction) stuff like machines taking over the world and Skynet (a fictional general AI system that seeks to eliminate humanity). And you have other philosophical problems that are good for cocktail party conversations.

But AI problems are becoming real. If you put AI in a car, sooner or later that car is going to have to make a decision that will harm a pedestrian or the driver.

Appian: There's a lot of commentary out there about the urgency of dealing with the problem of AI bias? What do you make of that argument?

Satell: The technical term for that is bias in the learning corpus. That's a really big problem. Do you have kids?

Appian: Yes.

Satell: when they were growing up, you worried about what they were learning in school, and who their friends were, and what TV shows they were watching, because you worried about what was influencing their learning process.

So, the question is what influences are our algorithms being exposed to? Who's watching that?

Microsoft Tay is a classic example. They put a (AI chatter) bot on twitter. And within 24 hours, it was exposed to Twitter trolls that had it repeating bigoted and misogynistic misinformation.

Even more subtle than that, there's a great book by Cathy O'Neal called "Weapons of Math Destruction", in which she makes the point that we don't know how any of these algorithms are being trained.

Appian:And they're already touching almost every aspect of our lives.

Satell: Yes, they're making decisions about who gets exposed to predatory marketing, who gets a job, who goes to jail, who gets paroled, who gets a mortgage.

So, the ethics question is a major, major problem. The big tech companies seem to be trying to get ahead of it.

When I wrote about AI and ethics two years ago, they were setting up something called The Partnership for AI to come up with ethics standards for AI.

Everybody's worried about who's influencing our kids. But we should also be worried about who's teaching our algorithms.