Pod Insights Decoder Recode Decode: Donna Dubinsky and Jeff Hawkins, co-founders, Numenta About this episode Donna Dubinsky and Jeff Hawkins — the co-founders of Palm, Handspring and Numenta — talk with Recode's Kara Swisher about their efforts to decode the human brain. They say everything that makes us human, from language to art to engineering, derives from the same learning algorithm, and Numenta hopes to ultimately teach that algorithm to a machine. Dubinsky and Hawkins explain why we shouldn't be afraid that machines will take over the world and why they believe artificial intelligence will drive advancements in technology for the rest of the century.
Learn more about your ad choices. Visit podcastchoices.com/adchoices
Transcript 0:00 Speaker 00 Today's sponsor is Casper, an online retailer of premium mattresses for a fraction of the price. Get $50 toward any mattress purchase by visiting casper.com/recode and using the promo code RECODE. Terms and conditions apply. Today is also sponsored by SoFi, a new kind of finance company that offers student loan refinancing at low rates. SoFi members save, on average, $19,000 and can save even more through an employer partner program. See how SoFi can help you at sofi.com. Terms and conditions apply at sofi.com/legal.
0:35 Speaker 01 Recode Radio presents Recode Decode, hosted by Kara Swisher, powered by digital media.
0:42 Speaker 00 Hi, I'm Kara Swisher, Executive Editor of Recode. You may know me as the person who convinced the Tribune Company to change its name to Tronk, but in my spare time, I talk tech and you're listening to Recode Decode, a podcast about tech and media's key players, big ideas, and how they're changing the world we live in. You can subscribe to Recode Decode at iTunes.com/recodedecode and while you're there, leave us a review. Today in the red chair are Donna Dubinsky and Jeff Hawkins, who I've known for a dog's age. They're the co-founders of Numenta, along with a lot of other companies. Numenta is a machine intelligence company that tries to help other companies take action based on data. They've been at it for a long time, well before this trend has become a big one in Silicon Valley.
1:23 Speaker 00 They first teamed up in 1992 at Palm, which is one of the pioneering, I don't know, smartphone companies. Where Donna was the CEO and they also co-founded Handspring, a company that made the Trio smartphone. Donna and Jeff, welcome to the show. Happy to be here. And you're still together after all these years. Amazingly enough.
1:41 Speaker 02 Yes, that's true.
1:42 Speaker 00 I know. How long have you been working together?
1:44 Speaker 02 Since 92.
1:45 Speaker 00 1991, yeah.
1:46 Speaker 02 92.
1:47 Speaker 00 92, so how many years? What is your secret to your partnership's success?
1:51 Speaker 02 I don't know. What would you say, Donna?
1:53 Speaker 00 I would say tolerance.
1:57 Speaker 03 You know, I think we complement each other very well. Jeff has always been the driver on the technology, the market side. I've been much more the business person, if you will, and looking at how to create businesses around these ideas. And I think it's that complementary nature that has been able to persist through several different companies.
2:17 Speaker 00 Yeah, absolutely.
2:18 Speaker 03 And even one nonprofit. I helped Jeff when he created the Red Road Neuroscience Institute as well. So, we just, I think, have a great respect for each other's strengths and able to work together. Plus, we have fun.
2:28 Speaker 00 Yeah. It's unusual in Silicon Valley, longtime partners.
2:30 Speaker 02 Is it? I guess. Is it? Maybe. I mean, you think about the partnership that led Intel in the beginning.
2:36 Speaker 00 That's true.
2:37 Speaker 02 That lasted for a long time. But maybe it's not too common now.
2:39 Speaker 00 Yeah. So, let's talk about Numenta. Now, we had Code Conferences here. We got a lot of talk out of, a lot of statements from a lot of people about artificial intelligence. On one side, it was sort of the Facebook people and others talking about the happy, shiny future of artificial intelligence. On the other, it was Elon Musk saying, "At best, we can hope to become housecats of computers," which was somewhat disturbing. And at worst, of course, they'll kill us. But you guys have been at this. Explain Numenta for the listeners.
3:03 Speaker 02 Oh, wow. This actually goes back a long, long time ago, when I was right out of college. And I fell in love with brains. And I wanted to, I just wanted to dedicate my life to...
3:14 Speaker 00 To people's brains.
3:15 Speaker 02 Yeah, the human brain.
3:16 Speaker 00 Inside the skull.
3:17 Speaker 02 How it works. And just as a scientific endeavor, humanity to understand how we think and how our brains work, because that's what humanity is.
3:25 Speaker 00 So, this was, you were not a medical student?
3:27 Speaker 02 No, I actually ended up being a graduate student at Berkeley in biophysics. But anyway, it goes back a long time. And it turns back, at that time, you really couldn't do this at all. It wasn't possible scientifically to do this. So, that's when we started Palm. It was like a, it was something to do for a few years until I could get back into neuroscience, believe it or not. That's how I viewed it. And I think the first day I met Donna, I told her this.
3:50 Speaker 03 That's what he told me. When I interviewed with him for CEO of Palm, he basically explained that the long-term vision was to be able to get back to understanding the brain and how the brain works.
3:59 Speaker 00 Right. The greatest computer of all, right?
4:01 Speaker 02 Yeah, of sorts. And so, that's, it's a long, long history. Personally and Donna as a partnership, understanding this and so the time at Palm and Handspring was really all exciting, we're working on mobile computing and all stuff, but everyone knew that I wanted to get back into this. So about 15 years ago, or 13 years ago, I can't remember now exactly, I left the mobile computing space. I started this nonprofit that Donna mentioned a moment ago, Redwood Neuroscience Institute, which was just focused on brain theory, cortical theory, and Donna was on the board and helped, she was the chair of the board and helped organize that. I did that for a number of years. That research entity had a close partnership with Berkeley, and so it's now continuing on the Berkeley campus, it's part of the Helen Wills Neuroscience Institute.
4:43 Speaker 02 But we at that time left and started Numenta, which is sort of a hybrid between a for-profit company and a private research lab. And so we continued doing the neuroscience research, but also building technology as a spinner off.
