Short & Sweet AI

Who will win the race to control Artificial Intelligence? China or the U.S?

I’m Dr. Peper and today I’m discussing the book AI Superpowers by Kai Fu Lee.

In AI Superpowers, Kai Fu Lee, an accomplished AI scientist, former head of Google China and now one of China’s top venture capitalists, makes some alarming predictions. In his book he says that within the next decade the U.S. which is ahead in AI with new discoveries, will be surpassed by China due to China’s better implementation. The two nations will then dominate as the AI superpowers and he describes the other nations being “left to pick up scraps while the AI superpowers will boost productivity at home and harvest profits from around the globe.”

According to Lee, the U.S. is in the age of AI discovery which requires brilliant scientists and innovative breakthroughs. Next will be the age of AI implementation which requires vast quantities of data to train the computers and super fast computing power. Here China has the lead: China has 800 million internet users (three times more than the U.S.) These users generate huge amounts of data which their government freely collects. And China is investing heavily in quantum computing which exponentially increases super fast computing power,

Kai Fu Lee also describes the additional factors leading to China’s ascendancy in AI as being the Chinese “hypercompetitive” business landscape and state support for developing and subsidizing AI industries.

And there is another comparison Lee makes that I and many other readers have found very disquieting. He describes the major difference between Chinese and Silicon Valley tech culture. His observation is Silicon Valley is mission driven to take an original idea and achieve an idealistic goal and change the world. Chinese businesses are market driven to make money and become rich.

AI Superpowers is written from the perspective of China rather than Silicon Valley and questions many of our assumptions about the US and AI technology. Sometimes it’s uncomfortable and that’s why you should read it.

https://www.washingtonpost.com/outlook/in-the-race-for-supremacy-in-artificial-intelligence-its-us-innovation-vs-chinese-ambition/2018/11/02/013e0030-b08c-11e8-aed9-001309990777_story.html (https://www.washingtonpost.com/outlook/in-the-race-for-supremacy-in-artificial-intelligence-its-us-innovation-vs-chinese-ambition/2018/11/02/013e0030-b08c-11e8-aed9-001309990777_story.html)

https://www.goodreads.com/book/show/38242135-ai-superpowers (https://www.goodreads.com/book/show/38242135-ai-superpowers)

https://www.nytimes.com/2018/09/22/opinion/sunday/ai-china-united-states.html (https://www.nytimes.com/2018/09/22/opinion/sunday/ai-china-united-states.html)

https://drpepermd.com/wp-content/uploads/2019/12/Facial-Recognition-Technology.docx (Facial Recognition Technology)

I’ve been hearing more and more about facial recognition technology and here’s what I’ve learned.

Facial recognition is a biometric technology using artificial intelligence to identify people. Biometrics is the measurement and analysis of a person’s unique physical and behavioral characteristics. Using a type of AI called machine learning, computers analyze large data sets in the form of photos and videos looking for patterns and learn to recognize facial features. The computer then compares new images to stored images in its databases to identify a face.

The technology is most accurate at identifying white, male faces since the algorithms have been trained on mainly white male visual data sets and is least accurate at identifying people of color and women.

Pro and cons of the technology center around its use to improve security and safety versus the threat to an individual’s privacy.

Important points and facts to be aware of:

San Francisco became the first U.S. city to ban facial recognition by government agencies.

Microsoft and Google reported they have denied facial recognition services to law enforcement while Amazon has faced pushback from its employees and shareholders for selling the technology.

There are new facial recognition smart glasses available only to police which allows users wearing augmented reality glasses to scan faces in a crowd which are then compared to a million image database. Positive matches are sent to a display embedded in the lens of the glasses.

And the New York Times reported facial recognition is growing stronger thanks to your face. Using images from social networks, photo websites and cameras placed in public areas, there is a growing database of collected photos in the public domain available for anyone to download and use for training facial recognition software.

