Plan For a Powerful Future
Andrew Benedict-Nelson of Insight Labs talked to a few technologists who have an eye on something bigger than the next product launch: Stefan Weitz, director of search at Microsoft Bing; Aaron Frank, manager of business development at Singularity University; and Harper Reed, former CTO of Obama for America. Here’s what we learned from them about where technology is headed next and what it means for organizations.
“Moore’s Law” — named after Gordon Moore, co-founder of Intel — states that the number of transistors that people can fit on an integrated circuit will double every two years. The law has held true since 1958. The implication is that computers become twice as fast every two years. While the specific ways in which computers will be designed may be uncertain, and their development may slow somewhat in the future, the overall trend described by Moore’s Law is still likely to hold.
Stefan Weitz, Microsoft Bing: “Faster computing power probably won’t continue to scale in quite the way it has so far. We keep pushing the envelope by pushing down what we call 'die sizes' — the way in which we actually etch circuitry on silicon substrate. We’re currently down to around 18 nanometers — a tiny number. But the challenge we’re starting to run into with traditional etched circuitry is that the electrons themselves are starting to get too close together. The challenges of actual physics are kicking in here.
So the traditional method of integrated circuits may getting to the point where we cannot fit many more transistors on a silicon die. That being said, there are a number of things people are doing to try and get around this problem. They are looking at three-dimensional chips where they stack chips logically or physically to get around the problem. We’ll see how far we can push this traditional chip architecture that we've been improving over the past 30 years.
There are also some other kinds of computing technologies that are being tried, though they are very far out. People are looking at organic processors that use chemicals instead of electrons. There is also some early work in quantum computing, though it all has a long way to go.”
The key thing to remember is that Moore’s Law is exponential. Many of the technological outcomes of it have been too. But individuals and institutions tend to think and plan in linear ways. In making future plans, organizations should at least consider the possibility that technology will be much more powerful very soon.
Aaron Frank, Singularity University: “What makes Moore’s Law so profound and so incredible is that it is an exponential growth curve. At the beginning of any exponential growth curve, it seems very linear. It doesn't feel like a very dramatic progression. But when you start to hit the ‘knee’ of the curve, it starts to accelerate dramatically, and soon you hit the asymptote. There’s an expression: ‘It took me 20 years to become an overnight sensation.’ That is the way a lot of these technologies operate. People and organizations need to think more about managing their expectations of what it will be like to hit that growth curve.
Many organizations interpret the world through a linear framework. As we help folks re-wire their interpretation of looking at technology, they find that there is a lot of room for excitement and optimism.”
The most important impact of Moore’s Law for average computer users is not likely to be the raw processing power or speed of computers. More important will be outcomes such as new kinds of devices made possible by miniaturization.
Stefan Weitz, Microsoft Bing: “When you think about where computing technology is now and where it could be in the next ten years, you realize that processing power isn't the main problem holding us back. We already have phones with quad-core processors that run screamingly fast, for example. At Bing, we have a few hundred thousand machines running. We can return results in a few milliseconds from hundreds of billions of documents. Is it going to get faster? Yeah. Is that fundamentally going to change how we think about computing and search? Not really. …
You can think of Moore’s Law as having two corresponding trends. The first is miniaturization. You’ll be able to make smaller devices that have similar performance to the best-performing chips that we have today. You might be able to put a quad-core processor on the size of a pinhead. So while the top-end processors that are running PCs and servers might be eight times faster, you would also see things like 'smart pens' or other new kinds of handheld devices that use smaller versions of previous-generation chips.
The increase in processing power will actually manifest itself in the fact that the devices we actually work with on a daily basis will be much more natural to interact with. Functions like voice and multi-modal interfaces (pointing at things and stuff like that) require a tremendous amount of power to work properly.”
The ability of computers to process such “real-world” information could lead to fundamental changes in the way we think about and use technology.
Stefan Weitz, Microsoft Bing: “What will be more important by 2023 will be systems’ abilities to understand the real world, for the Web to become a digital proxy of the physical world. That’s very different from how we think about the Web today. The Web today – or at least the Web of three or four years ago – was one of documents. It was basically a huge connection of documents linked to each other. The design of search was predicated on the information corpus that we were analyzing. Search was basically a keyword- or noun-based operation. You would search for 'Southwest Airlines,' not 'check in on Southwest Airlines.' We don’t do very well handling complex actions.
What’s happening now with the Web is that we’re looking at more 'graphs': the social graph, the geo-spatial graph, the object graph, the economic graph, the personal graph, the local graph — these are all pockets of information. Now, for the first time in history, things like the social graph — things like your friends, who you talk with on a regular basis, the interests you share with them — instead of just being locked up in your head, for many people that will all be machine-readable. That could lead to a profound change in how we think about computers and how they can help people.
I imagine that in 2023 if you land in Fez, Morocco, you would see your system come online, you’ll pay ridiculous amount of money for roaming — hopefully that will be resolved by 2023 — but you probably won’t be doing any 'searching' per se. Your device would know who you are and what your interests are and who you know and what their interests are. It would also know that there are 14,000 restaurants in Fez and they are open for certain hours. The end result would be something like your best, smartest friend who is with you at all times. That would be in part a result of greater processing power, but it’s mainly going to be a result of better programming. That’s a profound difference from the document-centered searching that we have today.”
