
Ray Kurzweil (USA) at Ci2019 - The Future of Intelligence, Artificial and Natural
What this covers
This is a one-hour conference keynote from Creative Innovation 2019 in the Asia Pacific, delivered solo by Ray Kurzweil, who also leads a machine-intelligence and language-understanding team at Google. The talk moves from a claim about how people misjudge change, through the history of AI and the brain, to a set of near-term forecasts. It begins with perception: a 2013 poll of 24,000 people in 23 countries found 70% believed extreme poverty had worsened over two decades, when it had fallen by half. Kurzweil ties this to the habit of ignoring exponential growth, and he explains how a chain of S-curves, each paradigm flattening as the next starts, produces a sustained exponential. He cites a 300,000-fold rise in neural-network compute since 2012.
The middle of the talk is about intelligence itself. He recounts the perceptron of the 1960s and the limits that Marvin Minsky's 1969 book proved, then contrasts it with today's neural networks. He describes the neocortex as a thin folded sheet that processes patterns hierarchically over dozens of levels, and uses a surgical case in which a patient laughed only when specific axons were touched to argue that abstract functions like humor are localized. Practical demonstrations include Gmail's smart reply and Talk to Books, which searches 120,000 books by meaning.
The last stretch turns to society and is where the claims get more contested. He projects renewable energy reaching full global supply in the early 2030s, falling costs in food, energy, and housing, and a shorter work week driven by automation. He expects a social safety net already funded at current levels to support comfortable living by the 2030s. Gene sequencing falling from $1 billion to about $1,000 and cancer immunotherapy illustrate his medical optimism. Kurzweil also warns that facial recognition and social credit systems could enable totalitarian control. He closes on education, arguing for learning by doing on real problems with teachers as mentors, and for the capacity of young people to tackle large problems.
Ray Kurzweil argues that exponential technological growth, driven by computational advancement across multiple paradigms, will transform human civilization through artificial intelligence, life extension, brain-computer integration, and radical improvement in quality of life by the 2030s-2040s, contrary to widespread pessimism about progress.
- Exponential growth in computing power has followed a consistent trajectory for 45+ years and will continue, enabling AI to master all human skills
- Historical data shows human welfare metrics (life expectancy, literacy, poverty, income, education) improving dramatically across all dimensions, contradicting public perception of decline
- Deflation in technology-enabled sectors (food, energy, housing via AI, renewables, 3D printing) combined with existing social safety nets will enable comfortable living standards without increased earnings by 2030s
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Most people ignore the exponential nature of technological change, which causes them to make inaccurate long-term predictions because exponential growth is not immediately visible in early phases and eventually reaches an asymptote (S-curve), after which a new S-curve begins, creating the appearance of sustained exponential growth across multiple paradigms.
“people ignore the exponential nature of technological change so their long-term predictions leave out the essence of change people point out the fact that exponential change is not go on forever so if you know exponential it eventually comes to an asymptote however so that's an S curve however we've seen that 1s curve leads to another s curve and this has happened already four times with computation and we're already on the sixth paradigm which is three dimensional growth of transistors”
Despite ubiquitous technological progress, human perception of global conditions is inverted from reality: a 2013 poll of 24,000 people across 23 countries showed 70% believed extreme poverty had worsened over the prior 20 years when it had actually decreased by 50%, and only 1% of respondents predicted the actual outcome.
“this is a poll of 24,000 people in 23 countries they were asked whether extreme poverty has gotten better or worse over the past 20 years 70 percent thought it had gotten worse only 12 percent thought it had gotten better the reality is that poverty had actually decreased by 50 percent which was the prediction of only 1% of the people who responded”
Between 1800 and 2015, life expectancy at birth improved from the 30s to approximately 70+ years globally, and even in the poorest countries today, life expectancy exceeds that of the richest countries in 1800, representing universal human welfare improvement.
