Early last year, the Institute of Electrical and Electronics Engineers revealed the International Roadmap for Devices and Systems, a group dedicated to “identifying trends and developing the roadmap for all of the related technologies in the computer industry.” Following a SXSW keynote on the future of the tech industry, Georgia Tech professor Tom Conte spoke with Free Press Houston to discuss the history of Moore’s Law, why the industry needs to move past it, and how we can protect ourselves from the invasion of Internet privacy.
Free Press Houston: I will admit, before preparing for this interview I had no idea what Moore’s Law is. Can you talk about it and its significance, or perhaps the importance that it used to serve?
Tom Conte: So, Gordon Moore is one of the founders of Intel, along with Boy Noyce. In 1965 [Moore] realized that the semiconductor industry was following a certain trend where the cost of transistors were getting half as expensive every two years. Sometimes the number changed to one-and-a-half years, and sometimes it’s been a little higher than two years, but it was usually pretty steady. For a long time, the way you made them cheaper was to make them smaller, and when you made a smaller transistor, you made a faster transistor. So Moore’s Law, for a lot of people, had been interpreted to mean computers get twice as fast every two years.
FPH: What exactly are you trying to do by “surpassing” Moore’s Law?
Conte: Well, let’s talk about the problem. In 1995, it turned out that, even though the transistors were getting smaller, wires were getting longer. You have to charge up a wire to use it, so the wires started to dominate. So we started to slow down of the transistor wires by doing tricks in microprocessor design. Those tricks would do crazy things, I’ll just put it that way. Nobody knew. I mean, only a thousand people or so around the world knew what we were doing. In 2005, the fun ended. We hit a wall where we could clock the transistors faster to a point where we were clocking in at 1GHz, and they were dissipating 2(W/cm)^2. So to put 2(W/cm)^2 into perspective, that’s the same power density as an operating nuclear reactor core. It’s hard to get rid of that much heat, right? So what we had to do with Moore’s Law kept on giving us more transistors for the same price. We weren’t using them any faster, so we started to put down microprocessors with multiple cores, and if you look, most laptops and cellphones advertise how many cores they have. But that didn’t really speed up general programs, it only makes things faster when you have two completely independent things. So Moore’s Law effectively — not Moore’s Law, but the computer performance — effective ended. They were detonated. Since that time we’ve been trying to figure out other fundamentally different ways to compute that we could use to continue this exponential growth in computer performance or anyway to want to continue that is fundamental improvement in the human condition in every aspect of life. It isn’t just your iPhone, it’s anything from medical diagnosis to telecommunications to how your car runs. I mean, every single aspect of life was improved because of that, so we need to get back back on track then, right?
So in 2012 I helped launch IEEE, the Institute of Electrical and Electronics Engineers. It was an effort to look at fundamentally different ways to compute. We call it rebooting computing. And since that time we’ve been holding summits of experts, trying to exchange ideas, and we’ve come up with some cool ways to compute. Things like computing using devices that operate similar to how the brain is wired. Another one that people have heard a lot about is quantum computing, so computing using, in essence, the parallel universe nature of quantum physics. There are a lot of interesting ways we can compute. The question now is, how do we build all of that? Because the semiconductor industry keeps giving us the same kind of transistors. We don’t want that, we want something different that will let us build these unique kinds of computers. So that’s what the group is about, it’s about taking applications that we care about, like voice recognition and self-driving cars, and figuring out how the novel computers, how they have to work, and then figuring out how individual transistors we need to make to make it work.
FPH: What exactly is the IRDS?
Conte: That’s the roadmap. That’s our entirety as a roadmap, the International Roadmap for Devices and Systems. By systems we mean computer systems; by devices we mean transistors and others things made of silicon or whatnot. We’re trying to link the two, and say, if we want to have a self-driving car that runs on a tablespoon of gasoline, or maybe 2 AA batteries — this is kind of silly, because they won’t — in other words, some extremely efficient way that’s extremely likely. We have to figure out how to get to build independent devices. We need to make that happen.