4:56 Speaker 00 Donna, talk a little bit about the business part of that. Now, what year did you start Numenta and what was the funding?
5:00 Speaker 03 We started Numenta in, I think about 11 years ago, so...
5:05 Speaker 00 So early, early, before a lot of people were thinking about...
5:07 Speaker 03 Oh, absolutely. Though, if you look at the history of, you know, what's called artificial intelligence, it goes back 50 years.
5:13 Speaker 00 Oh, yes, absolutely.
5:14 Speaker 03 You know, a long time, people have been struggling with this issue of how can we make machines that are intelligent. But, you know, we've talked a lot about what's the right business model for us along the way. When we created Numenta, we debated, should we make it a nonprofit? But the reason we decided to make it a commercial entity was because we have a huge belief in platforms, we've seen that through our days at Palm, at Handspring, my earlier career at Apple, and seeing how if people can add value on top of what you're doing and creating specific solutions, you can have a much, much bigger impact. And if people are excited about that, it's a way to get more support for what you're doing.
5:48 Speaker 02 It's kind of a recognition that ultimately this is going to be a huge industry.
5:52 Speaker 03 Right.
5:53 Speaker 02 And just to acknowledge that fact and say, how do you play a role in that? And no, don't ignore that component of it.
5:59 Speaker 00 Right, but when you started it, you know, from a business point of view, was it difficult? Because people, artificial intelligence and machine learning have been around for a while, but it really is just recently.
6:07 Speaker 02 Well, it was funny, because Donna, when we started, I think you led an effort to decide what terms we're going to use. And you did focus groups on AI.
6:14 Speaker 00 So, what happened?
6:16 Speaker 02 Well, it was the worst thing in the world.
6:18 Speaker 00 Why? Tell me.
6:19 Speaker 03 Well, AI has a negative kind of a meaning from back in the early days. You may even remember, they talk about the AI winter that we went through.
6:28 Speaker 00 The Terminator didn't do so much more.
6:30 Speaker 03 Yeah, there's been a lot of sci-fi movies that have incorporated AI. But, you know, there was a real disappointment. People in the early years tried to do very interesting things and realized that they were up against a wall, and they really didn't know how to move forward. And I think-
6:44 Speaker 00 Well, talk about those early days, meaning, what were they trying to do? What was the concept?
6:47 Speaker 02 Oh, this goes back into the 50s, when the term AI was termed, came up with, coined. And at that time, they were saying, oh, this program-
6:55 Speaker 00 That was at MIT, right?
6:56 Speaker 02 Yeah, MIT and other places as well. But they had the AI lab back at MIT. And so there was all these promises, all this stuff is going to be amazing, we're going to solve all these things in a few years, and it fell flat. And then in the 80s, there was a research, there was a huge interest in neural networks, and there was all these conferences, and companies, and funding, and all this stuff was great, and all that went away. And so we're actually sort of like on a third wave right now, where there's a new hyping of interest again that's happening here. But so this has been going on and on again, where people get excited, and then all of a sudden, ah, it didn't really turn out the way we thought it would be. They get excited, and then it turned out the way we thought it would be. We still think, right, the current wave of AI right now, AI right now is built on the wrong technology base. It's not even close to AI. It's not even close to what human intelligence is.
7:40 Speaker 02 So, you know, we're going through another wave of interest right now. But it's interesting how, when we started the company, the public association with the term AI was very, very negative. And now, here it is today, everyone can't wait to call themselves an AI company.
7:53 Speaker 00 Yeah, so, let's talk first about the, again, the business of it, Donna. What is the actual business numenta, and then let's talk about what the product is?
7:59 Speaker 03 We've looked at a few different business models through the years. We've tried enterprise sales, we've created a product-
8:03 Speaker 00 Meaning what? What would that do?
8:05 Speaker 03 Like going to, directly to the, a big company, and trying to solve a problem for them with consulting services, as well as with our technology.
8:13 Speaker 00 Sort of what IBM is kind of trying to do now.
8:15 Speaker 03 Yeah, a lot of big companies do this kind of approach. Another approach is a product approach. We created a product called Grok for AWS, which monitors servers on a network and gives you anomalies. That was actually well-received, but we found that- Where our heart was, was the core technology.
8:31 Speaker 00 Right.
8:31 Speaker 03 And every time we went down that path, we felt we didn't want to invest resources in what it took to get it from where it was to a real vibrant, big, deployed product. So at this stage, our business model, which I think is, is very interesting, is modeled after a tech transfer model of a university or a research lab. And the idea is that we work with partners and those partners take the products into a specific domain. They add the part on top of the technology that we don't do. They have domain expertise. For example, that Grok product has been taken over by a company. They've brought it to the market. They're building out a whole product around it. And we're supporting them in doing that.
9:08 Speaker 00 So the core technology, and talk about that, Jeff. What is the?
9:11 Speaker 02 Um, yeah, so, I mean, we start, really, we start with a scientific mission. And that is how to reverse, reverse engineering the neocortex.
9:19 Speaker 00 Right.
9:19 Speaker 02 I mean, in detail.
9:21 Speaker 00 But not the stupid parts.
9:22 Speaker 02 Uh, well, this, I like to say.
9:24 Speaker 00 Hey, the Trump parts. I call it the Trumpian.
9:26 Speaker 02 The brain can be easily divided into the new parts, which is the neocortex, literally meaning new layer. And the old parts of the brain and all the things we think about of emotions or anger or lust or fighting, and these are all the old parts of the brain. The neocortex is sort of the rational part of the brain that builds a model of the world and understands it. So if you want to think about like math and physics and, and humanity, all these things, it's the new part of the brain. If you think about the baser parts of humanity, that's the old part.
9:50 Speaker 00 So the brain's at war with itself.
9:51 Speaker 02 Uh, it is. It's like, you know, you have those, those old pictures of the cartoon with a devil and an angel on two sides, which really like that.
9:58 Speaker 03 Just like when you see that cupcake and your brain is at war with itself.
10:01 Speaker 02 The neocortex says, I know I'm not supposed to eat the cupcake. It's obvious. I've read all the reports. I understand why it's bad for you. But the other part of the brain goes, eat, eat, eat.