“these circles are countries and this is the state of the world in 1800 so life expectancy on the y-axis was in the 30s so life expectancies was in the 30s of years even in the wealthy countries”
The perceptron, a single-layer artificial neural network introduced in the 1960s, could recognize some patterns but had fundamental limitations, such as inability to solve the connectedness problem (determining whether shapes are spatially connected), which Minsky's 1969 book mathematically proved.
“this was a simple neural net it's organized as a single level it emerged in the 1960s in 1962 when I was 14 I visited dr. Rosenblatt at Cornell and he was pushing something called the perceptron... he said to me and this is actually quite significant if we were to create a perceptron with multiple layers not just one it would overcome the vagaries that a perceptron could not deal with”
Advanced surveillance technologies (facial recognition, social credit rating systems, AI tracking in China) create serious risks of totalitarian control and authoritarian enforcement, requiring ongoing political and social vigilance to prevent abuse.
“I've been accused of being an optimist I am somewhat optimistic but I've written about the downsides of these technologies and you could use these technologies to enforce a totalitarian system or an authoritarian system and we do see that... you have to fight that”
Neuroscientists have identified specific neural locations in the brain that detect humor, demonstrated through a case where a patient undergoing brain surgery laughed only when surgeons touched specific axons showing humor detection, proving that high-level abstract cognitive functions are localized and replicable.
“this is actually a young girl she was sixteen she was actually having brain surgery and the surgeons wanted her to be conscious and react to her experience so whenever they had touched the axons here shown in red she would laugh they first thought she's having some laughs reflex but they quickly discovered that no she was actually perceiving humor only when they touched these points they had found the points that actually detect humor”
Elementary and junior high school projects can involve real-world problem-solving that teaches reading, mathematics, and other skills as byproducts of solving meaningful problems, with teachers acting as mentors rather than content deliverers.
“it's changing my grandson is actually in a school that practices learn by doing and they do different projects where they will learn how to read and to different things in the real world... for elementary school where you actually go in the world interact with it you learn other skills not as the major thing but as a side product of actually learning how to change the world and then the teachers actually becomes kind of a guide mentor to doing that”
The number and types of intellectual pursuits available to humans will continue to expand with technological advancement, with entirely new fields of study emerging that don't exist today, as happened between 1900 and 2019.
“there's going to be many more types of pursuits that we can pursue I mean if you look at the types of things that people pursue now they weren't they didn't exist at all in 1900 or even 1950 and we're gonna keep pursuing increasing forms of new types of development”
Exponential growth in computational power has followed a consistent straight-line trajectory on an exponential chart from the early 20th century through present day, and this pattern will continue, supporting Kurzweil's long-term technological predictions.
“I noted the speed of these processors that is calculations per second for constant dollar for all the computer processors up to that year you see going back to the early 20th century the data forms a straight line on an exponential chart and the straight line and an exponential chart is exponential growth”
US social safety net spending has increased steadily both in absolute terms and as a percentage of GDP over the past 70+ years, and will continue to rise, with this baseline spending being sufficient for comfortable living by the 2030s even without further increases.
“here's the expenditures for the u.s. total social safety net spending so it started out almost at zero it's gone up steadily it just goes up and the percentage of GDP also Rises with it”
Deflation in DNA sequencing costs has enabled modern biotechnology, reducing the cost of decoding a human genome from $1 billion to approximately $1,000, making genetic medicine and personalized treatment economically viable.
“the deflation in DNA sequencing costs has enabled biotechnology and that's what really brought biotechnology to the fore decoding the first genome cost a billion dollars now it costs about a thousand dollars”
Learning by doing, where students work on meaningful real-world problems as the primary educational method, creates more engaged learners and better prepared students than traditional lecture-based education, and should be the foundation of education systems.
“I think we need to learn by doing and that's actually something that Danielle follows in fact the school she's in she gets a little assignment like when she's 11 to go solve the palestinian-israeli the israeli-palestinian problem which she does actually get it's an a-minus in the course so we may not want to assign that necessarily but we do want to assign real meaningful work where people can actually change the world”
Ordinary people, including children, can solve major problems that society faces, as evidenced by major companies founded by college students, high school students creating medical tests, and emerging technologies coming from younger populations.
