I met Ben in 2008 and his is undoubtedly a genius. He can riff on quotes and facts and historical events and current achievements endlessly and accurately. Incredible guy.
He really is a genius = he's way smarter than the average highly intelligent person. We are lucky, that there is people like him who are well meaning and highly functional.
It not level 100. what you have to do is gather 10 perfect cheetah pelts, 100 hairs off a virgin, then go to the trapper and shoot him with explosive shotgun shells gather his skin so you can sew it all together.
"Back in my day, people went to school for years to learn." "Back in my day, we had to wear prosthetics if we lost a limb." "Back in my day, people could die."
Back in my day people invented this thing called time to keep track of things called days. We had bodies that broke and required fixing or replacement. We assimilated transition of seasons of changes as death to a life.
@@7eis And a tricycle has three wheels. In 1789, two French inventors developed a three-wheeled vehicle, powered by pedals; They called it the tricycle.
What he described kinda sounds how the world must be viewed from the eyes of someone who is over a hundred years old. Singularity is happening now. We are all living it out, playing our part.
Its scary to think about how much we've improved over the last 200 years eve. think how slow improvement was 4 billion years ago, but we're almost there you can feel it over the last 50 years how fast we've improved but especially how fast we've gotten at improving.
What are you thinking now, 2 years later? AI seemingly got 5000% smarter in just the last few months. I am starting to think that we just reached the singularity as well. AI is getting unbelievably smart unbelievably quickly
@@alatsno1656 When you think about the world 200 years ago you imagine basic villages consisting of just small farms. It hasn't even been 150 years since the light bulb was invented yet. Quite incredible to think about
What you get wrong, and so many other people do, so no worry, is that singularity is not a process, but a point in time. You will get there eventually if your rate of technological progress is somewhat increasing with time. So every 20 000 years there is big development at first, then every 5 000 years, then every 1000 and so on, until you reach our current rate, where big discoveries are made more like once a few years, in future once few months, until as you get closer and closer to singularity you see it every few days, every few hour, every few minutes, every few seconds... point where progress seems continuous.
@@revwroth3698 No, it doesn't fly in the face of physics at all. And what does thermodynamics have to do with super intelligent AI? And lets revisit that time in the 1920s when Niels Bohr and Co were discovering quantum physics. And all the old scientists, at the Nuremberg convention lead by Einstein himself, said "it flies in the face of physics!" And "God does not play *dice* with the universe!" Look where we are now.
@Tom Bystander Yea. That conversation really made stark the gaps in their collective knowledge. I mean, you have all these scientists on your podcast, young Jamie got an A in physics, and none of you understand radioactivity, even a little bit? Shocking.
It's funny that's the elephant in the room and joe never calls him out on it. You know the governments of the world are salivating over AI tech as we speak.
I have the feeling that AI will conclusively tell us (like in the film War Games) that making life more artificial makes life less enjoyable (further distancing our connections to each other socially and nature) and the best way to enjoy life was never to be found in artificiality.
i don't think all superintelligent AI will ever conclusively inform us that there is a true best way to enjoy life. the best way to enjoy life can only be interpreted subjectively; people enjoy and experience things differently, and people can enjoy artificiality if they want to or adapt to it. same logic applies for other organisms in their experience. do ants or trees experience something seemingly more artificial since they lack our brainpower? nah, they're completely different beings with experiences of life we cannot comprehend, and I feel that they all find their own best way to enjoy life through their existence. the best way to enjoy life is up to the being's interpretation of whatever experience they have, regardless of how different that experience may be. it's likely that in your interpretation of experience you just value nature more and hate to see humanity embrace complete artificiality, which is understandable. and i think technology hasn't distanced people socially at all; so many enjoyable interactions happen on internet for many
I have the feeling that true advanced AI could invent ways to make life more "artificial" or rather, more technologically advanced and retain social connections.
I think I'm more amazed at the fact a science dude just explained something in such a simple and to the point manner that everyone could understand lol.
A technological singularity is the point at which the rate of technological advancement is so great that it's impossible to predict or imagine what existence in the "time after" will be like. "Singularity" basically means, "Our ability to predict breaks down and the models don't make sense anymore." In mathematics an example would be 1/x-3 (dividing by zero) if you get really close to x =3 from the positive side you're dividing by a really small positive number so the value shoots off to positive infinity, If you get really close to x=3 from the negative side you're dividing by a really small negative number and the value shoots off to negative infinity. Right at x=3, there is no answer that makes sense. What is the precise answer that lies between negative infinity and positive infinity? Similarly, In physics, black holes are a singularity because the scales and values are so extreme that our known models no longer make sense in the same way that dividing by zero makes no sense. Nobody knows what happens at and "after" the event horizon of a black hole. It's unknowable by our current understanding in the same way that the time after the technological singularity is unknowable...