On a more hopeful note, a start up company is working on a tool to let you check whether your image is part of an openly shared database of faces. Somehow I think they may be using AI to develop it.

https://www.nytimes.com/2019/07/13/technology/databases-faces-facial-recognition-technology.html (https://www.nytimes.com/2019/07/13/technology/databases-faces-facial-recognition-technology.html)

https://www.theverge.com/2019/6/10/18659660/facial-recognition-smart-glasses-sunglasses-surveillance-vuzix-nntc-uae (https://www.theverge.com/2019/6/10/18659660/facial-recognition-smart-glasses-sunglasses-surveillance-vuzix-nntc-uae)

https://www.forbes.com/sites/bernardmarr/2019/08/19/facial-recognition-technology-here-are-the-important-pros-and-cons/#1918805214d1 (https://www.forbes.com/sites/bernardmarr/2019/08/19/facial-recognition-technology-here-are-the-important-pros-and-cons/#1918805214d1)

https://www.stopspying.org/our-vision (https://www.stopspying.org/our-vision)

https://drpepermd.com/wp-content/uploads/2019/12/Meta-Trends.docx (Meta Trends)

The world can be exciting but scary as the way we live changes and has changed in just the last 10 years. Peter Diamandis, co founder of Singularity University, tells us why it’s changing even faster than before.

One meta trend he sees is increasing abundance. Products and services are cheaper and more available to everyone. The proportion of the world’s population living in extreme poverty is the lowest it’s ever been. And while goods are increasing, the costs of day-to-day things such as food, energy, communication, and transportation are trending down.

Another meta trend is people everywhere are becoming connected as 4 billion more people are brought online by 2025 all connected at gigabit speeds, all wanting to create, discover and invent. And with the meta trend of capital abundance there is more capital to invest in the newly generated ideas. Venture capital funding globally increased to total of 207 billion dollars in 2018, up from 186 billion dollars the year before.

Increasing human intelligence is another meta trend. Having data on everything will lead to something called just in time education where 5G speed combined with AI and augmented reality will allow you to learn something the moment you need it. Then there’s the crazy idea but actually predicted as reality that brain computer interfaces will allow us to connect our brains to the cloud where we will have any and all information available, at anytime.

A final converging meta trend is increasing human longevity. Even as our lifespan has doubled in the last 100 years, the expectation is to double it again. Using technologies such as CRISPR, stem cell therapy, genomic sequencing, 3D printed organs and AI digitized detection of diseases before they develop, it’s likely in the future people will live to 140.

With the negative news cycle wanting our eyes to be focused on their advertisers, the watchword is if it bleeds, it leads. Instead what is less obvious and what we need to understand with data is, the world is quietly getting better.

From Short and Sweet AI, I’m Dr. Peper.

https://singularityhub.com/2019/08/20/these-are-the-meta-trends-shaping-the-future-at-breakneck-speed/ (https://singularityhub.com/2019/08/20/these-are-the-meta-trends-shaping-the-future-at-breakneck-speed/)

https://futurism.com/scientists-genetically-engineer-mice-live-25-percent-longer/ (https://futurism.com/scientists-genetically-engineer-mice-live-25-percent-longer/)

https://futurism.com/new-1-terabit-internet-satellites-will-deliver-high-speed-internet-remote-areas/ (https://futurism.com/new-1-terabit-internet-satellites-will-deliver-high-speed-internet-remote-areas/)

https://www.brookings.edu/blog/future-development/2018/12/13/rethinking-global-poverty-reduction-in-2019/ (https://www.brookings.edu/blog/future-development/2018/12/13/rethinking-global-poverty-reduction-in-2019/)

https://drpepermd.com/wp-content/uploads/2019/12/Exponential-Technology.docx (Exponential Technology)

At Singularity University’s 2019 Global Summit, Peter Diamandis, one of the co founders, captivated the audience with his discussion of exponential technologies that have already been set in motion creating something called meta trends. A meta trend has been described as a trend that runs deeper and powers more specific trends, like a tidal wave that drives waves to the shore.

And an exponential technology is when the technology’s power or speed doubles each year and/or the cost drops by half. Exponential technologies are those that are rapidly accelerating and shaping major industries and all aspects of our lives, and that idea is actually the subtitle of Diamandis’ latest book, to be released in January 2020. It’s called The Future is Faster than You Think.

He describes many of the meta trends as unstoppable and even somewhat predictable. There are 2 things driving these trends to converge here and now. One is quantum computing which allows all the trends to get strong because of cheaper, faster computing power starting with the move to 5G networks. And the second is there is an acceleration of the acceleration. The rate at which technology is getting faster is itself getting faster. More people are connected than ever before with access to more and more information and the technology they’re using is cheaper than ever before.

So what are the meta trends?