“most of the problems that can be solved can be solved by ordinary people including children you know the major companies we've seen have been founded by college students we're now seeing high school students do fantastic things including new medical tests that could be done by junior high school students”
The average work week has decreased over time, particularly since the Great Depression, as a direct consequence of increased automation in production, enabling workers to earn good incomes while working fewer hours.
“a workweek has gone down it shot up slightly after the Great Depression but it's continued to go down since then this is one consequence of increased automation”
65% of American workers now work in information-based jobs that would be unrecognizable to workers in 1900, such as web designers, developers, and IT engineers, representing emergence of entirely new job categories from technological change.
“65 percent of American workers deal with information in jobs they would not be recognizable in 1900 for example webdesigner half developer IT engineer and so on”
Thirty doublings in exponential growth produces a billion-fold improvement (30 doublings = 2^30 ≈ 1 billion), and this is the magnitude of improvement seen in electronics and will be seen in every other product category as technology advances.
“if you leave out the exponential thirty changes gets you to 30 and that's why people in the long term leave out the usability of change thirty doublings in an exponential gets you to a billion that's what we've seen with electronics and we'll see it soon with every other product”
The neocortex processes information hierarchically across multiple levels simultaneously, with each ascending level adding conceptual abstraction, such that progression from recognizing shapes to understanding whether text is ironic or funny occurs across dozens of hierarchical levels.
“there are levels of the neocortex there are many dozens of them and as you go up the all the levels are actually processing simultaneously as you go up through the different levels each level adds some concept it's to what we're perceiving so here you can see the different levels and an understanding of what words are doing on a page if you go up another twenty levels you'd have axons that decide whether this passage is ironic or funny or deceitful”
The neocortex, which arrived with mammalian animals 200 million years ago and became dominant after the Cretaceous extinction event 65 million years ago, is a two-dimensional structure approximately as thick as a napkin with intricate folds that increase its surface area and enable hierarchical pattern processing.
“the neocortex is the outermost layer of the brain it's fin essentially a two-dimensional structure it's about as thick as a napkin with intricate folds with deep ridges grooves wrinkles that increase its surface area to be the bulk of the human brain it's responsible for our ability to deal with patterns of information in a hierarchical fashion”
Extreme poverty rates have declined steadily worldwide since 1820, and the decline has accelerated in recent decades across all geographic regions.
“declining poverty rates worldwide since 1820 decline in extreme poverty across different areas”
Immunotherapy for cancer—where patients' own blood cells are treated outside the body and reinjected to fight their specific cancer type—is less invasive than traditional treatments and offers superior patient safety prospects, with numerous trials showing complete remissions even in Stage 4 cancers.
“immunotherapy for cancer is where our patients blood cells are treated outside of the body there re injected into the patient and then light up that person's own immune system to fight their specific type of cancer this is personalized medicine it's far less invasive than traditional treatments and peratt's provides far more promise for patient safety this there have been a number of tests where every single person including Stage four cancers have obtained complete remissions”
AI systems capable of natural language understanding can provide intelligent email response suggestions, as demonstrated by Google Gmail's smart reply feature, which is powered by neural networks developed by Kurzweil's team.
“ai is also making progress in natural language if you have Gmail you'll notice that it provides an intelligent response giving three choices for your next email that choice is actually provided by my team at Google”
The 'Talk to Books' Google feature allows users to query 120,000 books' content (half a billion sentences) simultaneously, with the system reading all books in approximately one second and returning semantically relevant answers based on meaning rather than keyword matching.
“we've also put the same capability in something called talk to books so if you go to Google and just put in talk to books you'll get the link and then you go to that link and you can ask any question you want and the software will read all of a hundred twenty thousand books it actually reads all of 120 thousand books in a second it's a half a billion sentences and in a second gives you the best answer from the hundred twenty thousand books based on the meaning of your question”
Renewable energy capacity is growing exponentially, doubling every 4 years, and should reach 100% of global energy supply by the early 2030s, though even at that point only 1 part in 10,000 of available solar radiation will be utilized.
