
Andrew Viterbi, co-founder and former longtime chief technology officer of Qualcomm, the giant San Diego company that creates the guts of the world’s cellphones, devised the Viterbi Algorithm, a mathematical formula for separating signal from noise. The Viterbi Algorithm has made cellphone communication possible, as well as the tracking of missiles, and even high-speed DNA analysis.
In person, Viterbi is like your gentle Italian uncle, soft-spoken but with a steel-trap mind. He often finishes a thought with a chuckle and a smile.
In the boardroom of his offices at Viterbi Group, he spoke to Intellectual Capital about the future of technology: what could slow us down, surprise, and scare--
Q: What might the world of technology look like in 20 years?
A: My answer is very simple: look back 20 years. The year the Internet opened up was 1994, probably the second most important event in digital communications, after the transistor, and that was 1948. So 20 years ago, who would've thought we would have a Facebook?
The first question to ask is whether Moore's Law, the speed of chips, will continue to double every 18 months? I don't know. Economics could stop it. Foundries cost $10 billion now, when 10 years ago $1 billion was the norm. Whose got that kind of money? Google does. The Taiwanese, the Koreans do.
But if you can do the speed — and maybe we will double every four years, rather than two — and do it at reasonable cost, then who knows what new things are possible? Virtual reality — Oculus — is becoming a consumer product. Your refrigerator is going to be talking to the supermarket, and they’ll send over a drone with your groceries. We’ll have “the Internet of all things.”
Of course, you start concocting these great visions of the future, well, a number of them will never happen. We used to talk about automobiles that would sprout wings and fly. Even the drones — we’ll be having gridlock in the sky.
Q: So you can imagine a few bumps in the digital highway?
A: Historically, software has always lagged. In some cases, we squander opportunities. But if software can keep up, we’ll have unimaginable developments in the biological arena. Brain research, one of the few things the government has been able to fund more or less seriously — but whether it will all have curative consequences? That’s an engineering thing.
Science is figuring out what works and engineering is can you make it?
Q: Commercial opportunities grow out of technical challenges?A: Cybersecurity is going to be a big field, certainly. The NSA employs some of the best mathematicians in the country. I believe using smart people in our own defense is one aspect of the NSA we better preserve. I’m sure the Russians do, even though they’ve lost a lot of their best people.
Q: Like Google co-founder Sergey Brin and his parents.
A: Yes.
Q: Do you worry about privacy issues?
A: Even if from a civil liberties point of view some of what the NSA does is somewhat questionable, if we disable the NSA, we are cooked.
Q: What about ceding the U.S. role over the Internet to the United Nations, when the ICANN (Internet Corporation for Assigned Names and Numbers) contract expires in 2015?
A: The Internet was developed here. We shouldn’t give that up to the U.N.
Q: Why?
A: Because the other governments are all worse than ours.
Q: What else worries you about the future?
A: Nuclear threats. Proliferation, starting from Iran and North Korea and Pakistan.
Q: And at home?
A: I worry that our K-12 public school system does not prepare the kids in science, mathematics and engineering. Our private schools do better and we are all probably better than two to three years ago but still, engineering isn’t as popular, even today. And scientists, well, they are those people in the white coats that nobody understands. We have to import smart people, and we do. But we should be smarter about it.
Q: How?
A: The students we train here at our best universities — and our universities are recognized as the very best in the world — we graduate them, give them their Ph.D.s, and then tell them to go home. Used to be they would fight like hell to stay if a company like Qualcomm wanted them. Our success rate at keeping them was about 95 percent when I was there. But today, the Indians and the Chinese in particular are incentivized to go back with an American degree, and that’s a big degree.
Our immigration policy is very complex, and I tend to be on the liberal side, but these foreign-born scientists and engineers and CEOs are very productive, and we should value that.
Q: You’re on record as being bothered by the inequality gap between CEOs and workers.A: Yes, and the gap is getting worse. It’s not just the CEOs. It’s the tax policy. I’m not happy when I have to pay taxes, but I think it’s more appropriate I pay or Warren Buffett or Mitt Romney pays rather than somebody who is under the poverty line.
Q: So, Andrew, how did you come up with the Viterbi Algorithm, in a blinding flash of the obvious, in the shower, walking on the beach?