* Accelerating technology which is cheaper and available to everyone

* Increasing global abundance

* Everyone, everywhere is connected at gigabit speeds

* Everything, everywhere is connected

* You can know anything, anywhere, anytime

* Autonomous personalized transport which is fast and cheap

* Increasing human intelligence w/ AI support and brain computer interfaces

* Increasing human longevity

* Capital abundance with access to capital everywhere

* Globally abundant, cheap, renewable energy

I’ll discuss some of these details in the next episode but I want to leave you with this: people have no idea how fast the world is changing.

From Short and Sweet AI, I’m Dr. Peper.

https://singularityhub.com/2014/06/10/staggering-promise-of-exponential-technologies-in-a-succinct-5-minute-video/ (https://singularityhub.com/2014/06/10/staggering-promise-of-exponential-technologies-in-a-succinct-5-minute-video/)

https://su.org/blog/exponential-technology-trends-defined-2019/ (https://su.org/blog/exponential-technology-trends-defined-2019/)

https://singularityhub.com/2019/05/06/5g-is-here-what-does-that-mean-for-exponential-tech/ (https://singularityhub.com/2019/05/06/5g-is-here-what-does-that-mean-for-exponential-tech/)

https://xponentialworks.com/what-is-exponential-technology/ (https://xponentialworks.com/what-is-exponential-technology/)

https://su.org/blog/exponential-technology-trends-defined-2019/ (https://su.org/blog/exponential-technology-trends-defined-2019/)

Elon Musk, best known for Tesla, Space X, and Open AI, also has a company called Neuralink which works on developing brain computer interfaces.

An interface is a connection between a person and a computer like a keypad or touchscreen or when we use a mouse to control a computer screen. It’s been reported that Neuralink is a very secretive company but Elon Musk recently gave a big presentation to show off some of Neuralink’s revolutionary technology.

By way of background there’s already been a person with spinal cord paralysis who received a brain implant and was able to control a computer cursor to play a game using his thoughts. Another patient also with a brain implant was able to move a robotic arm in a limited fashion.

The Neuralink device represents a substantial advance over this older technology because it allows people to communicate more quickly with computers directly from the brain. Multiple flexible threads, each one much thinner than a human hair, are surgically inserted into the brain with very little damage but with the ability to transfer enormous amounts of data. Right now the implant connects to a computer via a USB cable, but a sensor is being developed that would sit behind a person’s ear to transmit information from the brain threads to the computer wirelessly. The idea is for paralyzed patients to use this interface to control phones or computers with their thoughts. These threads are so flexible many more can be inserted in the brain but their flexibility also make them more difficult to implant.

So Neuralink’s second big advance was to invent a neurosurgical robot that can insert the threads automatically, up to 6 threads with 192 electrodes per minute, which allows for an ultra high bandwidth brain computer interface. As Musk explained everything you perceive, feel, hear, think, it’s all impulses from neurons and these flexible threads can record from and selectively stimulate many, many neurons across diverse brain areas.

In the short term the goal is to treat serious chronic brain diseases and brain damage caused by strokes and trauma. But Musk has an even greater long term goal: to have humans merge with artificial intelligence which he feels is necessary to keep us from becoming irrelevant as AI rapidly advances. His response to why he founded Neuralink is “the existential risk is too high not to.”

https://www.theverge.com/2019/7/16/20697123/elon-musk-neuralink-brain-reading-thread-robot (https://www.theverge.com/2019/7/16/20697123/elon-musk-neuralink-brain-reading-thread-robo)t

https://interestingengineering.com/key-takeaways-from-elon-musks-neuralink-presentation-solving-brain-diseases-and-mitigating-ai-threat (https://interestingengineering.com/key-takeaways-from-elon-musks-neuralink-presentation-solving-brain-diseases-and-mitigating-ai-threat)

https://interestingengineering.com/neuralink-how-the-human-brain-will-download-directly-from-a-computer (https://interestingengineering.com/neuralink-how-the-human-brain-will-download-directly-from-a-computer)

#6 https://drpepermd.com/wp-content/uploads/2019/11/XR-What-is-Extended-Reality-1.docx (XR- What is Extended Reality?)

Have you heard of XR, extended reality? It’s something I came across just recently and want to share with you because it’s expected to be part of everyday life in the next 10 years.

Extended reality is an umbrella term for augmented reality, AR, virtual reality, VR, and mixed reality MR. These are technologies that merge our physical world with the virtual world, a world simulated by a computer.

Augmented reality augments your physical world by adding a new layer where virtual information and computer simulated objects are overlaid onto the real world by using AR glasses or a smart phone . Pokemon is AR where digital creatures are superimposed onto the real world as well as Snapchat which uses filters to superimpose a hat or glasses onto your head.