“we have an exponential growth of renewable energy it's done now this is renewable energy that will replace fossil fuels it's it's increasing it's actually increasing at an exponential rate its doubling every four years so it should reach a hundred percent by the early 2030s the growth rate for solar energy is even higher oh and by the way when we reach a hundred percent we'll only be using one part in 10,000 of all the sunlight that hits the earth”
Computational capacity for neural networks has increased 300,000-fold since 2012, with computational requirements doubling every 3.5 months, demonstrating exponential growth in AI training capacity that has only recently reached levels sufficient for human-level and beyond-human-level capabilities.
“this graph it looks like my other graph it goes up over time but it's it's not the doubling time is not one year the doubling time here is 3.5 months... we've basically been doubling the amount of computation in the neural net every 3.5 months so it's the the speed has increased 300,000 fold since 2012”
AI systems are improving on standardized tests faster than human performance improves; approximately 6 years ago AI exceeded third-grade reading comprehension benchmarks, and today AI exceeds average adult performance on adult reading comprehension tests.
“there are tests where you can read a paragraph and then the answer questions about your understanding of the paragraph you're probably taking this if you've been in grade school about six years ago AR computers were challenging a third grade test today they are exceeding the average adult score on the adult test”
The internet's power to connect humanity and enable access to global knowledge and communities based on shared values creates more benefit than harm despite current problems, and these problems (misinformation, polarization) existed before the internet in different forms.
“it's so powerful it affects so many aspects of our lives that it's bound to get people criticizing it we certainly see a lot of things that are wrong but there's certainly were things wrong before we had any internet people had many problems without an internet... we have access to whole human knowledge we can create communities based on common perception of of value across the world so it is complicated but I think in we gain a lot more benefit than then then problems that we're generating”
Telecommunications capacity (speed of data transmission per constant cost) has improved by a trillion-fold over the past century, from processing a few characters per minute via Morse code to modern high-speed networks, following the same exponential trajectory as computational power.
“the speed of telecommunications we processed a few characters per minute using Morse code over AM radio a century ago today it has multiplied a trillion fold for the same cost”
Income per capita in constant 2015 dollars was in the hundreds of dollars per year even for wealthy countries in 1800, whereas modern developed nations have per capita incomes exceeding $30,000-$50,000 annually in constant dollars.
“income per person in today's dollars was in the hundreds of dollars per year shown in today cells so even in the wealthy countries it was hundreds of dollars that you try to live on in a year”
Protein folding is a critical biological process where amino acid chains fold into unique three-dimensional shapes; predicting these shapes is extremely demanding, but AlphaZero achieved 43 correct predictions out of 99 compared to the best human achievement of 3 out of 99, demonstrating AI's potential for creating new proteins for medications.
“in biology we have all these cells the cells put out a chain of amino acids and they're in a straight line in a magic moment of biology this linear sequence of amino acids folds up into a unique three-dimensional shape so this is a key thing in biology and to try to predict what that three-dimensional shape would be is very very demanding a few people have learned it this very same neural net that learned go and learn chess and learned other things taught itself protein folding it was way ahead of the best humans at predicting protein folding the best humans got three shapes predicted out of 99 alpha zero got 43 out of 99 so this actually will be extremely valuable in creating new proteins for medications”
Of Kurzweil's 147 predictions made for 2009 in the late 1990s, 86% proved correct, as documented in a 150-page public record of prediction accuracy.
“if you go to how my predictions are faring if you put in how my predictions are faring on the web it will bring up 150 page document that goes through the predictions I've made in at different times over the last 20 years so for example I made a hundred and forty seven predictions for 2009 in the late 1990s and 86% of them were correct and they were recorded in the age of spiritual machines”
Manufacturing labor has declined from 25% of the US workforce in 1920 to 9% by 2015, yet overall labor statistics have improved, demonstrating that sectoral employment decline does not cause permanent unemployment.
