Bold (9 page)

“The cloud is democratizing our ability to leverage computing on a massive scale,” says Weston. “Today the computation speed that someone in the middle of Mumbai has access to outstrips what the entire US government had during the sixties and seventies. We're entering an epic period of global innovation where high-performance computing is abundant, reliable, and affordable.”

So what does all of this computing power buy you? An entirely new way to approach innovation, for starters. Consider brute force, a term that refers to our newfound ability to use infinite computing to literally surround problems. Imagine you wanted to solve a Sudoku puzzle. You could try and build an elegant mathematical approach—derive an algorithm that calculates the correct missing numbers—or you could simply ask a computer to try every possible number in every possible box and then select the one that works best. The latter approach is brute force.

On my tour of Autodesk's Pier 9 Design Center, as a way of illustrating brute force further, Bass points out an electric go-kart he's building with his fifteen-year-old son. “In the old days, when it comes to attaching this electric motor to that go-kart, I would try for an elegant solution—taking an educated guess on the thickness of brackets and best location, then run a few calculations to find out if what I was doing was adequate. Today I can create a computer model and know exactly the stress and strains at every location for my chosen design. But in the near future, with infinite computing, I could ask the cloud to run design simulations, experimenting with every possible location for the motor and a range of different materials and thicknesses, resulting in not just an adequate design, but the best design.”

And what Bass can do, you can do. If your passion is building better go-karts, the technology now allows you to build the best possible go-karts—and in a fraction of the typical time and for a fraction of the
typical cost.

And what is true for go-karts is also true for anything else one wants to create. Moreover, we all learn from our mistakes, but until recently, mistakes were too costly for entrepreneurs to make with wanton abandon. This too has changed. Infinite computing demonetizes error-making, thus democratizing experimentation. No longer do we have to immediately dismiss outlandish ideas for the waste of time and resources they invariably incur. Today we can try them all.

Infinite computing has led to a massive increase in design possibilities, though to really unleash this power, you still have to gather the data, feed it into the computer, then code the algorithms to analyze the data. But what if you didn't? What if you could just talk to your computer and your computer could perfectly understand your desires, gather the data for you, and analyze it in a fashion that would answer your question? Well, in the broadest sense, that's the capability we'll explore in our next exponential technology, the exploding field of artificial intelligence.

“Just what do you think you're doing, Dave?”

Strange fact: These words—their deadly calm and deep menace—were the absolute height of artificial intelligence for nearly fifty years. The line is spoken by HAL, the sentient computer onboard the spaceship
Discovery One
, in director Stanley Kubrick's legendary
2001: A Space Odyssey
, which he cowrote with Arthur C. Clarke.
²⁵
When not fussing about Dave, HAL supports the crew, acting as their interface to the ship, answering questions and helping analyze collected data (the ship is on a scientific mission). HAL's physical presence is depicted as an ominous red television camera eye located on equipment panels throughout the ship.

But that glowing red eye was so last century.

Move over, HAL, say hello to JARVIS.
²⁶
Short for Just Another Rather Very Intelligent System,
JARVIS first appeared in
Iron Man
as Tony Stark's personal AI. Programmed to speak with a male voice in a British accent, JARVIS handles everything from house security to Iron Man suit fabrication to running Stark's global multibillion-dollar business conglomerate—an enormous workload for an extraordinary system.

From a technological perspective, what makes JARVIS extraordinary is both its pervasiveness in Stark's life and its ability to understand natural-language instructions, even when the banter is laden with irony or humor. More technically, JARVIS is a software shell that interfaces between Stark's every desire and the rest of the world, able both to gather data from billions of sensors and to take action through any system or robotic device connected to the AI. In this way, the Internet of Things serves as JARVIS's eyes, ears, arms, and legs.

For sure, JARVIS has dethroned HAL, now holding the title for most recognizable AI in the world, but what makes his dominance more spectacular is that unlike the never-actualized HAL, key elements of JARVIS are starting to come into existence in laboratories and companies around the world.

AI expert and Singularity University cofounder/chancellor Ray Kurzweil
²⁷
explains: “In the 1960s, when Arthur C. Clarke conceived of HAL,” explains Kurzweil, “it was clearly science fiction. Fifty years ago, we knew very little about AI. Today it's a different story. Many aspects of JARVIS are either already in existence or on the drawing board.”

Kurzweil would know. Bill Gates called him “the best person I know at predicting the future of artificial intelligence.” Larry Page hired him as Google's director of engineering, where Kurzweil is leading efforts to develop an AI with natural-language understanding, meaning he's teaching computers to understand the subtle nuances of the spoken and written language, allowing us to ask our machines far more complex questions than “Siri, where can I find a cup of coffee?”

“It's a shift from computers having only logical intelligence to ones that also have emotional intelligence,” says Kurzweil. “Once that occurs, AIs will be funny, get your jokes, be sexy, be loving, and even
be creative.”

Along these lines, in March 2013, I stood on stage at TED, alongside TED curator Chris Anderson, and announced our intent to join forces and design an AI XPRIZE.
²⁸
“Here's the concept,” said Anderson. “An XPRIZE for TED to be awarded to the first artificial intelligence to appear on this stage and present a TED talk so compelling that it commands a standing ovation from you the audience.”