In contrast to augmented reality, in virtual reality you are fully immersed in a computer simulated environment using a VR headset or a head mounted display. The headset generates realistic sounds and images and engages all five of your senses to create a world you can interact with as if you were at a live concert, scuba diving in the ocean, or walking on the moon.

Mixed realty or MR is a combination of virtual and augmented reality. The real and virtual worlds are blended to create an environment that mixes physical and digital elements. Like AR, mixed reality can superimpose digital content onto a real world and like VR, in mixed reality, you can move objects and interact with everything around you.

There are challenges to this technology such as the vulnerability of having large amounts of very detailed, personal data collected about what you do, what you look at, even your emotions. And there are difficult technical and hardware issues to get the display, power, motion tracking, and connectivity to give a realistic, immersive experience. Also there’s a high cost to implementing this technology.

Despite these problems, XR will include any future realities, too. It’s a fundamental shift in the way you’ll carry out your daily life. Even as you attend a morning conference in China, go over business plans in New York, and host clients in Brazil for drinks after work, all from your office in LA, you won’t refer to the technology at all because your experience will be seamless. Someone has called it the end of distance.

From Short and Sweet AI, I’m Dr. Peper.

https://www.forbes.com/sites/bernardmarr/2019/08/12/what-is-extended-reality-technology-a-simple-explanation-for-anyone/#d03b03872498 (https://www.forbes.com/sites/bernardmarr/2019/08/12/what-is-extended-reality-technology-a-simple-explanation-for-anyone/#d03b03872498)

https://www.theverge.com/2019/1/28/18197520/ai-artificial-intelligence-machine-learning-computational-science (https://www.theverge.com/2019/1/28/18197520/ai-artificial-intelligence-machine-learning-computational-science)

https://www.visualcapitalist.com/extended-reality-xr/ (https://www.visualcapitalist.com/extended-reality-xr/)

https://www.youtube.com/watch?v=MHz2Ib0JeJY (https://www.youtube.com/watch?v=MHz2Ib0JeJY)

https://www.youtube.com/watch?v=lbJ-IKPn2l8 (https://www.youtube.com/watch?v=lbJ-IKPn2l8)

Creativity…. in this age of artificial intelligence it is thought to be the thing that distinguishes us from machines.

#5 https://drpepermd.com/wp-content/uploads/2019/11/Creativity-and-AI.docx (Creativity and AI )

Yes AI can be used to perform routine tasks faster with less mistakes, drive cars, read medical images and even talk to us, but machines do not have the inspiration or originality which is the uniquely human ability to be creative. That has all changed with a type of AI called machine learning which is forging a new exciting era for the arts.

In 2018 the auction house Christie’s sold the first piece of art created by a computer. The painting, which ultimately sold for $430,000, was created by a group of artists and AI researchers using a computer algorithm call Generative Adversarial Networks or GAN. They trained the GAN algorithms on 15,000 portraits dating from the 14th century to the current day. The way a GAN works is when two algorithms compete against each other, that’s the adversarial part, one is the generator which creates content and one part is the discriminator which judges the content. The generator makes a new image and then the discriminator tries to spot the difference between a human made image it trained on and one created by the generator. With the two competing with each other, something new is created, a new sort of art. One that is signed with a mathematical equation at the bottom right instead of an artist’s signature.

There’s also been an AI generated movie. The original goal was to see if a computer could win the 48 hour sci-fi London film Festival. The film created from the screenplay made it into the top 10. The following year the same director and creative technologist working with an algorithm Bot that named itself Ben, won third place. Last year the Bot was placed in charge of all aspects of the movie making from writing the screenplay to selecting the score and stringing sentences together using voice recordings from the actors. The result was a film called Zone Out, Silicon Valley and was generally agreed to be a crazy mess. But it demonstrated the potential to generate a complete film using artificial intelligence.

Whether it’s a painting or a movie, AI can generate new art and in the process is redefining what it means to be human.