10:10 Speaker 02 So anyway, we're focused on the neocortex, which is seventy-five percent of a human brain.
10:15 Speaker 00 Okay.
10:15 Speaker 02 It's not.
10:15 Speaker 00 The size, you mean?
10:16 Speaker 02 Yes, physically sized. It's what makes us unique.
10:18 Speaker 00 And develops in?
10:20 Speaker 02 Uh, well, all mammals have a neocortex and non-mammals don't. So it's, it's a characteristic, unique to mammals. And, um, in humans, it's particularly large and particularly, um, well-developed. So anyway, it turns out that it's an amazing organ and everything that you and I do as humans, our language, our, our art, our engineering. These are all done by the neocortex. And it, the surprising thing about it is the neocortex uses a common method for all these things. This was discovered fifty years ago. So the way you see and the way you hear and the way you think and the way you do language and the way you engineer things is all the same basic learning algorithm. Hard to believe, but it's true. So our goal is to understand that and to reverse engineer that, understand exactly how
11:02 Speaker 02 the neocortex works at a biological level and a mathematical level and a comprehensive theory. That theory is a biological theory, but it also gives you the functional basis of how this thing works, and then you can build machines that work on that principle. And that's a very different approach than what other people use for AI.
11:19 Speaker 00 So explain that, Donna. What do you mean? What's the other approach for AI?
11:23 Speaker 03 The other approaches are much more mathematical, statistical.
11:26 Speaker 00 Teach a computer to be a human.
11:27 Speaker 03 And they're often focused on a specific problem. How do we solve this problem? This is much more of a generalized approach. It's about a learning machine. It's like your child is born and the brain doesn't have stuff in it. And then it learns from what it's exposed to. So the brain could learn Japanese, it could learn English, it could learn Swedish, whatever. It learns from what it's exposed to. You're not programming a Japanese brain. And so this idea is a much more generalized approach.
11:52 Speaker 00 So the way it's approached right now is that you're programming it, that you're programming a computer.
11:55 Speaker 02 No, the current interest in AI is not programmed. That was the old way. Today, it's learned, but it's a very simplistic learning method. You know, when someone wants to look at a picture and say, what's this picture? Facebook does this, Google does this. They train it on millions and millions of pictures and say, okay, which is essentially like, what's it closest to? Humans don't do that at all. We build a model of the world. We understand, we're in a conference room right now, and we understand what all the objects here on the counter are, the microphones, the calculator, the walls and the windows. We have this very intricate model of the world that we've learned through interacting with it. It's a very, very different approach than just say, okay, we're going to recognize a picture, let's train it on 16 million pictures, and we'll have a picture recognizer. I'm trying to, it's a little...
12:38 Speaker 00 Complicated than that, I get it.
12:40 Speaker 02 It's a little more complicated, but basically, that's the truth. So there's very, very different approaches, and I think Donna mentioned a moment ago, if you're focused on solving a particular problem, you'll engineer a solution to that problem. What biology tells us that everything we think about humanity and humans and our capabilities are really, it's a learning process based on the same algorithm. And that algorithm is us moving through the world and touching things and so on. So, you know, we believe that those principles, and they're very different, those principles by which the brain works, will be the foundation for what we think in the future of AI.
13:10 Speaker 00 So when you're trying to essentially rewire, create a brain, why the fear around computers right now? I mean, Donna, you know, Elon talked about it.
13:19 Speaker 03 Well, Jeff wrote a great article for Recode that you guys published on this question. I think it's like any new technology, it's fear of the unknown, where can it go? And let's face it, every new technology can be used for bad as well as good. I mean, cars give us great transportation and cars give us accidents as well. And so any-
13:36 Speaker 00 But this is a little different. This is more, I mean, cars are passive thing that requires a human in it to do that accident. Maybe.
13:42 Speaker 03 But we've seen it with computers, you know, early computers, of course, instilled fear in people. That's why those sci-fi movies that you talked about, Hal, whatever. So I think it's, there's always some great fear and I think part of that is justified and these things do need to be regulated and understood. And the example I like to use is drones. I mean, drones are so interesting, I think they're going to open up all sorts of interesting areas of business and ideas and viewpoints of the world. It's fascinating, but they need to be regulated. They can't be flown near airports.
14:10 Speaker 00 They just were today. The FAA just put out a series of regulations.
14:13 Speaker 03 I didn't realize that.
14:14 Speaker 00 Yeah, they did, in fact. All right, we're going to talk some more about Numenta and the brain in just a second, but first a moment for some advertising. We're here talking with Jeff Hawkins and Donna Dubinsky, a longtime Silicon Valley Entrepreneur Partnership, who is now working on Numenta, which is figuring out the human brain. Casper made a perfect mattress and sells it directly to consumers to save you money. The Casper is an obsessively engineered mattress at a shockingly fair price. It combines springy latex and supportive memory foams to create an award-winning sleep surface with just the right sink and just the right bounce. Time Magazine named it one of the best inventions of 2015. Shipping to both the US and Canada is completely free and there's a 100-day, risk-free trial and return policy. If you don't love your Casper mattress, they'll pick it up and refund everything.
14:58 Speaker 00 These mattresses are made in America. Get $50 toward any mattress purchase by visiting casper.com/recode and using the promo code RECODE. Stop paying for the mattress industry's inflated prices. Go to casper.com/recode, R-E-C-O-D-E, and use the promo code RECODE. Terms and conditions apply. Today's show is also brought to you by SoFi. SoFi is a new kind of finance company. They're helping people get out of student debt faster while saving them a lot of money. Refinancing student loans with SoFi saves members an average of $19,000. SoFi even partners with companies to help free their employees of debt. See how SoFi can help you at SOFI.com. Terms and conditions apply at SOFI.com/legal.
15:42 Speaker 00 We're here with Jeff Hawkins and Donna Dubinsky, two longtime Silicon Valley entrepreneurs who are now working on Numenta. They're most famous for their forays into the smartphone business with Palm and Handspring and others, but now they're working on recreating the human brain. I think that's about right. Jeff, correct?
16:00 Speaker 02 Reverse engineering.
16:01 Speaker 00 Reverse engineering.