This concept demands that a key number of AI's abilities either equal or surpass human abilities. When this will happen has been a famous and longstanding debate. Kurzweil himself has pegged the date when AIs will do everything better than humans at 2029.
²⁹
(As explained in
Abundance
, his predictions are based on exponential growth curves and have an amazing track record for accuracy.) Certainly, for most entrepreneurs, 2029 is a date too far out to serve as a basis for a business. But no need to wait, as AI is yet another technology transitioning from deceptive to disruptive, and about to become ubiquitous in our daily lives. Consider, in fact, our daily lives. Today, in America, 80 percent of jobs revolve around the service industry,
³⁰
which in turn can be broken down into four fundamental skills: looking, reading, writing, and integrating knowledge. How far has AI progressed? Computers can now perform all four of these skills and in many cases, better than humans.

Let's take a closer look.

All four of these skills have emerged from a branch of AI known as machine learning—which is literally the science of how machines learn. And one thing for certain, today machines are learning faster than ever.

Looking, the first category, has long been a task better performed by humans than computers. “The first time that a machine learning algorithm was able to ‘see' at a level of accuracy similar to humans was in 1995,” explains Singularity University's head of machine learning, Jeremy Howard.
³¹
“That year a US Postal Service competition was won by an algorithm called LeNet 5, which was able to recognize numbers in a zip code and help sort
the mail.”

LeNet 5 algorithm recognizing a handwritten “2”

Source:
http://yann.lecun.com/exdb/lenet/

Progress remained steady (but unremarkable) until 2011, when a series of major breakthroughs put the machine-learning world on high alert. In Germany, an annual competition pits humans against machine learning algorithms in an attempt to see, identify, and categorize traffic signs. Fifty thousand different traffic signs are used—signs obscured by long distances, by trees, by the glare of sunlight. In 2011, for the first time, a machine-learning algorithm bested its makers, achieving a 0.5 percent error rate, compared to 1.2 percent for humans.
³²

Even more impressive were the results of the 2012 ImageNet Competition, which challenged algorithms to look at one million different images—ranging from birds to kitchenware to people on motor scooters—and correctly slot them into a thousand unique categories. Seriously, it's one thing for a computer to recognize known objects (zip codes, traffic signs), but categorizing thousands of random objects is an ability that is downright human. Only better. For again the algorithms outperformed people.
³³

Similar progress is showing up in reading. Today, there are AIs that can accurately and consistently decipher everything from high school student essays to complicated tax forms far faster than humans.
Take legal documents, a linguistic quagmire if ever there was one. Yet, as John Markoff wrote in a 2011 article for the
New York Time
s
:
³⁴
“Thanks to advances in artificial intelligence, ‘e-discovery' software can analyze documents in a fraction of the time for a fraction of the cost. . . . Some programs go beyond just finding documents with relevant terms at computer speeds. They can extract relevant concepts—like documents relevant to social protest in the Middle East—even in the absence of specific terms, and deduce patterns of behavior that would have eluded lawyers examining millions of documents.”

In our third human-skill category—writing—a January 2014 Deloitte University Press report
³⁵
explains that AI is making a dent here too. “Intelligent automation, though still rapidly developing, has matured to the point where it has penetrated nearly every sector of the economy. [In the writing category], Credit Suisse uses a technology from Narrative Science to analyze millions of data points on thousands of companies and automatically write English research reports that assess company expectations, upside, and risk. The reports help analysts, bankers, and investors make long-term investment decisions and has tripled the volume of reports produced while improving their quality and consistency compared with analyst-written reports.”

Integrating knowledge, our fourth skill, represents the much more complex ability to pull together information from many sources and reach accurate conclusions. Here we find arguably the most important breakthrough and the greatest entrepreneurial opportunity. Remember IBM's Watson, the supercomputer who bested humans on
Jeopardy
³⁶
in February 2011? Well, as of November 2013, IBM has uploaded Watson to the cloud, making it a development platform available to anyone, especially entrepreneurs. As Michael Rhodin,
³⁷
the senior vice president at IBM in charge of Watson, says, “Putting Watson on the cloud is aimed at spurring innovation and fueling a new ecosystem of entrepreneurial application software providers—ranging from start-ups and emerging venture-capital-backed business to established players. We've even established a new $100 million venture fund to back
start-ups using Watson.”

One example of a new start-up backed by Watson is Modernizing Medicine. Back in 2011, Modernizing Medicine launched as an iPad-based, specialty-specific electronic medical records platform with a cool crowdsourced twist.
³⁸
For example, all dermatologists who sign up with Modernizing Medicine have their outcome data—that is, what was wrong with a patient and what treatment they prescribed—de-identified (meaning patients' names are removed) and aggregated. This information then becomes available to every dermatologist on the network—some 3,000 of them, or 25 percent of all dermatologists in America—thus significantly improving quality of care. But, in 2014, Modernizing took a huge step forward and partnered with Watson. Since winning on
Jeopardy
, Watson has been sent to medical school—loaded up with millions of journal articles, textbooks, patient outcomes, scientific papers, and the like. By combining their structured patient outcome data with Watson's unstructured research data, doctors on the Modernizing Medicine network have access to incredible levels of point-of-care information. “It would be impossible for us humans to replicate what Watson does in health care,” says Modernizing Medicine CEO Daniel Cane.
³⁹
“Not only can it answer questions pulled from millions of individual documents, it can instantly cite the source and confidence level. Beyond empowering physicians with the most powerful Q&A tool ever created, it will fundamentally change the practice of medicine.”