FPH: Do you feel that right now younger generations are where they need to be in terms of enthusiasm about tech innovation?
Conte: That’s a good question. You know, I teach at Georgia Tech, which is one of those “egg-head” places. I teach freshman and sophomore up to graduate students, and the enthusiasm I feel is really strong among them. There’s a whole generation of tinkerers and hackers coming forward. These are people that started with gaming console, building a gaming computer, or just playing on their phones and figuring out how to go further. I mean, they are doing incredible things. It started with Lego Mindstorms and building robotics, and again, they kept on pushing the limit. It’s amazing, the kind of enthusiasm that this generation has. I think it’s fundamentally higher than others, and I’ve been in academia for coming on two decades. There’s no stigma attached to playing with computers. It’s not only fine, it’s cool to to try to make computers do something interesting. Back in the day, it was for the ultra-nerds who put on a beanie hat with a propellor on top; now, this generation has really embraced it. It’s great.
FPH: When the excitement occasionally lowers, what keeps you motivated to continue what you do?
Conte: I really think it is the students. When I’m feeling down I go talk to some of the students or walk through a lab and see their enthusiasm. It makes me feel better, because, like I said, the enthusiasm in the next generation is really high.
FPH: I wanted to get your opinion on the recent controversy of congress and internet security, and the urgency of setting up a VPN; Is the problem as severe as the media has made it out to be? Is our privacy really for sale?
Conte: So, the President, as you might now, signed a law that lets our ISP’s collect our browsing history, so for me, that’s an incredible invasion of privacy. I don’t know about you, but I’m running through a VPN, so that they can’t get ahold of it. I don’t want anyone to collect information about what I’m browsing. It just doesn’t make sense. They might as well put a meter in my bedroom. I mean, it’s way too private for me.
FPH: How is that information sold? Who is the typical buyer?
Conte: Marketers. If they know what you’re browsing, they know what you’re interested in. They can start to do even more targeted marketing. You would contact AT&T, for example, and say “Who in Northeast Atlanta is interested in this? What are they interested in? Tell me what this one particular house is interested in, and we’re going to try to get him to buy targeted products.” I don’t want someone making up my mind for me, do you know what I mean?
FPH: It seems as if people like Elon Musk are modern rockstars. Do you think there is anything wrong with how quick he is trying to change the game of conventional innovation, whether that be with Tesla and SpaceX, perhaps even SolarCity?
Conte: Well, it’s a self-fulfilling — I don’t want to say prophecy, it’s more anticipation. Having computers get faster and faster does more and more things, and people become to expect it. There is a lot of enthusiasm from trying to figure out how to keep that going. It’s a great time to get into computers because everything is about to change. We’re trying out novel things, stuff we haven’t really changed since the 40’s It’s an exciting time in the sense of people wanting to push the limit. Let me tell you another story. I tell it every now and then. Do you remember when I said since 2005 computers haven’t really been getting that much faster? I’ve had members of my family come to me and say “I’ve bought a new laptop, and it’s no faster than the one from five years ago. Why?” It’s finally dawning on people that, “Hey, there’s something going on.” I think it’s starting to sink into the zeitgeist that we have this bubbling catastrophe coming soon, so we have to do something about it.
FPH: Besides Moore’s Law, what is something that is critical for the public to do more research on that media is not covering?
Conte: I think, really, the security issues we have, and just securing communications and transactions. You don’t need an alarm system today in your house, because that’s not where the money is, it’s in your phone and computer. People were cracking into our devices and stealing, and I think the awareness needs to increase to the public, and in the industry as well, to try to be fundamentally secure things that can’t be cracked. That’s the other tsunami I see: security.
Pros and Cons of Tech in 2017: An Interview with Prof. Tom Conte This was reposted for my personal reading use