“we've had a similar trend with manufacturing labor going from 25 percent in 1920 to 9 percent today”
AlphaZero, released by DeepMind in 2016, taught itself to play Go using only the rules, without human-provided moves, and achieved performance far beyond the best human players, uniquely demonstrating that the same neural network architecture could master multiple domains (Go, chess, protein folding) with no human guidance on strategy.
“consider alpha zero this came a little over a year ago from deep minds Google subsidiary it taught itself go with only the go rules it was not given any moves by humans it's performance is way past the best you you take the very best humans”
Average income per hour earned by American workers has continuously increased in constant dollars since 1929, contradicting widespread reporting that average hourly income has stagnated.
“this is the average income per hour actually earned by each person in the United States in constant dollars since 1929 it just goes up this is in today's dollars so we hear a lot of reporting that average income per hour has stabilized it's not going up but that's obviously not the case”
Brain imaging capabilities (both destructive and non-destructive) are constantly improving, enabling observation of individual nerve cells forming in real-time; these imaging advances allow formation of realistic models of how the brain works.
“our ability to image the brain both destructive and non-destructive brain imaging is constantly getting better we can now see individual nerve cells forming in real-time with non-destructive imaging as a result were able to form a realistic idea of how our brains work”
Suicide rates and mental health problems are serious concerns, but improved brain imaging and neuroscience understanding will enable better diagnosis and treatment of mental illness, as current psychiatric models are vague and lack precision about underlying causes.
“I'm not sure that suicide is actually increasing but it is a serious problem I think as we go forward we'll actually have better ideas of how mental illness occurs right now we have very vague psychiatric models of Howard Kurtz something goes wrong and we're not really sure why and there's something to do with serotonin and but we don't really have a precise method of actually understanding what happens with mental illness we are getting the tools now to see inside the brain and with continued research using these new tools we'll have a much better idea of what actually goes wrong in mental illness and we'll be able to fix it in many cases”
Future food production will move away from animal agriculture through development of plant-based and lab-grown meat alternatives that taste like traditional meat, reducing human dependence on animal agriculture and enabling more humane relationships with animals.
“it will be less dependent on animals we're ready of various types of meat that don't require meat but that tastes like meat and we're moving I think away from reliance on animals for food”
Kurzweil envisions direct brain-computer interfaces where the top layers of the neocortex will be extended through connections to web-accessible AI, adding new concepts and capabilities while preserving natural intelligence, creating a hybrid intelligence greater than biological or artificial intelligence alone.
“my scenario is that at the top level of our neocortex will extend we will connect with extenders that access the web... the fifth layers are not like this they're actually like a pyramid in the very top level increases detects the things that are most sophisticated... that's the level that we want to extend to add new concepts so it's the top layer that needs to be extended we would then have a combination of our natural intelligence with the super normal intelligence that we are creating”
All human skills and intelligences can be represented as patterns that neural networks can learn to recognize and replicate, meaning neural networks will be able to master everything that humans can do as computational capacity expands and training examples become available.
“everything that humans can do is explained by these types of examples some will require some navigation to find because they're not always written down we might have to actually monitor the world to find what those things are but over the next few years we'll have these Nets explore everything intelligent that humans can do”
Due to deflation in necessities (food, housing, energy) and existing social safety net spending levels, most people will be able to meet their life needs by the 2030s even without increased income, representing a fundamental shift in the relationship between work and subsistence.
“as prices for food and other necessities of life come down this will be enough for most people's life's needs by the 2030s... it actually will be very very enough for most people's life's needs as we get to the 2030s and certainly by the end of the 2030s”
Universities will continue to exist and evolve because their core function—embodying expert practitioners who guide students in specific domains—remains valuable, even as technologies for learning expand and students gain more direct access to tools.
“I think the universities is going to continue because really what a university does is embodies somebody it was really an expert on certain modality and gets into that and pursues it and we're going to continue to do that with increasing power of tools to expand our our minds we'll also have the ability ourselves to learn new things but we'll have certain people that really guide us to pursue each food”
Previously it was thought that axon-to-dendrite interconnections were a chaotic mess, but higher-resolution scanning reveals that all axons are already pre-connected with other axons; when two axons want to remain connected, the interconnections become permanent through a synaptic strengthening process.