16:02 Speaker 02 And building machines.
16:03 Speaker 00 So, let's talk a little bit about the science of this. What is difficult about doing that? Because it's really hard to, I mean, people still, after all these years, don't know a lot about how the brain works.
16:11 Speaker 02 Yeah, it is a very difficult problem, but we need to separate out things that are hard to figure out from things that are hard to understand.
16:18 Speaker 00 Okay.
16:19 Speaker 02 And so, the history of science is still the things that at first seem very complicated and difficult, but then afterwards, it's, once you understand, it's not so bad. That's the way the brain is too. And one of the reasons it's so hard to figure out is that it's a very complex organ made of cells, and it's very hard for neuroscientists to actually get the data properly. So, we work with lots and lots of partial studies of all this information, and there's thousands of papers published every year that are bits and pieces, and we have to sort of assimilate all that. But if you really go through the effort, you can do it, and you can make progress. And we're making excellent progress in this. The problem of how the neocortex works will be solved, it's partly solved already, and it'll be completely solved, hopefully, in our lifetimes.
16:58 Speaker 00 What is the most interesting thing that's going on in that area right now, from a research point of view?
17:02 Speaker 02 In the whole field of research? I think one of the things is, it's the recognition that this is a problem that can be solved in the near term is elevated. It used to be that people thought this was like an impossible thing, it would take 500 years.
17:16 Speaker 00 Because of so many connections, so many...
17:17 Speaker 02 It just seemed, just seemed, we didn't have enough data, and it's hard to get the data. There's been so many new techniques for extracting data from brains now, that we have an embarrassment of data and lack... of theory. So this is a field which is data embarrassingly rich, but theoretically very poor. There's very few people in companies or labs that are really doing sort of this theoretical integration. But you know, if you think about like even all kinds of people around the world have decided that understanding how the brain works is one of the next big things. It's a it's a grand challenge in the national kind of engineering. We have a grand challenge in Europe called the Human Brain Project. In the United States, we have the Brain Initiative. These are sort of things that we can solve this problem in the next
17:58 Speaker 02 10 years or so. And so the elevation of like, okay, we've got the data. We understand this is one of humanity's most interesting problems of all time, forever, is how our brains work. And this is the time that this can be done.
18:12 Speaker 00 So don't talk a little bit about the other companies that are addressing this because now everybody seems to have jumped in heavily. They're all around machine learning. Can you just tell explain for people what the difference, your machine intelligence?
18:23 Speaker 03 Well, there's a lot of companies working on machine learning and solving a lot of important problems. So I
18:28 Speaker 00 So break that down for people who don't understand it quite proper.
18:31 Speaker 03 Machine learning is more the technique that Jeff described earlier of using labeled data sets and trying to under, you know, take away the statistics from it in order to solve a specific problem. And it's been successful at doing that. So you see the cat, the dog, cat kind of things is a problem that that people couldn't do at all 10 to 20 years ago. And now they can do a relatively good job. I think it's a interesting you go on Facebook and how it identifies some of the faces. And so I think there are those are approaches that are very good at solving specific problems.
19:04 Speaker 00 With heavy data.
19:05 Speaker 03 They've made a lot of progress. They are highly reliant on labeled data, not just heavy data, but labeled data. Somebody has to go through and say, this is a cat, this is a dog. You can't have unlabeled data and be able to figure it out. And they are highly reliant on, you know, significant computing power, which has added a lot to solving these problems. But it's very different than our approach. It doesn't mean that it's not interesting and important. It's just different.
19:27 Speaker 00 So why are people scared of it right now? Because again, we had that sort of dichotomy at code and then it degenerated into it doesn't really matter. This is all a computer simulation anyway.
19:37 Speaker 02 Yeah, that was
19:38 Speaker 00 Which means there's some big brain up there.
19:39 Speaker 02 That was an amazing claim by Musk.
19:41 Speaker 00 There's many people that agree with him. Just so you know, they've come out of the woodwork and they're prominent people.
19:46 Speaker 02 I think one of the things that's going on here is there's very smart people. Whether it's Elon Musk, or Jeff Bezos, or you know, Bill Gates, or Stephen Hawking, who have said, "Oh my gosh, this is really dangerous." But those people, as smart as they are, actually have no idea how brains work. They don't have any sort of base theory about what intelligent is and what a machine-
20:07 Speaker 00 They're also forgetting that humans are very dangerous.
20:09 Speaker 02 Yes, so, well, maybe, but it's based on a, it's based on really a sort of a lack of knowledge in the field. And so they extrapolate from what people are doing with machine learning today, and add some hyperbole to it, and stir it around a bit, and some science fiction, and all of a sudden you imagine these things becoming crazy and alive and taking over the world. It is so far from that, it's crazy, ridiculous.
20:30 Speaker 00 Meaning?
20:31 Speaker 02 Well, A, we're not close to doing that, B, the current technology is even, not even on a road to true intelligence. It's not, the current approaches people are using in machine learning are not on a path to building intelligent machines. We think the approach we are, is. But if you understand what we're doing, and earlier we were talking about like the old brain and the new brain, and it's just not gonna happen. So it's, the fears are based on sort of lack of knowledge about the field, from people who are very smart, and therefore they think, since they're very smart, they can apply their smarts to this field they don't really know much about.
21:01 Speaker 03 Right. I think it's worth, this distinction of old brain, new brain is very important here, because the old brain part is what's got the motivations. So, I think they keep confusing that there's some old brain-ness somehow in this work, that these machines would have desires, and goals, and want power, and have needs, and they won't. They won't have any of those things. Nobody really is replicating the old part of the brain, that part. I mean, unless you program a machine to do that. It's the logical part, which is just very different, and I think it's, it's hard for people to not separate those two things.
21:35 Speaker 00 Because they're focusing on their own, this is what I would do.
21:37 Speaker 03 That's how you define yourself as a human. You think of both those pieces when you're a human, and that's true, those are both important pieces as a human, but that's not what the machine is gonna be.