From Short and Sweet AI, this is Dr. Peper

https://www.christies.com/features/A-collaboration-between-two-artists-one-human-one-a-machine-9332-1.aspx (https://www.christies.com/features/A-collaboration-between-two-artists-one-human-one-a-machine-9332-1.aspx)

http://www.obvious-art.com (www.obvious-art.com)

https://www.thecreativepenn.com/2019/03/11/ai-and-creativity-with-marcus-du-sautoy/ (https://www.thecreativepenn.com/2019/03/11/ai-and-creativity-with-marcus-du-sautoy/)

https://www.youtube.com/watch?reload=9andv=LY7x2Ihqjmc (https://www.youtube.com/watch?reload=9andv=LY7x2Ihqjmc)

https://www.wired.com/story/ai-filmmaker-zone-out/ (https://www.wired.com/story/ai-filmmaker-zone-out/)

Algorithms, neural networks, data … it involves machines and it’s deep.

https://drpepermd.com/wp-content/uploads/2019/10/Are-Machine-Learning-and-Deep-Learning-the-same-thing-as-AI-1.docx (#4 Are Machine learning and Deep Learning the same as AI? transcript)

Just when I thought I was beginning to understand AI, a friend asked me a question I struggled to answer. He asked me if machine learning and deep learning are the same thing as AI. Here’s my explanation that I couldn’t give to him then but that I can share with you now.

Think of AI, machine learning, and deep learning as circles within each other. AI is the biggest circle and then there’s a little smaller circle that is machine learning and within that is another circle that is deep learning.

So AI uses something called algorithms which are essentially a set of instructions of computer code that tell a computer what steps to follow to solve a problem or reach a goal. In machine learning computers take the data you’ve give them and learn from it and make decisions based on that learning. They do this by using something called an artificial neural network that is inspired by how the human brain works. The algorithms behave as though they are interconnected brain cells creating these artificial neural networks which process information the way the brain does. We know from cognitive scientists that the brain is arranged in different layers to process different pieces of information. The information comes into the brain and each level or layer of neurons processes the information, provides insight and passes it on to the next more senior level. That’s how the human brain learns and that mechanism is how these artificial neural networks work. They take the information, learn from it and pass on to the next, more senior level which processes it further and adds to the learning.

So the machines learn from the data you give them and that’s called machine learning. The artificial neural networks are connected one layer to the other and when you have many layers of these neural networks several layers deep that is what is called deep learning. For example Google uses a 30 layered artificial neural network to power Google photos and Facebook has what it calls its DeepFace algorithm which uses deep artificial neural networks to recognize faces with 97% accuracy.

I love it when people ask me questions that I can’t explain because when I figure it out I can share the explanations with you. From short and sweet AI, I’m Dr Peper.

https://www.forbes.com/sites/bernardmarr/2018/09/24/what-are-artificial-neural-networks-a-simple-explanation-for-absolutely-anyone/#53f6f8c21245 (https://www.forbes.com/sites/bernardmarr/2018/09/24/what-are-artificial-neural-networks-a-simple-explanation-for-absolutely-anyone/#53f6f8c21245)

https://blogs.nvidia.com/blog/2016/07/29/whats-difference-artificial-intelligence-machine-learning-deep-learning-ai/ (https://blogs.nvidia.com/blog/2016/07/29/whats-difference-artificial-intelligence-machine-learning-deep-learning-ai/)

https://hackernoon.com/deep-learning-vs-machine-learning-a-simple-explanation-47405b3eef08 (https://hackernoon.com/deep-learning-vs-machine-learning-a-simple-explanation-47405b3eef08)

https://towardsdatascience.com/clearing-the-confusion-ai-vs-machine-learning-vs-deep-learning-differences-fce69b21d5eb (https://towardsdatascience.com/clearing-the-confusion-ai-vs-machine-learning-vs-deep-learning-differences-fce69b21d5eb)

Mainstream AI and all existing AI is called narrow artificial intelligence. Take 3 minutes and get the skinny on it.

https://drpepermd.com/wp-content/uploads/2019/10/The-Skinny-on-AI.docx (#3 The Skinny on AI transcript)

Artificial intelligence is understood to mean machines that mimic human behavior and intelligence. But actually there are 3 different types of AI: narrow AI, general AI, and super AI.

What we recognize as AI in the everyday world such as Google ranking pages, Amazon knowing what we like, and voice assistants such as Alexa, is considered narrow artificial intelligence or weak AI. It’s considered narrow because the ability to mimic human behavior works within a very limited narrow context and it can’t take on tasks beyond it’s field.

