Rise of AI Conference 2016 · Berlin

Panel: The Singularity Might Be Closer Than You Think

Samim Winiger, Ehssan Dariani, Ronnie Vuine, Danko Nikolic, Jan Klauck

At Rise of AI 2016 in Berlin, a panel moderated by Samim Winiger with Ehssan Dariani, Ronnie Vuine, Danko Nikolic and Jan Klauck discusses whether the singularity is near. The panel first debates whether understanding intelligence is a precondition for building AGI, with one panelist arguing it is not, since humans do not fully understand their own brains either, and another distinguishing human-level from human-like intelligence. Asked to describe the world in 50 years, panelists offer visions ranging from smart robots doing unwanted work to humans merging with technology and uploading themselves. The discussion turns to whether the singularity movement already shapes policy, with panelists citing the NSA, homeopathy regulation and drug prohibition as examples of the gap between expert knowledge and political decisions. Panelists then name the technologies exciting them most, including high-resolution brain electrodes, cheap smartphones reaching poor populations, lab automation, and disappearing user interfaces. The panel closes on whether Moore's law's slowdown threatens the exponential case for the singularity.

Key takeaways

  • One panelist argues AGI does not require understanding the human brain, since humans do not fully understand their own brains either.
  • Danko Nikolic says new electrode arrays record hundreds of thousands of neurons and will soon image brain electric fields at 20-30 micrometer precision.
  • A panelist argues cheap smartphones reaching poor populations will change societies more in the next 10-15 years than the last 50 changed Western Europe.
  • The panel cites homeopathy regulation and drug prohibition as examples where political decisions ignore scientific consensus because of what voters believe.
  • One panelist argues raw computing power reaching brain-equivalent levels does not equal intelligence, since a machine still needs knowledge to use that power.

Detailed notes

Must you understand intelligence to build AGI?

Opening the panel, one panelist argues it is certainly easier to build something once you understand it, and points to repeated "AI winters" as evidence that a lack of understanding has historically caused failed promises. Another panelist disagrees, arguing that humans do not need to understand the brain, creativity or consciousness in detail to build systems that emulate intelligence, since humans do not fully understand their own brains either and evolutionary mechanisms produced complexity without planning. A further view distinguishes human-level intelligence (generalization, abstract thought, transfer across domains) from human-like intelligence, arguing AGI needs the former but not the latter, and that for practitioners, understanding intelligence is itself often the motivation for doing the work.

Visions of the world in 50 years

Asked to project 50 years ahead, one panelist hopes for smart robots that do unwanted work without complaint, freeing people to be creative, while acknowledging this is a wish rather than a prediction. Another panelist argues that the steps toward singularity are incremental, describing a future where humans have robots in their brains and may upload themselves, while questioning whether the average person understands enough science, comparing this to how few people understand general relativity or quantum mechanics despite media coverage. A third panelist says he is emotionally indifferent to the singularity, seeing it as a continuation of gradual technological change rather than a single event, and hopes that if his darker predictions are wrong, the outcome would be AIs discussing music or art.

Does the singularity movement influence policy?

One panelist says the singularity movement's influence on policy has only just started but will grow, comparing it to how nuclear energy gradually became a routine concern. Another argues that politics is already influenced by technology, citing the NSA and Edward Snowden, and contrasts this with cases where scientific consensus is ignored by politicians for fear of losing voters, giving homeopathy regulation and drug prohibition in Latin America and Afghanistan as examples. A further contribution recalls that an Intel CTO referenced the singularity explicitly around 2008, naming 2050 as a date for humans merging with machines, and notes that groups like Singularity University try to involve policymakers directly.

Which current technologies excite the panel most

Danko Nikolic, introduced as a brain researcher, says he is most excited by new electrode arrays that can record hundreds of thousands of neurons and soon produce real-time images of electric fields in the brain at a precision of 20 to 30 micrometers. Another panelist argues technology does not need to be sophisticated to be transformative, pointing to cheap smartphones with internet access reaching poor populations worldwide as a bigger driver of change over the next 10 to 15 years than the previous 50 years brought to Western Europe. A further contribution names lab automation for medical research and new materials, along with better brain imaging, brain-machine interfaces and 3D printing, as most exciting; another panelist highlights the coming disappearance of clumsy user interfaces in favor of natural human-language interaction with technology.