“at first it was thought that the interconnections if I want to create a connection from this axon to this dendrite was a huge mess but we were not scanning it with enough resolution it was too low now they were scanning it with high resolution what we see is that all axons are already connected with other axons and when two axons want to remain connected the interconnections are made permanent”
Political difficulty in discussing future jobs comes from the fact that new job categories are invented in fields that don't yet exist, making them impossible to discuss or advocate for in present-day policy debates until they have already begun to emerge.
“it's very hard and it's actually a political disadvantage because we can't really speak about these jobs because I haven't even been invented yet in there in a field that doesn't exist yet”
Timing is critical for technological success and innovation; inventors like Thomas Edison and Larry Page succeeded not just because they had good ideas but because they had the right ideas at the right time in history when technological prerequisites were met.
“I realized a timing was critical to being an adventurer Thomas Page Larry Thomas Edison Larry Page other inventors you've heard about we're in the right place with the right idea at the right time timing was critical for everything from making investments to romance”
If in 1870, when there were only 65,000 college students in America, someone had predicted that 20 million people would be studying poetry and quantum mechanics by the 21st century, it would have seemed absurd, yet this is exactly what happened, demonstrating that new job categories are difficult to predict.
“if I were to say at the end of the 19th century that 20 million people would be studying poetry in quantum mechanics by the beginning of the 21st century that would seem absurd to a country that had only 65 thousand college students in 1870”
Historically, humans have used technology to amplify their own capabilities rather than be replaced by it—machines amplify muscle capacity, and future AI will amplify cognitive capacity—making the 'humans vs. AI' narrative from dystopian movies unlikely because technology integration, not competition, is the historical pattern.
“people's imagination imply that it will be humans versus a eyes we see that in dystopian movies of the future but that is not how we've used technologies thus far we've used it to amplify our own capabilities for example try building the building that you're in be very hard to do with your bare muscles but we have machines that go way past our muscles so with those machines were able to build a great structure now we will use the increasing strength of non-biological thinking to extend our own thinking”
Meaning and purpose in human life come from intrinsic motivation (wanting others to read one's book, enjoy one's music, appreciate one's ideas) rather than solely from economic compensation, and reduced pressure to earn income will actually increase human freedom and allow pursuit of these deeper motivations.
“I mean we get a tremendous amount of I mean if you publish a book you want people to read it you want people to appreciate the ideas that's not just a financial perspective that's something that comes from the heart that you want to connect with people people have different ways in which they appreciate knowledge music for example is something that is meaningful because you play it so it's not just earning a living”
AI can now be applied to diverse domains beyond single specialized tasks, with a single neural network architecture capable of mastering multiple different problems through variations in training, creating a general-purpose intelligent system.
“it's really one thing and it can be applied to everything and we've already shown it to be applied to great many different types of things and it requires a very advanced form of computation but that's what we have now”
Without radical life expansion (new experiences, enhancements, and virtual reality), people would get bored living extremely long lifespans; however, with continuing enhancement of our minds and virtual reality capabilities, humans will have a much more enhanced and extended life.
“people say well aren't we going to get bored if we live that long we're doing the same things over and over again and I actually think that's true we would get bored if we lived a very long time if we had radical life expansion without radical life extension so we're going to be doing things like virtual reality is the previous speaker just spoke about and a lot of other things that will enhance our minds right now we have about ten of the fourteenth calculations per second in our brain we're actually going to expand that so we're gonna live a much more enhanced life as well as a extended life”
There are approximately seven things that limit human lifespans to around 120 years, all of which are addressable; through advances in medicine, AI, and biotechnology, humans will overcome cancer, heart disease, and other limitations, eventually backing up human minds into non-biological intelligence systems.