A: Took me a long time and I wasn’t even looking for it. Many months of focused hard work, knowing the field and manipulating a few equations. But a lot of that goes on in your head. I was only about 30, 31. Your thought processes are much more well-defined and swift then they are at age 80
I was looking for the proof of a theorem involved with certain error-correcting codes, a class of codes called convolutional codes which are embedded in many examples in nature. For example, voice recognition, how you’re talking to me right now. Your phonemes [the smallest unit of speech] come out one at a time, each dependent on the previous one, and maybe on a few others but not on a lot. That’s called a Markov sequence, after Andrei Andreyevich Markov, a Russian mathematician at the turn of the last century. And my algorithm just fit. It was a step in the proof. I was looking for what most likely was sent, ones and zeroes, physical signals--that’s the purpose of coding, of all digital communications effectively.
Q: Right. I mean, I think so--
A: When I found it, I knew I had something, but you know, it took a couple of years. I was excited enough that I sent the manuscript to a number of people. The first comment I got back was from a friend at Notre Dame, turned out to be the reviewer for the journal, but he didn’t tell me until much later. And another friend was one of the guys who I consider today to be the best in the field. He said it was, “very interesting.” Later on he said it was a break-though. Anyhow, it got very quick publication.
Q: I’m still not quite sure I get it--
A: It’s like predicting the winner in a tennis match from the beginning. A better example might be “March Madness” where 64 teams are matched up in a tree structure and the 32 winners of the first 32 matches are matched up in pairs to play 16 matches and so on. My algorithm might be used before the start of the tournament to find the most likely winner at the end, all based on each player’s likelihood of winning each game against his/her opponent.
Q: Now that is a very practical application!
Currently, the Viterbi Group is invested in On Ramp Wireless, a San Diego company which aggregates sensor readings for utilities, pipelines, and farming, and sends using CDMA technology; a Seattle company in the RFID chip business (reading tags electronically); and an Israeli company which does traffic monitoring and reporting wirelessly, without the need for sensors in the road.
Michael Ma and Subin Ryoo helped with research for this column. Steve Chapple’s Intellectual Capital covers game-changing people, ideas and perspectives.
(c) Steve Chapple ALL RIGHTS RESERVED
In person, Viterbi is like your gentle Italian uncle, soft-spoken but with a steel-trap mind. He often finishes a thought with a chuckle and a smile.
In the boardroom of his offices at Viterbi Group, he spoke to Intellectual Capital about the future of technology: what could slow us down, surprise, and scare--
Q: What might the world of technology look like in 20 years?
A: My answer is very simple: look back 20 years. The year the Internet opened up was 1994, probably the second most important event in digital communications, after the transistor, and that was 1948. So 20 years ago, who would've thought we would have a Facebook?
The first question to ask is whether Moore's Law, the speed of chips, will continue to double every 18 months? I don't know. Economics could stop it. Foundries cost $10 billion now, when 10 years ago $1 billion was the norm. Whose got that kind of money? Google does. The Taiwanese, the Koreans do.
But if you can do the speed — and maybe we will double every four years, rather than two — and do it at reasonable cost, then who knows what new things are possible? Virtual reality — Oculus — is becoming a consumer product. Your refrigerator is going to be talking to the supermarket, and they’ll send over a drone with your groceries. We’ll have “the Internet of all things.”
Of course, you start concocting these great visions of the future, well, a number of them will never happen. We used to talk about automobiles that would sprout wings and fly. Even the drones — we’ll be having gridlock in the sky.
Q: So you can imagine a few bumps in the digital highway?
A: Historically, software has always lagged. In some cases, we squander opportunities. But if software can keep up, we’ll have unimaginable developments in the biological arena. Brain research, one of the few things the government has been able to fund more or less seriously — but whether it will all have curative consequences? That’s an engineering thing.
Science is figuring out what works and engineering is can you make it?
Q: Commercial opportunities grow out of technical challenges?A: Cybersecurity is going to be a big field, certainly. The NSA employs some of the best mathematicians in the country. I believe using smart people in our own defense is one aspect of the NSA we better preserve. I’m sure the Russians do, even though they’ve lost a lot of their best people.