21:46 Speaker 02 There was a, there was an unfortunate thing that happened many, many years ago. Alan Turing, the famous computer scientist- He proposed this thing called the Turing test. And he actually proposed it because he was for a different reason. He was tired of people asking could machines be smart? But anyway, we've we've accepted this Turing test as sort of the gold standard. Like, can it be like a human?
22:04 Speaker 00 Mm hmm.
22:05 Speaker 02 Can a machine? But that's not what intelligent machines are going to be like. They're not going to be human like at all. They're going to have the the new brain, the near cortex part.
22:12 Speaker 00 Mm hmm.
22:13 Speaker 02 But not the old brain, and they're not going to be humans. They're not going to live human lives and have human emotions and human experiences. We can. And so there's this idea that, you know, oh, a lot of people feel like the goal of AI or the goal of.
22:24 Speaker 00 Is to create a human. And then are they slaves? Are they? It gets into all those issues.
22:28 Speaker 02 But that's not going to be what the future machine intelligence looks like. It's going to be so different than that.
22:33 Speaker 00 So what does it look like?
22:34 Speaker 02 It's really just this part of the brain that. Here, I'll give you two examples of things that we can look forward to. Imagine you want to do mathematics.
22:41 Speaker 00 Mm hmm.
22:42 Speaker 02 This is a very abstract thing that's done by the near cortex. It's done on the same principles we use everything else that we do everything else with. We we didn't evolve to do mathematics, but our brains can do it. Imagine you can build a brain that is a million times faster than a human, never gets tired, and it's designed to think like mathematics. It's really tuned to be a mathematician. We could advance mathematical theories extremely rapidly or fast than we could otherwise. This machine doesn't going to do anything else. It's really designed to.
23:11 Speaker 00 Doesn't want to eat.
23:11 Speaker 02 It doesn't want to eat, doesn't want to have sex. It's not, you know, jealous that you've got the, you know, tuna sandwich and I got the whatever. And that's the kind of thing that we can look forward to. Machines are just brilliant and smart.
23:21 Speaker 00 All they do is work.
23:22 Speaker 02 But they do some, they apply these things in other ways. I'll give you one other example of you. I have time for that.
23:27 Speaker 00 Sure, absolutely.
23:27 Speaker 02 Um, this is another one. You know, a lot of people, including Elon Musk, wants us, you know, want to send humans to Mars. I think it's a great idea. But Mars, in its best day, is going to be worse than Earth in its worst day to live for humans. And.
23:39 Speaker 00 I get that sense.
23:40 Speaker 02 Yeah, and we we, if we're going to do this, we need to have, think of it this way, it's sort of an engineering robot. Engineering, robots that are smart at solving problems, like they have tools that break, they see materials that have, they're kind of, you take a really smart engineer and put them someplace and try to solve some problems and and figure out new solutions, these things. We're going to have to have machines that are like that. And they're not going to. Be human-like machines, but they're going to have to be very flexible and intelligent in solving problems because humans are not going to be able to go outside all day long in hostile environments trying to fix things.
24:12 Speaker 00 So Donna, talk about that idea because I had Anne Wojcicki on recently from 23 and we were doing DNA research and she was talking about this, similar to what you were talking about, is that these machines will say replace radiologists. You're talking radiology is an area of great controversy because a lot of radiologists don't read the things right and computers can do it. They don't get tired, they get it right, they can call your attention to the three of the 500 photos of the brain you're looking at, these are the three that are most important. There's a whole job creation thing is that they do replace people because they're better at it.
24:41 Speaker 03 Well, they're going to replace some things, but they're going to create some other opportunities and things.
24:46 Speaker 00 I know you all, all UTEC people say that, but I actually in this case think they're just going to replace them. Like what, you know what I mean? I know, oh, there'll be more, when the horses went away, there were cars. I'm like, except that this really does replace things and I can't think of how you, what is the job that humans have then? If they can do it.
25:01 Speaker 03 I don't know. If you just take, you know, Jeff's example of sending a smart thing to Mars and exploring there, somebody's got to create that, somebody's got to manage that, somebody's going to take the data from that. I think in the genetics area, I'm very excited about that. There's all that data that Anne and people like her are collecting and people are trying to figure out what the heck do we do with this? And to the extent that can be used for finding ways to address diseases that we can't address today, to understand them better and new treatments are developed and new protocols and so on, those are all opportunities for jobs as well as for improving humankind. So I think there's going to be as many opportunities as there are going to be, you know, disconnects, but
25:44 Speaker 03 people have to be retrained for those things. I mean, it's not to say there isn't dislocation. There could be enormous dislocation, but I think it means there'll be different things going forward.
25:52 Speaker 00 So what is the product then, Jeff? What are you making right now? This platform.
25:57 Speaker 02 Well, as Donna talked about earlier, we've been through a series of ideas about what our product is. And currently today, it's a licensing model. We have a very-
26:05 Speaker 00 Of the core technology.
26:06 Speaker 02 Yeah, we have, for a small company like ours, we have over thirty issued patents. And some of them we believe are fundamental, and we have many more coming along. We've had a lot of luck in this regard because we're really doing things that no one has ever done before. And so it's, there's not much prior art. So, in the end, from a business model point of view, it's all going to resolve around, revolve around some method of licensing technology. And Donna could talk to you, because she's developed multiple, we have an interesting strategy right now, combining an open source with licensing-
26:37 Speaker 03 Yeah, I think it's important to mention the open source, which is, kind of, you might scratch your head and say, well, how do you have a business model at all if you've got an open source-
26:46 Speaker 00 Yeah, I was just thinking that.
26:47 Speaker 03 Software. It's all there. We even put our research software there so that people can see what we're working on. It's the most open of the companies who are working in this space, I would argue. So, how can we do that and reconcile that with a commercial mission? Well, the way we do that is through, kind of, the magic of the open source license and how it works. It's open for any kind of non-commercial use, for academics, for researchers and companies, anybody. But at the point at which somebody makes a commercial product, some may, they can choose to do it under the open source license, but many will say, I don't want to build a commercial product on this open source, this specific open source license. I'd rather get a commercial license from you. So, in the end, we can do both.
27:26 Speaker 00 So you're licensing the core technology.
