Professor Leonard Kleinrock
If you missed Part One of our interview with Leonard Kleinrock, check it out here!
Leonard Kleinrock is the Distinguished Professor of Computer Science at UCLA. He developed the mathematical theory of packet networks while a graduate student at MIT. In September 1969, his host computer at UCLA became the first node of the internet. On October 29, 1969, he directed the transmission of the first message over the internet.
What did sending that message look like? Check out his explanation below:
Our conversation with Professor Kleinrock was be published in two parts. You can find Part One here. This interview has been lightly edited for clarity.
You talked about a system that disappears into the environment. Our CEO, Dan Myers, mentioned this conundrum to me: when you look at ads for Smart Home and IoT devices, you never see the power cable. Flair set out to make an almost zero-energy-consuming device with high HVAC energy-saving potential. Think low energy, high impact.
If IoT growth continues its trajectory and smart home devices become part of our built environment, you’re going to have a huge energy problem. We’ve thought about this. Do you think others are thinking about that? Or is the market focused on making high-power Google Homes and Amazon Echoes?
Well you are raising a very good question. I think that the innovation—and I’m using that word with caution—that we are seeing today in devices and apps is not revolutionary. They are not thinking long-term. They are solving a focused problem in a focused environment. And so, you are right, nobody else is thinking about the huge demand for energy.
That kind of approach builds a chaotic system with no overall architecture. There hasn’t really been any global approach to some of these harder problems. In the early days, in the 60s, 70s and 80s, we were solving really-hard problems at a grand scale. That innovation that kind of creativity—we haven't seen much of it. Maybe quantum computing is an example and neural networks begin to approach that as well. By the way, these neural networks and the whole AI world started back in the 50s and 60s at MIT. It’s been around for 60 years or so.
The really-hard problems were funded by some wonderful funding agencies and I'm thinking specifically of ARPA, which is now called DARPA. They provided a funding culture that enabled these great breakthroughs to occur.
How did (D)ARPA help breakthroughs?
Their approach was this: they come to you and they know you and that you’ve done some really great work. In return, they gave you a pile of money, they know you are an expert say in IoT, and they say go do something crazy. They are not going to tell you what to do, they are not going to lecture you. It's okay take a high risk, high payoff approach. Open your mind.
That culture of funding was enormously powerful. It came out of Eisenhower, by the way, he created in ARPA in ’58 after Sputnik went up in ’57. And forming ARPA, which was an agency within the Department of Defense, they brought in some magnificent people: some new PhDs and some great scientists to help them identify the great researchers to throw money at.
They let them do their thing without bothering them. There weren’t a lot of project reports, not a lot of site visits, not a lot of demos, just go ahead and do it we trust you to do something good. Well philosophy was enormously powerful, however, in the ’90s and the 2000s and it started to go away.
Does that have anything to do with the commercialization of the internet in the ’90s? In 1994, you see the launch of Netscape and the internet just explodes after that. It almost moves from a government-funded works project with lots of independent contractors doing long-term projects to a gold rush.
That gold rush had its consequences. The traction and the investment grew enormously but the level of innovation went down. Facebook is not a great breakthrough! It’s not a great dream.
At any rate, because the funding approach diminished, faculty researchers were faced with the opposite approach. They reached for just a little bit of money, for a little bit of time, and had to compete very heavily for that bit of money. The funders were going to tell you what they want you to do with it. You were not allowed to fail. You had to do it quickly.
That the researchers and the faculty had little choice but to accept this horrible indictment is bad enough. The bigger problem is that at a university, the research money was being fed to the PhD graduate students who learned that this is the one way do research. Of course it isn’t, but these young researchers are the faculty of the next generation. There's a long-term effect on the way people view research and innovation. Now that’s turning around a bit, but it’s a very slow process.
So, my point is you asked about what your CEO was talking about with energy and the whole idea of where this vision is going to go. It's going to take some open-minded innovative people to think about it. They have to be motivated for the long-term solution. Short term—six months, a year, two years, or three years—isn’t enough.
IoT is a step in the right direction. It’s moving along. But you mentioned the energy problem. It’s quite right and it’s not being addressed by humanity with a high-level of funding. It’s been piecemeal. You made a solution for your problems, but it hasn't been solved yet for other technologies.