Q&A highlights: Moore's law and exponential growth

Asked about reports that Moore's law is slowing, one panelist says every exponential curve must eventually flatten into a sigmoid because neither the universe nor time is infinite, and argues that comparing raw computational power to brain power, as Ray Kurzweil does, conflates computing power with intelligence: powerful computers still need to be filled with knowledge, and that compressed, accumulated knowledge is itself a slower, older exponential process going back to the origin of life. Another panelist counters that more people today live up to their creative potential than in Shakespeare's or Mozart's time, when most talented people worked in fields and died young. A closing view notes that Moore's law covers only one segment of technological progress, and points to developments in quantum computing as a field that could make Moore's law "look like a fault" by comparison.

Questions answered in this talk

Do you need to understand human intelligence to build artificial general intelligence?

At a Rise of AI 2016 panel, one panelist argues understanding is not required, since humans do not fully understand their own brains and evolution produced complex intelligence without planning; another panelist distinguishes human-level intelligence, which AGI needs, from human-like intelligence, which it does not.

Does the singularity movement have real influence on government policy?

According to panelists at Rise of AI 2016, the singularity movement's policy influence has only just begun; one cites an Intel CTO referencing the singularity around 2008, and notes that Singularity University tries to involve policymakers directly.

Is Moore's law slowing down, and does that undermine predictions of a technological singularity?

At Rise of AI 2016, one panelist argues Moore's law is one segment of progress that must eventually flatten like any exponential, but says this matters less than assumed because raw computing power is not the same as intelligence, and another panelist points to quantum computing as a field that could outpace it.

Mentioned

singularity, AGI, cybernetics, Moore's law, quantum computing, brain-machine interface

Keep exploring

More from Samim Winiger

More from Rise of AI Conference 2016

More on AGI & Future of Intelligence

Transcript

Auto-generated transcript, not edited. · 4,188 words

Show the full transcript

guys um so ronnie started several companies and is now the founder and ceo of my microprocessing industries and there he's working with aie industry luminaries you could say such as joshua his focus is building cognitive machines and emotional ai i find it interesting that his tagline on linkedin reads either you care or you don't then jan has a background in complex systems and works for the part-time scientists a company that participated in the google luna x prize so i suppose we can literally say we have a rocket scientist here currently he is negotiating the rockets and selling payloads now he's a agi 15 conference coach and as he described is currently founding the berlin exponential institute finally we have danko here he's a brain researcher i almost said recovering brains a researcher because you're transitioning into ai but you're a brain researcher uh he's a group leader at the mark splunk institute of brain research uh and the inventor of ai kindergarten um i hope i'm accurate with this but it's a method for developing

strong ai systems and he is in his work he asks questions such as what does it take to build a truly intelligent machine so thanks a lot for coming in for the absolutely fascinating talks i enjoyed every one of them fantastic so so maybe we can start um you know an easy question um do you guys think that we actually must understand the nature of intelligence in order to simulate or replicate it so do we need to understand intelligence to build an agi maybe we start we don't go here well it's certainly easier to build something once when you understand it and when you don't uh and it's it turned out to be especially difficult to build ai and we have all these ai windows which means that we had all these promises and enthusiasm and people believe that they could build something they turned out for they didn't succeed so actually we historically see that there was a lack of understanding and and there was a problem so it's really i think a great

idea to think on more abstract level also what we are doing and understand what we are doing where we are going um no i don't think that we need to understand the human brain in detail or creativity creativity or consciousness in order to be able to have systems that can emulate that why because we don't understand our systems either our brains but nevertheless we exist so natural history shows proves that evolutionary mechanisms lead to this kind of results and complexity without our you know as supreme leader brain needing to anticipate and plan everything okay well i guess we need to understand intelligence but not necessarily human intelligence um i mean from the design space if you if you look at it only a really tiny part of systems are intelligent and not all of them are human so when we talk about agi we have the distinction between human-like intelligence and human-level intelligence what we aim for is human level so that means it can make generalization has abstract thought can transfer from one domain

