All Quotes

"What gets me excited about this space is the convergence of all these different trends together, so when you think about data, new channels, and artificial intelligence, the ability to then personalize for our customers and give them relevant, meaningful interactions"

"I recognized the power of when somebody comes into your orbit, how you react to them and what they see sets the tone for the rest of their experience"

"We exist because our clients still want to do business with us"

"As part of working with them, we owe them that transparency and that responsibility as a brand. As marketers who are telling the stories of the brand and who are engaging with our consumers and clients, we have no choice"

"We're very careful about how we spend our marketing"

"We have very sophisticated models. There are very few things that don't go through a rigorous exercise and funnel. And in a world with all the digital tools we have, we can do that"

"We know experiences are a way to add value, which can be a competitive advantage"

"The reality is, our mass-affluent consumer has pretty low awareness that JPMorgan Wealth Management can serve them"

"They may have a credit card with us, a savings account with us. We talk to them about things you wouldn't imagine a financial services company would talk about: We may talk about home lending; they talk about home ownership. It's a way to provide another service to them as they think about their lifespan and well being"

"Whatever your first issue of concern, media had better be your second, because without change in the media, progress (in your primary area) is far less likely."

"[T]he closer a story gets to corporate power and corporate domination of our society, the less reliable the corporate news media is."

"Homer and Shakespeare lived in very different times, more than two thousand years apart, but they both captured the same idea; we are not all equally intelligent. I suspect that anyone who has failed to notice this is somewhat out of touch with the species."

"The study of intelligence is not an isolated academic topic; our intelligence has social consequences. We want our leaders to be intelligent, and exhibit concern if we think they are not. There were politically motivated attacks on the intelligence of Presidents Lincoln, Truman, Harding, and Ford. Seri¬ ous concerns about mental competence"

"My training was basically as an ecologist—for is largely applied ecology—and, since serving in Ghana for 14 years in the 1930s and 1940s. I have been back 14 times to East, South and West Africa. Even when I started work, it was already clear that the arid land such as the could not be safely occupied other than by s, who with their flocks and herds of grazers and browsers can use small patches of mixed vegetation scattered over a wide area, seldom causing anything more than limited short-term degradation, for the nomads had worked out a '."

"... by the third millennium the ns had developed a type which is shown in art of that period. Egyptian monuments of at least one thousand years earlier show dogs very like the modern and others like the ."

"Some animals lend themselves to use in idioms—the goose is a good example of this. As long ago as 1547 a simpleton was being referred to as a goose; presumably a domestic goose, for a wild goose is anything but a fool."

"When Attenborough joined the , in 1952, what were then called “animal programs” were presented by the superintendent of the London Zoo, a Mr. George Cansdale. Once a week, Mr. Cansdale would bring a selection of more or less compliant creatures to the BBC studio. He “put them on a table covered with a doormat,” Attenborough remembers in the preface to his new book, Adventures of a Young Naturalist, “where they sat blinking in the intense lights.”"

"the best decision I have ever made."

"It doesn’t really matter what you decide to do, but until you have tapped into what drives you every single day to get up, then the company isn’t really going to matter, right?"

"I’m inspired by the younger generation, who deserves the opportunity to live in a clean, healthy, and safe environment. This is why I participate in, and support, initiatives and efforts focused on mentorship and development of youth who are passionate about climate change and clean energy. The ultimate goal is to develop leaders who focus on sustainability and conservation of our trust resources."

"a Neural Network is a set of mathematical expressions that are really good at recognizing patterns in information, or data."

"A NN accomplishes this through a kind of human emulated perception, but instead of seeing, say an image, like a human would, the NN expresses this information numerically contained within a Vector or Scalar (a Vector only containing one number)."

"MNIST can be seen as the ‘Hello World’ dataset because it is able to demonstrate the capabilities of NNs quite succinctly. The dataset is made up of handwritten digits, which we will train our NN to recognize and classify."

"To facilitate our implementation we are going to be using the Keras framework. Keras is a high-level API written in Python which runs on-top of popular frameworks such as TensorFlow, Theano, etc. to provide the machine learning practitioner with a layer of abstraction to reduce the inherent complexity of writing NNs."

"She taught me for only one course, but for some reason, I was always in her presence,” Owens said in a 2014 documentary about her life and career. “She had finished North Carolina Central herself and she had gone on to get her Ph.D. from the University of Michigan. The fact that she had gotten a Ph.D.—she became my role model."

"but for some reason, I was always in her presence,"

"was interested in a more inclusive environment at Duke,” Owens said in 2015. “And I just happened to be in the right place at the right time."

"Go for it!"

"My research explores human vision through the creation of computational models of visual processing and observations from perceptual experiments."