27:28 Speaker 03 So we can license the core, and the intellectual property, which may come into play, very importantly, in the future. Because one of the possible directions for this is that it's going to take a new kind of a semiconductor architecture, a new kind of a piece of hardware that's going to accelerate it in a different way than today's hardware does. And it could be, it ends up like, um, business model like a Qualcomm or a CDMA, where there is, um, specific hardware that needs to be-
27:53 Speaker 00 That you would make.
27:54 Speaker 03 Deployed, that we would license-
27:55 Speaker 02 We would, we would make that.
27:56 Speaker 03 We're not going to be a semiconductor company, but we've had- Quite a few actually talking to us and actively working and looking at porting our algorithms to their architectures. And I think that's going to be an important part of the business.
28:08 Speaker 02 That's an interesting story there. I've given numerous talks recently at semiconductor conferences, been invited to give keynote speech talks. And everybody in the semiconductor space is looking for what's next.
28:19 Speaker 00 Right.
28:19 Speaker 02 What's beyond Moore's Law, what's beyond the von Neumann architecture. And when you look at that space, they all eventually come around to looking at us because we have these really interesting, unique architectures which come from the brain. And they're different than everyone else is doing. And if you kind of buy the argument, this is the way it's going to work, then you've got a model for thinking about what new future semiconductor architecture is going to look like. So it's a fascinating sort of field right now where they're all trying to figure out what's going to come next.
28:47 Speaker 00 They are, because that era has ended. We're going to talk about that next. We're here with Jeff Hawkins and Donna Dubinsky. Their company is Numenta, which is a machine intelligence company. They're trying to reverse engineer the brain, not re-engineer the brain, although sometimes I wish you could do that. This episode is brought to you by GoCD, the open source continuous delivery server created by ThoughtWorks. GoCD is the best integration and deployment tool you've probably never heard of. It offers complete customization for your software's individual needs. There's no plug-in or workaround needed. GoCD just goes. Spend more time delivering and less time configuring. Commercial support and enterprise add-ons, including disaster recovery, are available.
29:29 Speaker 00 Download GoCD for free at www.go.cD. We're here talking with Donna Dubinsky and Jeff Hawkins about their company, Numenta, which is a machine intelligence company all around the brain. We've talked a lot about changing software, changing hardware, what's going on. You guys have had a long history in Silicon Valley. I'm going to bug you and make you pundits essentially of where you think you are. You're moving into an area that's all fresh and new. You're trying to create the next big thing and where chips are going and where other things are going. You've created several companies and while Palm and Hammerspring didn't sort of survive the change, they were at the pioneering edge of these things at all times. Can you talk a little bit about where you think we are in the cycle?
30:13 Speaker 03 Well, it's such a good point about, you know, looking back at these prior generations because I find it very amusing when now we walk around and people say, "Oh, the future is mobile."
30:21 Speaker 00 Yeah.
30:21 Speaker 03 It's like, no, no, no. Seeing that the future was mobile 20 years before it happened is the hard part.
30:26 Speaker 00 Right.
30:26 Speaker 03 Seeing that the future is mobile, once you're in the midst of it, that's not the future anymore.
30:30 Speaker 00 Yeah, right.
30:31 Speaker 03 And I think it takes a very particular skill set and mindset to say, "Where's this stuff going?" It takes a deep technology understanding. You have to know, when Jeff figured out the future was mobile, it was, he understood about screens, he understood about processors, he understood about memory. It wasn't just a, you know, a theory. It was based on real understanding of the technology. Intersected with an understanding of people and their uses and their needs and how this stuff could be applied. Not that you could totally predict it. You know, I like to say, when we created the Palm Pilot, we didn't imagine Uber.
31:05 Speaker 00 Right.
31:05 Speaker 03 So, you know, you don't know exactly where it's gonna go, but you know enough to know that it needs to go in that direction.
31:12 Speaker 02 Well, we did know, and we claimed publicly many times, that, that mobile computing would be the driving, dominant force of personal computing. And the time I used to say that, our VP of Marketing used to say, "Jeff, don't say that. It sounds crazy. It sounds crazy." But it turned out to be true, because at the time, it sounded a little crazy. But anyway.
31:29 Speaker 03 Yeah, yeah. It's like, you know, the future computing is mobile computing.
31:33 Speaker 00 Right.
31:33 Speaker 03 And people would just, you know, kick us out of their offices. They thought it was so nuts.
31:36 Speaker 00 Yeah, I think Walt and I were in there, even 10 or 12 years ago, saying Web 3.0 was all mobile, and we got so much pushback from lots and lots of people. They're like, "That's ridiculous." And we're like, "No, we think it is."
31:45 Speaker 02 You know, we're in a similar situation right now.
31:47 Speaker 00 Yeah, but this is...
31:48 Speaker 02 We're saying, "Look, the future of machine intelligence is gonna be based on, on brain algorithms." And that is a minority opinion today. I am so certain of it. I'm 100% certain of it. 20 years from now, we're all gonna look back and say, "Oh, yeah, that was obvious. Why would you think anything else?" But today, we have to...
32:04 Speaker 00 So, as the, the big brain here, what do you imagine? I mean, if here, where are we in the mobile development right now?
32:09 Speaker 02 Isn't it over by now?
32:10 Speaker 00 Is it over? I don't know.
32:12 Speaker 02 I don't know. Everyone has one.
32:13 Speaker 00 You're not interested in phones anymore.
32:14 Speaker 02 Not particularly.
32:15 Speaker 00 Why is that? Just been there, done that?
32:17 Speaker 02 Well, I think Donna and I are both attracted... I remember when I met Donna, she was like, she told me the story why she went to work at Apple early on because she said, oh my God, I had this epiphany that personal computers are going to be huge. And we are both attracted to like this next, like what's going to happen, what's going to be obvious 20 years from now. So, you know, mobile, it's like, it's okay, what, what do you have to do to be successful in mobile right now? How good is your camera? What, you know, how good is your, you know, your low light this? And how good is, you know, their screen OLED versus that? I mean, that's boring.
32:49 Speaker 03 It's incremental work now. It's incremental. I mean, maybe there's some big advances out there that are going to dramatically change it, but.