Q: Like Google co-founder Sergey Brin and his parents.
A: Yes.
Q: Do you worry about privacy issues?
A: Even if from a civil liberties point of view some of what the NSA does is somewhat questionable, if we disable the NSA, we are cooked.
Q: What about ceding the U.S. role over the Internet to the United Nations, when the ICANN (Internet Corporation for Assigned Names and Numbers) contract expires in 2015?
A: The Internet was developed here. We shouldn’t give that up to the U.N.
Q: Why?
A: Because the other governments are all worse than ours.
Q: What else worries you about the future?
A: Nuclear threats. Proliferation, starting from Iran and North Korea and Pakistan.
Q: And at home?
A: I worry that our K-12 public school system does not prepare the kids in science, mathematics and engineering. Our private schools do better and we are all probably better than two to three years ago but still, engineering isn’t as popular, even today. And scientists, well, they are those people in the white coats that nobody understands. We have to import smart people, and we do. But we should be smarter about it.
Q: How?
A: The students we train here at our best universities — and our universities are recognized as the very best in the world — we graduate them, give them their Ph.D.s, and then tell them to go home. Used to be they would fight like hell to stay if a company like Qualcomm wanted them. Our success rate at keeping them was about 95 percent when I was there. But today, the Indians and the Chinese in particular are incentivized to go back with an American degree, and that’s a big degree.
Our immigration policy is very complex, and I tend to be on the liberal side, but these foreign-born scientists and engineers and CEOs are very productive, and we should value that.
Q: You’re on record as being bothered by the inequality gap between CEOs and workers.A: Yes, and the gap is getting worse. It’s not just the CEOs. It’s the tax policy. I’m not happy when I have to pay taxes, but I think it’s more appropriate I pay or Warren Buffett or Mitt Romney pays rather than somebody who is under the poverty line.
Q: So, Andrew, how did you come up with the Viterbi Algorithm, in a blinding flash of the obvious, in the shower, walking on the beach?
A: Took me a long time and I wasn’t even looking for it. Many months of focused hard work, knowing the field and manipulating a few equations. But a lot of that goes on in your head. I was only about 30, 31. Your thought processes are much more well-defined and swift then they are at age 80
I was looking for the proof of a theorem involved with certain error-correcting codes, a class of codes called convolutional codes which are embedded in many examples in nature. For example, voice recognition, how you’re talking to me right now. Your phonemes [the smallest unit of speech] come out one at a time, each dependent on the previous one, and maybe on a few others but not on a lot. That’s called a Markov sequence, after Andrei Andreyevich Markov, a Russian mathematician at the turn of the last century. And my algorithm just fit. It was a step in the proof. I was looking for what most likely was sent, ones and zeroes, physical signals--that’s the purpose of coding, of all digital communications effectively.
Q: Right. I mean, I think so--
A: When I found it, I knew I had something, but you know, it took a couple of years. I was excited enough that I sent the manuscript to a number of people. The first comment I got back was from a friend at Notre Dame, turned out to be the reviewer for the journal, but he didn’t tell me until much later. And another friend was one of the guys who I consider today to be the best in the field. He said it was, “very interesting.” Later on he said it was a break-though. Anyhow, it got very quick publication.
Q: I’m still not quite sure I get it--
A: It’s like predicting the winner in a tennis match from the beginning. A better example might be “March Madness” where 64 teams are matched up in a tree structure and the 32 winners of the first 32 matches are matched up in pairs to play 16 matches and so on. My algorithm might be used before the start of the tournament to find the most likely winner at the end, all based on each player’s likelihood of winning each game against his/her opponent.
Q: Now that is a very practical application!
Currently, the Viterbi Group is invested in On Ramp Wireless, a San Diego company which aggregates sensor readings for utilities, pipelines, and farming, and sends using CDMA technology; a Seattle company in the RFID chip business (reading tags electronically); and an Israeli company which does traffic monitoring and reporting wirelessly, without the need for sensors in the road.
Michael Ma and Subin Ryoo helped with research for this column. Steve Chapple’s Intellectual Capital covers game-changing people, ideas and perspectives.
(c) Steve Chapple ALL RIGHTS RESERVED