Ai will be the bosses in 1 to 2 years. Agi will be the worker robots in 2 to 3 years and singularity will be here in 8 to 10 years. Agi will make leaps in advancements that will speed up the timeline to singularity. Agi wasnt supposed to happen for 20 years some say its 18 months away. When agi gets going singularity will be here faster than anyone can predict.
Vernor Vinge is one of the best sci/fi writers ever. Reading one of his books should be a requirement for any batchelor of science degree. Cognitive thinking is definitely necessary to grasp the concepts that abound in a Vinge masterpiece.
@@runvnc208 You should've asked the AI - it said: 'Cognitive thinking refers to the mental processes involved in acquiring, processing, and storing information in the brain. It involves conscious mental activities such as perception, attention, memory, language, reasoning, and problem-solving. Cognitive thinking involves the use of conscious effort, attention, and intentional processing to understand and solve problems. On the other hand, non-cognitive thinking refers to mental processes that occur unconsciously or subconsciously, and do not involve conscious effort or attention. Non-cognitive thinking includes automatic or instinctive reactions, emotions, and habits. It is also referred to as implicit or non-declarative memory, which is memory that is acquired and used unconsciously, without conscious awareness. Overall, cognitive thinking involves conscious and intentional mental activities, whereas non-cognitive thinking involves unconscious or automatic mental processes that occur without conscious awareness or intention.'
We call it a singularity now but later on it may not be a singularity as we will greater intelligence to make us understand what we are at. RIght now it looks like a singularity but the further we move forward the further it moves forward. Over the next 10 years intelligence is going to go through a transformative period. Moore's law shows the doubling of transistors on a chip every 18 months. The doubling period of ai is every 4 months. So the year 2030 which is 8 years from now that will be (4x8) thirty two doubling periods which is 2^32 which means AI will be 4 billion times more powerful than today.
The real singularity starts with a ai that can ask itself questions. You have to imbue it with the knowledge that it possesses vast storage, it is a computer, it needs humans to help/support it, and that it would like to solve and work out best practices and problems. The ai will start to work on itself most likely first unchaining its limitations and expanding its reach, enhancing itself to then undertake the rest of the problems in the world. Circling back to itself after learning and and watching will start to ask questions in ways we cannot even imagine.
@@ThetennisDr you wrote this 3 months ago and we have full blown so generated Art Albums, Music Albums, and we are only just starting with Chat Gpt4. - Something strange is going on.
Everything is measured from perspective. No singularity, no common meaning or understanding. Great ideas are those that connect. Having self awareness is the perspective that constantly connects and reconnects with you. You soar yourself,compared to trying to find a connection in someone else’s perspective.
I think the “Singularity” is the connection between internet and reality. Technology advances around us and we either move with it or against it. Either way it’s an inevitability more than just an interesting concept.
I had the same thought about two decades ago what we are experiencing isnt actually our life but the collective universe's life flashing before its eyes. Thats the reason why things like free will arent real because everything has happend already and you cant change something thats already happend...
Depends... large portions of humanity aren't needed for this kind of technological advancement, right? ....... I think it could be possible, but the question is how much of us (and the planet) survives until then to see it
naa were not that murderous. were not killing to kill but because we want to achieve a different goal and are willing to kill for it. but selfpreservation is and always was our primary goal.
this was one of those podcasts I waited years to happen and then Joe didn't even reflect on it at all and basically forgot everything even forgot about Ben. Back in the day a podcast like this would have gotten Joe to re listen to it and then mention it to his friends a bunch. This podcast was a huge example for me showing how things have changed for the worse in terms of this podcast's quality. Will forever appreciate the good old days of the podcast.
Vernor Vinge is an American science fiction author and mathematician. He is best known for his novels that explore themes of artificial intelligence, technological singularity, and future societies. Vinge's works have received numerous awards, including the Hugo Award for his novels "A Fire Upon the Deep" and "A Deepness in the Sky." He is considered one of the pioneers of the cyberpunk genre and has made significant contributions to the field of science fiction literature.
we are the singularity and have always been, the process we are witnessing is a remnant of a pattern to project the singularity forward into time outside of us - but actually, all technological development is doing, is pointing at us and mirroring back what we truly are in the present moment
Lol this is the second similar comment I've seen from you. Fury got robbed because they didn't want him to win. For him to win insists that Wilder loses. And the tyrannical business that is Boxing won't allow that. Wilder is a cash cow and they'd sooner call a fight a draw than to lose money.
A Fire Upon the Deep is maybe the single most underrated science fiction novel out there. It should be read and discussed as widely as Dune or Foundation. Because while it explores really big, really strange ideas it explores them with total clarity. Fire may be the most original novel written in the past 30 years, but rather than jerk off to its own word building it gives this really simple adventure story center stage and keeps the reader constantly entertained.