In theory, one consequence of this new integration between the Web and the world will be an enhanced ability to “off-load” low-level tasks to machines. But human beings will have to think about the ways they actually want to use this ability.
Stefan Weitz, Microsoft Bing: “There’s a ton of research around off-loading things out of your cerebral cortex to some other substrate that can take care of things on your behalf. You’d literally be freeing up some of your own processing power to focus on higher-level tasks, making you more creative and effective. There’s a lot of debate about whether that is actually going to happen. One could argue that calculators should have freed us from the tyranny of long division and made us more creative, but I fear that most of us spend our time reading TMZ and wondering about the name of Angelina Jolie’s next kid. But in theory, we should be able to off-load more rote and routine tasks to devices.
Though it won’t really be devices anymore — this is what I’m really excited about. Everything will be in the cloud, and devices will just be ways of accessing it. There will be more and more things that we will just put there and forget about it. The classic example is phone numbers — no one sees the need to memorize phone numbers anymore because our phones have them all. But we as humans will have to choose how to direct those powers toward noble goals.”
Harper Reed, Obama for America: “The strange thing about the role of a chief technology officer is that we don’t actually deal with a lot of technology — we deal much more with people. I don’t think that a faster computer necessarily means that that role changes. It just means that the tools we use will change. It’s like asking a carpenter whether things would be better if his saw were faster.”
In conjunction with the Internet, these new technologies could also lead to a new kind of cognitive democratization. More and more people will have ways of completing complex tasks that could once only be accomplished by those with access to significant amounts of money, labor or time.
Harper Reed, Obama for America: “All the Internet has done really is make the haves more equal to the have-nots. People are more likely to have access to data and media. An Obama volunteer in ’08 had access to tools that were probably not available to the campaign staff for Bill Clinton. That’s not due to more powerful computers so much as it is due to everything that the Internet brought us: a way of distributing tools more widely and distributing the work more granularly. Think about the introduction of crowd-sourcing and crowd-funding. These were no new technologies that allowed those to happen – it was a change in people’s attitudes about technology.”
Aaron Frank, Singularity University: “Many of these technologies are going to enable us to do things that are so powerful at a price point that is so affordable that it will really allow you to do things that were only available to really large organizations or governments. Small teams could take on much bigger challenges. That should give organizations cause to be optimistic about what they’ll be capable of.”
Law firms and other companies in the legal industry may have to grapple with the fact that in the near future, low-level tasks may be handled more effectively by machines than human beings. Cognitive democratization could also mean that through crowd-sourcing or access to new software, many people can receive some legal services for an extremely low cost.
Aaron Frank, Singularity University: “In a profession like law, much of the work is input/output. Here are the documents that need to be prepared. Here are the filings that need to be completed. One technology that would affect this would be artificial intelligence. We've already seen algorithms that can prepare legal filings or legal paperwork. I think anything that can be automated in the process will be. That should free up a lot of bandwidth for lawyers, helping them think more creatively about problems and then deploying tools like artificial intelligence that can automate much of their current work. Algorithms like Siri on the iPhone will be fairly crude compared to what will come in the future. You could see a more subtle artificial intelligence performing many of the tasks that paralegals currently do, for instance.”
Harper Reed, Obama for America: “Their whole world could be shaken by crowd-sourcing. Imagine if someone uploads a non-disclosure agreement and could quickly get a Wikipedia-level understanding of what that agreement actually says. They wouldn't need to go to a lawyer for that ‘first pass.’”
But the legal industry will also have to deal with much greater amounts of data — not just formal documents, but the kind of “real-world” information that would be gathered by widely distributed miniature computers and devices.
Stefan Weitz, Microsoft Bing: “Let’s imagine a divorce proceeding. Let’s say there’s a custody issue. The ability of a lawyer to understand a person’s social actions or financial actions would be greatly enhanced. The challenge for attorneys will be to sort through all this data. ... The amount of data created this year will be equivalent to all data created from the dawn of civilization to 2003. So 'discovery' will take on a whole new meaning. If you think that it’s bad now for a corporation to have to ship 74 boxes of materials in a class-action suit, imagine ten times that for even the most minute case. That information probably already exists today — it’s just that nobody knows how to find it yet.”
In the end, the impact of new computing technologies on the legal industry may not be so different from the impact they had on newspapers or record companies. Lawyers need to consider what it will be like to operate in a world in which many of the tasks for which their firms currently bill clients will no longer be reliable sources of revenue.
Aaron Frank, Singularity University: “I think you’ll see many of the same effects for law firms that you’ll see for other large organizations. I think they will be smaller and leaner. The need for the work done by the legal profession will probably grow while the number of people actually performing those roles will shrink.”
Harper Reed, Obama for America: “Lawyers are like a lot of businesses that have chosen not to change. A lot of businesses have already been forced to change. But I think that like other businesses that have not chosen to change, lawyers are going to have a pretty unfortunate path going forward. They are probably not going to like what they have to do. How that actually turns out will be fascinating. I am pretty sure they are going to have to be the ones who make the move, but they think that we will be the ones that bring them – ‘we’ being everyone who is not lawyers. I think they are waiting for their sherpas. But technology doesn't care. … They’re not the first kind of business to go through this. This is actually pretty normal for this time in our world. This shouldn't be scary. It should make everyone better at what they do. We just have to figure out how to get them to not stifle it.”