"My work focuses on how we analyze the three-dimensional structure and movement of objects in the environment through the integration of cues such as image motion and stereo vision, and use this information for the visual guidance of tasks such as steering a vehicle and catching a ball."

"As we move through complex scenes of the sort that arise when we drive or play sports, we quickly perceive distinct objects that we can recognize, track, or interact with in some way. A critical step in this process is to locate object boundaries in the rapidly changing image that reaches the eye."

"I teach an advanced course on vision that combines the study of state-of-the-art computer vision systems with observations from psychology and neuroscience about the visual processing strategies used in biological systems."

"I serve on the advisory committees for the Neuroscience program and the Cognitive and Linguistic Sciences program. Through these interdepartmental programs, students engage in a broad interdisciplinary study of the nervous system and the mind."

"They can pursue an area of concentration that includes coursework in computer science and research that incorporates the perspective of computational modeling."

"didn't go into office to get money–to be rich–I went into office to give a service."

"It’s really simple: I’ve tried to do what my mother and father told me and what the good book says. … And I got this from my daddy, regardless of how high you go, always keep your feet on the ground. So I’ve tried to live by that."

"We were adamant about changing things down at City Hall,” she said. “… I worked hard for the East Ward because I look like them and they needed it."

"In individual emotional development the precursor of the mirror is the mother's face."

"When girls and boys in their secondary narcissism look in order to see beauty and to fall in love, there is already evidence that doubt has crept in about their mother's continued love and care. So the man who falls in love with beauty is quite different from the man who loves a girl and feels she is beautiful and can see what is beautiful about her."

"At the end of the day, we are all human beings and we all know what it means to be in love, to feel betrayed, to feel misunderstood."

"It is strikingly plausible that by 2027, models will be able to do the work of an AI researcher/engineer That doesn’t require believing in sci-fi; it just requires believing in straight lines on a graph"

"Imagine if when asked to solve a hard math problem, you had to instantly answer with the very first thing that came to mind,” he writes. “It seems obvious that you would have a hard time, except for the simplest problems. But until recently, that’s how we had LLMs solve math problems. Instead, most of us work through the problem step-by-step on a scratchpad, and are able to solve much more difficult problems that way. “Chain-of-thought” prompting unlocked that for LLMs. Despite excellent raw capabilities, they were much worse at math than they could be because they were hobbled in an obvious way, and it took a small algorithmic tweak to unlock much greater capabilities."

"GPT-4 has the raw smarts to do a decent chunk of many people’s jobs, but it’s sort of like a smart new hire that just showed up 5 minutes ago: it doesn’t have any relevant context, hasn’t read the company docs or Slack history or had conversations with members of the team, or spent any time understanding the company-internal codebase. A smart new hire isn’t that useful 5 minutes after arriving—but they are quite useful a month in! It seems like it should be possible, for example via very-long-context, to ‘onboard’ models like we would a new human coworker. This alone would be a huge unlock."

"Look. The models, they just want to learn. You have to understand this. The models, they just want to learn…"

"Well, no. There are many other changes. I mean, of course the speed and the memory size of the computer is phenomenally different, I mean, but the parallel isn’t that the distributed computing is also tremendously big difference and the display systems. At that time we used – when I worked at Stanford we didn’t have even a raster display. It was all just vector displays that displayed only the contours. At the end of my stay at Stanford we got the first raster display, which allowed you to display all the pixels, but in the beginning it was all just displaying contours."

"I don’t think they had internal divisions or titles or anything. My advisor was a neuroscientist, called Emilio Bizzi. He was interested in how the brain worked, obviously, but in particular how the brain controlled movement and how you coordinated movement. At that time robotics was starting to happen at MIT, I guess in a more visible way, and he reached out to two people, a guy named Neville Hogan in the Mechanical Engineering Department and a guy named John Hollerbach who was at the MIT AI lab and they collaborated trying to ask the question, “What can robotics tell us about how the brain works?” Because clearly the body is a robot, so the brain has to be a robot controller and maybe some of the theory that’s being developed for robot controllers could be useful, and that was the context in which I oper – I was in graduate school. First year I studied human movement, second year I – yeah, it’s all coming back to me now – I studied more human movement or human movement in a slightly different way, I’m happy to go into more detail on that kind of stuff later, and we also did some monkey studies and then at that point I decided, “Oh, let’s go try to understand robots, they’re a lot simpler,” and that’s what I did. So, my thesis was on the question of, “If you pick something up, how do you know how much it weighs?” The example I loved to use was, “When you reach into a refrigerator and pick up the milk, how do you know it’s frozen?” Now, that’s not a question I actually answered in my thesis, but that was a motivating question."