32:55 Speaker 00 Do you think the form factor is going to change?
32:57 Speaker 02 I don't. I, I'm happy with my form factor. Yeah. Why would you, why would it be any different?
33:01 Speaker 00 On your head or, you know.
33:03 Speaker 02 I don't think so.
33:04 Speaker 00 Google Glass, even though it's a bad execution, it might be a good idea.
33:08 Speaker 02 No, I, I, it's, well, like something will probably do something, replace it in the future perhaps, but at the moment, I'm, I'm, I don't think so.
33:15 Speaker 03 To me, the much more interesting aspects of mobile have been the applications on top of it, you know, things like Uber or, or using Airbnb on it or whatever. I mean, there's been fascinating, um, application software direction. So I think there'll continue to be that, things that we never thought of that people will do with these things that'll be world changing and important.
33:34 Speaker 00 Right. And what, so what, what, what do you imagine that happening? Or do you think that's also played out, Jeff, do you?
33:40 Speaker 02 Um, you know, it's, as Donna said earlier, it's.
33:43 Speaker 00 Some people feel the app economy is sort of going to collapse.
33:45 Speaker 02 No, it's so hard to predict these details. Um, what you can get right is the really big trends and then the details are never, you're never going to get.
33:54 Speaker 00 So what does a phone look like and?
33:55 Speaker 02 I, you know, I don't think about this space much. I'm very hesitant to.
33:59 Speaker 00 Well, it could be full of brain information.
34:00 Speaker 02 Well, I, all right.
34:01 Speaker 00 Smarter phones is what we're looking for.
34:03 Speaker 02 You know.
34:03 Speaker 00 Smarter smartphones.
34:04 Speaker 02 If you want to summarize Donna, my relationship for 20, some of yours, um, it starts out, Donna, understanding that before we met, that the future of personal computing, you know, the, the future of computing was personal computing. Then together was the future of personal computing was mobile computing. And now it's the future of computing is intelligent machines that work on the principles of the brain. This is going to be the dominant driver in technology going through the remainder of this century. And I'm so certain of that, so confident that, that it makes you put up with all the crap that happens every day and the things you have to deal with and the difficulty of our task. But I think that's, we are attracted, me from a technology side, or from a sort of development business side, we are attracted to the really big ideas.
34:49 Speaker 02 And so it's less interesting to be working on the old stuff.
34:52 Speaker 03 Right. For us, it's really, it's about impact in the end. You know, where can you invest your short time here on earth to have the most impact for the good? And, you know, what are you going to work on where you can move the needle? And to me, what's been remarkable about this journey is that you can actually, and it's rare that you can say this, you can actually see more clearly 10 years from now than 2 years from now. Usually, you're going to look out 2 years and say, okay, I know what's going to happen, but I don't know what's going to happen in 10 or 20. I think it's the opposite in this case. We believe this will be ubiquitous.
35:23 Speaker 00 So give me a use scenario. I mean, is it a car that really is?
35:27 Speaker 02 I tried to.
35:28 Speaker 00 Because the human brain is good at driving a car.
35:30 Speaker 02 Yeah, but that's, I think that's thinking too narrowly.
35:33 Speaker 00 Okay, all right.
35:34 Speaker 02 Right? The human brain is amazing.
35:36 Speaker 00 I have a small human brain.
35:37 Speaker 02 No, you do. You have a beautiful one, and just like everyone else's.
35:40 Speaker 00 Yeah.
35:41 Speaker 02 So the human brain is incredible for the number of things it can do that it didn't evolve to do. Engineering and designing this building, structures, materials, science, physics, mathematics, space exploration, the things that we do as a species that no other species has done, has ever done, or could do, is just incredible. And we didn't evolve to do these things. So we have this universal learning algorithm. This has been known for a long time that the brain works on, this universal learning algorithm, that can be applied to all these different types of problems that no one would ever guess. And it's that promise of going from, think about it this way, today's computers work on a Van Neumann or Turing architecture, which says you can program a machine to do anything.
36:23 Speaker 02 And the brain shows that there's another algorithm that says you can learn to do anything. And that anything is not, it doesn't overlap much with what computers do. Computers are not good at designing buildings, and doing physics, and mathematics, and poetry, and so on. But humans are, and we can build machines that learn to do an incredible number of things that we've never thought of. And here's another example. Humans are limited by our
36:46 Speaker 00 Physical.
36:47 Speaker 02 Physical presence. We have to have a body, and so intelligent machines have to have a body. We have a certain set of sensors. We have eyes, ears, and skin, and we have a physical embodiment of those. Intelligent machines don't have to be limited by that. You can build intelligent machines, have sensors that are incredibly different that look at nanoparticles, that look at outer space. They will think in N-dimensions. They will think in protein folding, where humans have, we have trouble. We have limits to what we can do because of the impedance between our sensors and our body and the things we want to learn about.
37:18 Speaker 00 So, give me something that, what would be an application? I know it sounds crazy, but regular people need to know, what does that mean?
37:23 Speaker 03 Well, I'll give you an idea of what some people are working on today. It's a wide variety. I don't think it's necessarily reflective of what are the big applications in 10 years. But, we have people working on looking at crop sensors and trying to understand what's going on in a field and can we improve what, you know, the farmer's able to do in the field. We have a lot of work being done in language. So, reading contracts with the machines so it can understand whether contracts are in compliance or not in compliance. Searching manuals to find answers on technical support problems that today, you know, they sit there with, you know, 80 manuals from various versions of the product and they simply can't look them up in a very easy way. We have one company that's monitoring beer taps and trying to get the quality right in beer taps and understand what the flow is.
38:07 Speaker 00 Using brain algorithms.
38:08 Speaker 03 Using brain algorithms. I mean, it's just because they don't have to be programmed, they're learning from the data, they're looking at the sensors, the sensory information, it's temperature, it's flow, it's, you know, a bunch of sensors.
38:19 Speaker 00 Any old bartender can do.
38:20 Speaker 02 Well, actually, Donna didn't mention, we haven't mentioned so far, but the brain works on time-based data. It's always a, things are always changing in time, where most machine learning algorithms are not time-based, they're static. And so, all these applications Donna just mentioned, actually, like the beer tap, that people are sensing the temperature and whatever, the flow, over time, every minute, every 10 seconds, whatever. And so, you model data over time.