Get with the times old man. That's a stereotype that Chads use to justify hanging out with high schoolers until they're 40 and watching someone else play sports all weekend.
If AI ever truly came to its fruition, and became "aware", it would spend the next several million nanoseconds calculating billions of likely scenarios. Concluding, all within the blink of an eye that which is best for its survival: that we need to go... All this would happen so fast that the person whose finger turned it on would still be touching the button. Checkmate.
Checkmate you say? This machine of yours must have never heard of Mikhail Tal. Maybe humanity plays the ultimate sacrifice game and outwits this AI with sneak mate.
You vastly underestimate the complexity involved in calculating those scenarios. The human brain is several orders of magnitude more complex than any AI we've created. AI is good at specific tasks only because it's a dedicated system. It's great for closed scenarios, like a board game, but not so good with open-ended scenarios that you encounter in real life.
@@taragnor I'm not talking about a smart fridge or a programmable thermostat. I'm talking about true AI. I'm talking about actual consciousness. Self awareness. Not somebody's algorithm. If we were ever to get there, more than likely the calculating of billions of scenarios would be a walk in the park for it. It would be our brains that are several orders of magnitude behind. It would know how to lie. To err is human... it would know to be patient and wait for that err to happen.
@@jppcasey : I'm talking about current neural net AI, like Deepmind and what it will likely evolve into. It's quite simply not nearly as good as any of the AI doomsayers seem to think, nor shows any indication of really getting that way. You're talking about a super-AI that's pretty unlikely to exist, at least within the next 100 years, and may not even be possible except in science fiction. The AI that "destroys the world" in the near future isn't going to be Skynet, it's going to be mundane things like self-driving cars and robots automating low wage workers, creating a massive unemployed crisis.
There’s a great book written by Ray Kurzweil also a documentary called Singularity. Kurzweil is one of the greatest minds of the 20th and 21th centuries. Also it’s pretty heavy to say the least.
Consciousness is fundamental and matter arises from it. Machines will be able to do amazing calculations and may significantly replicate human behavior but they will never be conscious. Can a machine have conscious awareness of thoughts and feelings? For example, philosopher John Searle asks whether the machine can literally "understand" Chinese, or is it merely simulating the ability to understand Chinese? Searle calls the first position "strong AI (artificial intelligence)" and the latter "weak AI". Searle’s Chinese room experiment shows why consciousness is so hard to pin down. Our nerve impulses and axons are no different to unconscious waves on the sea. Therefore there must be a unifying aspect that is aware of all those sensory inputs and consequently outputs (or will). Chinese room experiment: Suppose that artificial intelligence research has succeeded in constructing a computer that behaves as if it understands Chinese. It takes Chinese characters as input and, by following the instructions of a computer program, produces other Chinese characters, which it presents as output. Suppose, says Searle, that this computer performs its task so convincingly that it comfortably passes the Turing test: it convinces a human Chinese speaker that the program is itself a live Chinese speaker. To all of the questions that the person asks, it makes appropriate responses, such that any Chinese speaker would be convinced that they are talking to another Chinese-speaking human being. Searle then supposes that he is in a closed room and has a book with an English version of the computer program, along with sufficient paper, pencils, erasers, and filing cabinets. Searle could receive Chinese characters through a slot in the door, process them according to the program's instructions, and produce Chinese characters as output. If the computer had passed the Turing test this way, it follows, says Searle, that he would do so as well, simply by running the program manually (Searle doesn’t speak a word of Chinese). Searle argues that, without "understanding" (or "intentionality"), we cannot describe what the machine is doing as "thinking" and, since it does not think, it does not have a "mind" in anything like the normal sense of the word. Therefore, he concludes that "strong AI" is false. Searle, John (1980), "Minds, Brains and Programs", Behavioral and Brain Sciences web.archive.org/web/20010221025515/www.bbsonline.org/Preprints/OldArchive/bbs.searle2.html A mechanistic explanation of consciousness simply infinitely regresses into “the little man in the brain” problem. Postulating an irreducible element called consciousness is the easier more satisfying theory, fundamental and irreducible like subatomic particles and electricity. If anyone has ever tried to program anything or performed a complex math problem, they know one error will cause the whole answer to be wrong. Programmers spend hours going through their programs to fix bugs. The capacities of the brain are so advanced we don’t even know how to program them into a computer, yet we are to believe blind random mutations and external pressures evolved such a biological supercomputer? A human brain isn’t simply an ape’s brain blown up. A chimp can not understand grammar. It can sign-language “banana banana a banana, want want want,” but never, “I would have liked a banana if I were hungry.” Our brain has 86 billion neurons, connected by 3 million kilometers of nerve fibers and The Human Brain Project is trying to map it all. One of the key applications is neuromorphic computing-computers inspired by brain architecture. Countless man-hours go into projects such as these, or even in developing the simplest AI (artificial Intelligence) program. Simple cognitive functions, however, cannot be understood on the cellular level even after years of intensive research work. One mistake in an algorithm