"So, a big underlying question at that time was it seemed that robots or the way to make robots useful was to develop something called a general purpose controller, which meant that whatever you were asked to do you had some planning machinery that basically operated in the same way and just did it, and a signature of that kind of approach is no matter how you moved, you’re always gonna do things in the same way, so maybe you move in straight lines or something like that. So, the obvious thing to do is to then look at humans and apply that kind of test and see when we ask humans to do different things, is there some commonality or are they coughing up some sort of special purpose solution to everything they do? And it turned out most of the time you ask people to move from point A to point B, and we developed some measurement system to do that, they move with a hand in a roughly straight line and more importantly with what’s called a bell-shaped philosophy profile, so they start up in a particular way and they slow down symmetrically. So, it’s a nice smooth movement. And that was actually an argument that whatever’s going on in your brain, there seems to be a general purpose movement system, at least for move from point A to point B in a human with your arm. Later, the bell-shaped philosophy profile was extended to all sorts of things like jaw movements, tongue movements, leg movements, so it actually seemed that whatever’s going on, it generalized across limbs as well."

"I’m going to focus on what I think is the last part of your question because I could parcel that as two questions, you know, how did the study of humans affect what we did with robot? At about that time there was a lot of emphasis on industrial robotics and doing things fast and so the idea was if you want to go from point A to point B, you want to hit your arm as hard as possible, turn the motors maximally on, accelerate maximally, get up to your maximum velocity, do what’s called “coast,” move at a constant speed, and then when you’re close to your target you want to reverse the process and blast your arm with your motors in the opposite direction so you slow down as fast as possible and then you stop on a dime. Okay, the public image of robots at that time was that they move jerkily, like this and if you ask anybody to imitate a robot, you’ll see them do jerky movements and in fact I believe there’s even a robot dance called “Do the Robot” which is all about that. That had nothing to do with robots, it had everything to do with this choice of movement strategy. Humans could move like that but they don’t want to. They instead use this bell-shaped velocity profile, which means we move with a slow acceleration and in fact, let’s see, there’s something called position, you’re changing position, there’s something called velocity, which is how fast you’re going, there’s a derivative of velocity called acceleration, which is how fast you’re changing your acceleration, then there’s a derivative of acceleration. Guess what that’s called? It’s called “jerk,” okay, because when you change your acceleration rapidly you induce a lot of vibrations in the system and it makes it look like you have high jerk. What humans seem to do is whatever they’re doing in their head, they seem to be moving as if they’re minimizing the square of jerk.When you apply that to robots and get them to move in the same way, they look alive, they look like a human and they do not look jerky."

"Okay, there’s a basic split, there are two kinds of robotics people. There are people who want robots to be useful and there are people who want robots to be like humans, and those are pretty much disjoint sets of people and they fight with each other and I’m clearly on the side of, “I want to understand what it means to be a person,” robots are a way for me to do that understanding and my idea of a good robot is it’s like a person, behaves like a person, looks like a person. Isaac Asimov, and of course many, many other science fiction writers and movie generators, so, “Metropolis,” the robot looks like a person. It’s a lot easier, you have an actor be your robot so they naturally look like people, but a lot of these people believe that that’s what a robot should look like and so they’re on my side. In the 1970s I can remember as a graduate student, or I guess 1980s, a professor named Warren Seering in the Mechanical Engineering Department wrote an article for I guess the MIT alumni magazine, Technology Review, which basically said if you wanna build VCRs your robot should not look like a person, it should look like a VCR making device, and he was absolutely right but the politics of this message were really bad. It basically said, “Don’t fund all these crazy people who want to build people or artificial people, fund people who are going to do something useful to society,” because at that time the reason robotics was hot was because we were trying to deal with the Japanese and the perception that automation was making them much more competitive relative to us and our trade balance was therefore terrible. That fear has receded now and we have terrible trade balances with places like China, which has little or no automation, so there are other issues going on as well."

"Well, in those days there were no CCD cameras, so it was a regular camera like yours and then attached was a digitizer, which provided you digital images, and then you worked offline on these digital images. And the challenge there was that first of all the memory in computers were very small. We worked on small images, so the resolution was not very good, worked on 64 by 64 pixels or at best 128 by 128 pixels, so the resolution was rough, and in texture you really need a higher resolution in order to capture the local properties, so that was a challenge. Technology just wasn’t there. Everything had to be made in the lab. We had to build digitizers and deal with swapping the memory to disk all the time, because the memory was typically 64K or so or even – then later on I worked on PDP-11, which had a 32K memory, the cache, so it was very different time, and therefore the approaches you took were quite different than today people use millions of pictures and do statistical analysis, very different."