“well we're addressing the things that cut short our lives there's about seven things I could point out that limit our lives to like 120 not that we actually get that far but all of those are actually addressable we're actually addressing things like cancer and heart disease before we get to that and we're going to ultimately pass our ideas on to non-biological intelligence which will work with us and so we'll actually be able to back ourselves up as you go further and further in time we'll be able to live really longer and longer”
Kurzweil is releasing a novel called 'Danielle' in two weeks; it is more than fiction, serving as a vision narrative, history, and roadmap for readers of all ages; it includes the ideas from his talk but also companion non-fiction books: 'How to Create a Mind' (a guide with 280 essays) and 'How You Can Be a Danielle' (providing pragmatic ideas for action to bring about a better world).
“one is called how you can be a Danielle and in this book they're pretty heavy it's a chronicle of ideas which is 280 essays on the ideas in the book”
Kurzweil chose a young woman (Danielle) as his protagonist not because the message is primarily for girls, but because nine years ago (before this talk) he was searching for a compelling young person character, and she emerged as the conception; the message is genuinely for both young women and young men and all adults.
“why do you think educating young women in particular to be entrepreneurs is important well it really is for young women and young men like we have movies that show both young women readers and young men readers and I don't know why it's a young woman I mean nine years ago I was trying to find some idea of a young person they would be able to bring these do these developments and she just turned out to be a young woman named Danielle and but it really it's for young women young men and and adults as well”
Child labor has declined substantially worldwide since 2000, representing continued moral progress despite ongoing concerns about exploitation.
“the client of child labor in the world since 2000”
Democracy has expanded from covering a very small percentage of global population a century ago to approximately 50% of the world today, representing significant but incomplete progress toward democratic governance.
“spread of democracy it was very small maybe a century ago really began with the good bird printing press beginning to distribute information to be continue through the 20th century with radio TV and computers it's now up to 50%”
Marvin Minsky acknowledged near the end of his life (a few years before his 2016 death, approximately 50 years after his 1969 perceptron book) that neural networks were indeed a very promising research direction, reversing his earlier position.
“shortly before he died a few years ago and this is 50 years after these interactions he acknowledged Minsky acknowledged that the neural nets were indeed a very promising road ahead”
Frank Rosenblatt in 1962 predicted that multi-layer perceptrons could solve problems that single-layer perceptrons could not, without ever having built a multi-layer perceptron, showing remarkable prescience about the potential of layered neural architectures.
“Rosenblatt's concepts was indeed very prescient because he had never done a multi-level perceptron before and yet he figured out that would actually solve the problem and that not only would it solve the problem but it would ultimately solve problems that humans are unable to do”
Homicide rates in Western Europe have declined exponentially since 1300, showing long-term decline in violence as societies have developed economically and institutionally.
“homicide rates in Western Europe since 1300 he's in one live in 1300”
US education spending per capita has increased steadily since 1949 in constant dollars, enabling greater educational investment per student and expanded educational access.
“us education expenses per capita in constant dollars since 1949”
The novel Danielle includes 24 color illustrations by New Yorker cartoonist Amy Kurzweil (Ray's daughter), with one illustration for each year of her life.
“the novel includes 24 color illustrations by New Yorker cartoonist and my daughter Amy Kurzweil that's one for each year of her life and here Danielle shows her sister her book at age one she's not yet speaking but she will speak”
Kurzweil's novel 'The Danielle Future' features a protagonist who uses technology to solve major global problems, serving as a thought experiment about how individuals can contribute to positive technological change.
“I've thought about how can I present these ideas in a way that will resonate with everyone young and old so I thought about a compelling story of a young person who uses her intelligence and accelerating technologies to take on the most difficult problems confronting the human race so Danielle it's the name of my character she's a remarkable character who does what every talented ambitious person wants to do she takes action and changes the world for the better”
The novel 'The Danielle Future' will be released in two weeks (from the time of this talk) and includes the novel itself, two companion non-fiction books (available online), and 24 color illustrations by Kurzweil's daughter Amy, who was a New Yorker cartoonist.
“in my view this novel which I'm actually releasing in two weeks is a thought experiment... it also has a couple of other things it's more than a work of fiction it's a vision peril history a road map for readers of all ages”