Other examples of narrow AI are Google translation where a book written in Mandarin Chinese can be translated to English in 30 seconds or self-driving cars which are a combination of several different narrow AIs. So you can see narrow intelligence is not low intelligence and because it can do many jobs faster with less mistakes, narrow AI is threatening to displace many human jobs. But even as narrow AI displaces humans from the routine jobs, it’s still not considered human level artificial intelligence.

Narrow artificial intelligence is where we are, general artificial intelligence is where we are going.

General artificial intelligence or strong AI refers to machines that can perform any generalized task which is asked of it, much like a human. This is the sort of AI seen in movies like “Her” and other science fiction movies where machine and operating systems are conscious and driven by emotion and self- awareness.

General AI would be expected to be able to reason, solve problems, understand uncertainty, integrate prior knowledge in decision-making, all at faster processing speeds than humans as well as be innovative, imaginative, and creative and experience consciousness. Most experts believe general AI is possible but since one of the world’s fastest supercomputers took 40 minutes to simulate a single second of brain neural activity and since we have 100 billion neurons in the human brain, I wouldn’t hold my breath.

Artificial super intelligence or super AI is when AI surpasses human intelligence in all aspects from creativity, to problem solving, to general wisdom. Super AI is something futurists speculate about and which we will discuss in a future Short and Sweet AI segment. Until next time, I’m Dr. Peper.

https://www.computerworld.com/article/2906336/what-is-artificial-intelligence.html (https://www.computerworld.com/article/2906336/what-is-artificial-intelligence.html)

https://interestingengineering.com/the-three-types-of-artificial-intelligence-understanding-ai (https://interestingengineering.com/the-three-types-of-artificial-intelligence-understanding-ai)

The hottest topic in AI this year is something called voice computing. In fact it’s been described as the latest disruption in technology.

https://drpepermd.com/wp-content/uploads/2019/10/Voice-Computing.docx (#2 Voice Computing transcript)

Voice computing at its simplest is when you use your voice to control your devices like what you’re doing right now getting this flash briefing, rather than touching your keyboard or screen. Now we use tools to hold, touch or swipe for information or to perform tasks. With voice computing these tools disappear and are replaced by voice assistants on smart speakers who respond to voice commands. The entire notion of what a computer is and does is different because of voice computing humanized with names such as Alexa, Siri, Cortana or Bixby.

Computer speech technology crossed a threshold in the last five years. We know human beings can understand and comprehend an average of 95% of what another human being says. Compare that to artificially intelligent computers which can now understand and comprehend 99% of what a human says to them.

Let’s look at the numbers. There are 2 billion desktop and laptop computers in the world and 5 billion mobile phones. By the end of this year there will be more than 200 million smart speakers worldwide, a number expected to grow to 8 billion by 2023, but there are devices such as home gadgets and appliances, 100 billion of them, expected to connect to these smart speakers and that market is exponentially larger than even mobile.

The days of going to a browser to type in a query or starting at a homepage on a screen are disappearing. Instead we will have voice discovery of information. Content will be filtered through a voice assistant which is able to understand the semantics and underlying intent and context of what we want because they have data of everything we’ve done, the purchases we’ve made, the podcasts and music we’ve listened to, the vacations in our calendar, the email sequences from a friend. All of this is discoverable by voice assistants and will be brought to bear in understanding the nuances of our requests.

Screens and smart phones won’t be eliminated just as the jet airplane didn’t kill off the car but people will gravitate to the more natural interface of voice. Computers will follow us around rather than us needing to go to them. This is called a voice first future.

From short and sweet AI, I’m Dr. Peper.

https://www.theverge.com/2019/5/20/18537019/artificial-intelligence-alexa-siri-cortana-google-voice-computing-james-vlahos-talk-to-me (https://www.theverge.com/2019/5/20/18537019/artificial-)

https://www.theverge.com/2019/5/20/18537019/artificial-intelligence-alexa-siri-cortana-google-voice-computing-james-vlahos-talk-to-me (James Vlahos: “Talk to Me” | Talks at Google – YouTube)

https://www.theverge.com/2019/5/20/18537019/artificial-intelligence-alexa-siri-cortana-google-voice-computing-james-vlahos-talk-to-me (https://www.youtube.com › watch)

https://www.theverge.com/2019/5/20/18537019/artificial-intelligence-alexa-siri-cortana-google-voice-computing-james-vlahos-talk-to-me (intelligence-alexa-siri-cortana-google-voice-computing-james-vlahos-talk-to-me)

https://techcrunch.com/tag/voice-computing/ (https://techcrunch.com/tag/voice-computing/)