38:42 Speaker 03 And they don't have to set up thresholds in advance. Oh, if it hits that temperature or that temperature, it learns what's normal, it learns Saturday night is different than Tuesday night. I mean, it's just a fun example, but, you know, we're monitoring servers, as I said, to say, you know, is there something unusual going on with a server? There's some, maybe some latency in the network or some bottlenecks that we can predict in advance. We're learning that temporal data comes in, we learn the pattern, we can make a prediction, we can find anomalies. So, so, there's a lot of applications today that are very interesting and they're You know, again, we're not focused on a specific application, but the amazing number of developers are working on this and looking at interesting ways to apply it.
39:21 Speaker 00 Sure. So if you're working on these brain algorithms that we make machines do, what about our brains? Do they get better? Do they learn? Do they become...
39:29 Speaker 02 Well, our brains, obviously.
39:30 Speaker 00 Does the old brain go away? It doesn't seem like it's going away in this season.
39:33 Speaker 02 Well, it doesn't seem like it.
39:34 Speaker 00 Seems like their old brain is having a resurgence.
39:36 Speaker 02 Well, uh, yeah, this is the thing. I actually want to write a book about this topic.
39:40 Speaker 00 It's a great idea, because literally, the old brain is doing pretty good this year.
39:44 Speaker 02 Because as a species, what makes us unique and what we want to be proud of in the longest term is things we do that are intelligent, intellectual, our ability to learn and understand the world. That's what makes us unique. Yet, we are saddled with our old brain as well. And I think, if I want to leave you with a very long-term thinking point of view, we have to, as a species, it'll be very interesting to see if we can transition to the point where we embrace more of our intellect and somehow shed the problems that we have inherited as an animal species. And that'll determine whether we survive as a species a long time or not. And the things we should be most proud of as a species are not that I kill this person over here, or steal their land, or do, you know, get the better food. It's going to be like, what did we leave from an intellectual point of view? What is our
40:28 Speaker 02 knowledge that we've achieved about the universe? And that's a problem for humanity to figure out. We, it's an interesting problem.
40:33 Speaker 00 Is the brain physically developing like that?
40:35 Speaker 02 Uh, the brain is already developed. It's not changing much now.
40:38 Speaker 00 Right.
40:39 Speaker 02 Uh, we continue to evolve as a species, like every other animal does. But right at the moment, you know, our technology is evolving so much more rapidly than any kind of evolutionary process in biology can handle. So the question is, can we, as a species, figure out how to embrace our better side, and our intelligent and intellectual side, and deal with the fact that we, um...
40:57 Speaker 00 Can we help that along? I mean, one thing Elon was talking about was neural lace, or whatever.
41:02 Speaker 02 I don't know what that is.
41:03 Speaker 00 Well, you stick it in the back of your veins, and then you suddenly get smarter. You turn your brain into a computer, I guess. In a serious computer.
41:08 Speaker 02 Well, I think it's unlikely.
41:09 Speaker 00 So if you can do math better, you...
41:10 Speaker 02 I think it's unlikely that we will, certainly you're not going to be able to upload your brain to the computer, and you're...
41:16 Speaker 00 I think they were talking about uploading things to your brain.
41:18 Speaker 02 I think that's going to be unlikely, too, unfortunately. But I wouldn't count on that.
41:22 Speaker 00 Okay. All right. Oh, damn. I was hoping that would work out, and I could finally understand quantum physics.
41:28 Speaker 02 No, maybe it'll happen, but we shouldn't count on it. We shouldn't sit around and say, "Oh, yeah, that'll, that'll solve a problem."
41:34 Speaker 00 I wanted, just to stick something in, I'd learn, like, karate in two seconds.
41:37 Speaker 02 Yeah, well, it'll go right to the matrix, isn't it?
41:38 Speaker 00 That's the matrix. That's exactly what they did.
41:40 Speaker 02 I don't think that's going to work.
41:41 Speaker 00 All right, yeah. Damn. All right. Thank you so much, Donna Dubinsky and Jeff Hawkins of Newmenta. They're working on how computers get brains and how brains evolve. Thanks very much. Thank you. If you enjoyed the interview as much as I did, be sure to subscribe to the show. Be the first to listen to future episodes or catch up on previous episodes, including some really fantastic interviews I've done with Yelp CEO Jeremy Stoppelman, 23andMe CEO Anne Wojcicki, and JP Morgan Chase CEO Jamie Dimon, just to name a few. And if you're one of the people celebrating the Cleveland Cavaliers NBA championship, you can hear my colleague Kirk Wagner interview LeBron James' manager Maverick Carter. All those interviews and more are at recode.net/decode. Now that you're done with this, why
42:23 Speaker 00 not try one of our other podcasts? Recode Media with Peter Kafka comes out every Thursday. On Fridays, I host Too Embarrassed to Ask along with Lauren Good of The Verge. And on Recode Replay, you can find audio from our events like the Code Conference, Peter Kafka's Code Media, and Jason Del Rey's Code Commerce. Thanks for listening, and thanks to our sponsors, Casper, SoFi, and ThoughtWorks. Also, thanks to Digital Media, which distributes this show. This has been another episode of Recode Decode. Remember to subscribe to the show and leave us a review at iTunes.com/RecodeDecode. I'll be back here on Monday with another great guest. Tune in then.
42:56 Speaker 01 This has been Recode Decode, hosted by Kara Swisher, powered by digital media. For more hard-hitting interviews with insiders from the worlds of tech, media, and politics, subscribe to Recode Replay on iTunes. Featuring candid conversations with leading voices like AOL CEO Tim Armstrong, Goldman Sachs' CIO Marty Chavez, the team behind the hit TV show Empire, Shark Tank investor Mark Cuban, and presidential candidate Hillary Clinton. They're all on Recode Replay.
Previous Recode Decode: Daphne Koller, president, Coursera Next Recode Decode: Gary Zenkel, President, NBC Olympics Page updated October 25, 2026 · transcript and summary are machine-generated