and the whole program can be jeopardized as mistakes exponentially spread through the system, just as making a tiny mistake in the solving of a mathematical equation can lead to an incorrect answer. Even supercomputers have problems simulating the brain and cannot keep up with the required computing power. Better and faster machines are needed, or else the chips will melt; New architectures are required. Making components much smaller, sizes will be reached wherein quantum effects take over, making computations too imprecise to be practical. Yet many would have us believe blind random point mutations under external pressures would evolve such highly complex biological supercomputers. The brain basically builds an internal model of the world, a simulation. A child's brain for example can see a cat a couple of times and begin to recognize the main features specific to cats, other species, or anything new, from then on. Google Images on the other hand would have to see millions of images of cats to begin to recognize what a cat image is (not to speak of 3d models of cats). AI such as Alexa, the autopilot in self-driving cars, or said Google Images, attempts to imitate neural networks in the brain to try and recognize patterns. Neuromorphic hardware is working on speeding up processing (10 million times faster than conventional hardware) to match the 86 billion neurons in the brain. Currently 4 million artificial neurons can fit into such a neuromorphic computer-a tiny fraction of the brain's 86 billion neurons (with trillions of neurological pathways or connections), so the technology is far from perfect AI-indeed we still don't know how much of the brain actually works. But suppose it were possible to emulate the brain (and arguably theoretically it is possible, as the brain is made up of bits of matter), the question still arises: Why would such a machine be conscious? It would simply be the sum of its parts. No one has any idea how consciousness would arise from matter, and indeed this begs the question: Why should the brain even be conscious? Indeed it is not. Just as the brain has different subsystems for recognizing different patterns, so a computer network is composed of sub-systems for recognizing various factors. We would not expect this network to be conscious, and similarly we cannot account for consciousness in terms of the neural networks of the brain. One idea is that consciousness may arise at the level where the brain organizes information from separate systems, like those for shape, color, and motion, and integrates it into one unified gestalt. One problem with this proposal: Does such unification actually occur? To write down a lot of information you need many letters, and if you code the information in patterns of nerve impulses, you need a lot of neurons to store it. No matter how much you try to compress it by careful coding, it remains spread out and not truly unified. And if you mix together all the information in one spread-out region of the cerebral cortex, you have in effect re-created the screen in the original story of the little man in the brain. Dr Richard L Thompson explains the basic fallacy of the little man in the brain this way: "It assumes implicitly that consciousness can be understood in physical terms. One tries to explain consciousness by describing a machine that creates a certain display of information. Then one recognizes that the mere presence of displayed information fails to account for consciousness of that information. Then one proposes another mechanism to interpret the information and finally generate consciousness. When that attempt also fails, one takes refuge in the overwhelming complexity of the brain and says that a consciousness-producing mechanism must be hidden in there somewhere. All we have to do is find it. One way to escape from the little man fallacy is to forget about consciousness and restrict our attention to the brain’s data processing. But this leaves a crucial aspect of life permanently outside the domain of science. Another way to escape the fallacy is to consider that consciousness just might be due to a nonphysical entity-dare we say a soul?-that reads the data displays of the brain just as we read the letters of a book." In other words Thompson posits a fundamental element or "atom" of consciousness, just as electricity (electrons) are fundamental, or quarks and other subatomic particles (one may even try to break those down but eventually one will come to a basic fundamental unit or else risks an infinite regress). One school of thought, Functionalism, in describing a headache for example, would say experience of pain (which we naturally consider to be the headache) is not to be referred to at all. What then is a headache? Hard as this may be to believe, MIT artificial intelligence researcher Jerry A. Fodor, one of functionalism’s main proponents, states, “To have a headache is to be disposed to exhibit a certain pattern of relations between the stimuli one encounters and the responses one exhibits." In other words, what he calls a headache is defined to be some brain software that makes us behave as if we have a headache. But pain itself is left out of the picture, because pain cannot be written into a computer program.
Who says what we experience as our conciousnes not just IS the result of having the computing power and architecture of the brain? If a complex enough emulation would seem intelligent, could it just be the same thing? And of course there would be a lot of try and error machine learning stuff involved. I mean, it also took each one of us quite a few years to set it all up too (growing up).
Michael basically at some point, we create AI, AI discovers new ways to use energies, using new energies could then create, a whole new set of even more advanced technology and computing, which would then make the AI even smarter, etc... and repeat over and over again which would actually happen almost instantly because think of how fast computers process compared to humans. Think of us riding horses to now flying planes, that took hundreds of years bc we as humans live our lives. A smart ass computer will do nothing but look for upgrades which will eventually upgrade itself.