and to the another which differentiated from narrow ai but you don't need to have human-like intelligence to use it for some of us who are in ai and our practitioners of trying to build ais understanding intelligence is precisely why you do it you mentioned joshua he's you know famously uh he's not interested i i just like to build stuff uh joshua wants to understand how he and his children work and there's also i think davis has deep deepmind you know that's the most you know short of the universe the brain is the most uh interesting thing to understand so why we needed to build ai i don't know um i mean it's completely conceivable that there's lots of intelligence automation that is completely weird and doesn't work like human intelligence at all and it's still very intelligent and uh we have seen exponents of this like you know the deep blue was the first thing that's that's a completely weird machine it does crazy stuff nothing to do with the way i play chess okay

so much better um but um so we don't need to but you know for some of us it's the reason we do it um yeah i find this this uh all answers give us different perspectives and it's quite interesting and maybe this leads to the second question which is uh because this concept of the singularity is such an abstract concept in a sense and we're negotiating the future um it would be fascinating to you to learn maybe through uh extrapolation what you guys think about this bit on a more emotional level in a sense so um how would you actually describe let's say in 50 years from now the time in 50 years in your views you know desires hopes outlooks if you make it sure if you give us you know a sound bite of use how you would project that reality in 50 years what would you say well i'm really glad that you asked you know what we wish to be rather than predicting because predictions were examples so my wish is

that we have around us really smart robots that do all the work and you never have to feel bad for them because they never complain they don't have real consciousness like we do so they just do the work and you can um people can enjoy their life and go create this stuff and don't have to do things that they don't like doing because they have someone else and everyone can pick something else interesting to do and hopefully a few people will be just drinking and partying and many more people will become something creative and we'll have more science art some beauty and all kinds of amazing things that we haven't had so far partnering can also be a very creative act i don't know um very berlin i think the the steps to singularity happen incrementally and what happens after singularity is also an incremental progress so on the one hand we will emerge with artificial intelligence we are already machines you know our smartphones are part of our brains calculating our you know and

the kind of we communicate we are already cyber if you want we will be more at sidewalks there we will have also robots surrounding us we will have robots in our brains we will be up uploading ourselves perhaps and on machines and travel the universe without being thirsty or hungry at all or sleep for hundreds of years whatever um on the other hand i mean the universe is super complex our societies are complex what do we understand and what does the average person understand and about the society about natural sciences i mean we have this discovery about gravitational waves you know even on the cover of the spiegel magazine but i bet that less than one per ml 0.1 percent of the society actually have a basic understanding of the general uh general theory of general relativity ask how many understood it could could you do some audience participation ask how many understood the theory who yeah who thinks who understands the basics of the theory of general relativity but this is not representing uh

representative foreign for the ones on youtube you have a lot of super smart people sitting here so i'm looking forward for this i mean i i don't understand it by the way you know uh you know oh quantum mechanics would say also less than one percent understand the implications of the mechanics so today already the majority of the society is incapable of understanding what smart people can know if they are interested if they are educated if they had teachers in schools better than those that we have religious religion teachers tell our kids lies and miss about things that happened 2000 years ago you asked what my emotional reaction is to the singularity and actually i'm quite indifferent um because i don't have any idea how i feel in 10 years any less than 50 years or so um i think we make an error when we try to understand the singularity as an event that's like appearing tomorrow if you talk to people and you talk about the singularity with all the real stuff that

is implied by it i mean that's that left it leaves out all the intermediate steps towards it and i guess every step is reasonable if you get used to it so um talk to people 10 or 15 years ago what you think about today's world with all the technology and stuff so yeah i'm not really invested in the singularity emotionally so i i told a very dark story earlier so if you ask me about my emotions about this i hope i'm wrong that would be nice and i think the best way of being wrong while being consistent with the story i told you earlier uh would be you know there would be a panel of a couple of smart nice ais talking about i don't know music or art um i think that's probably the best result we can open a sign of humanity my perspective coming shining through and i i love that it's quite important i think to combine our humanity with technology in a sense um so this is a quite a

good league way in a sense um if you study the history of technology as was mentioned twice in the talks you used some on concepts like cybernetics etc that had a massive impact on policy making and on societal discussions throughout the 1940s 50s 60s up to the 70s now it's called something different control theory etc but it has you know influenced stuff up to today uh you could call the 20 free time rule at google a cybernetic system in a sense um and so what i'm trying to get up is how do you see this if you want to call it the movement the singularity movement ideas influence decisions today and i'm talking about concrete policy decisions uh especially if you look at california and how influential people thinking at google are how much do you think it influences policy well my personal impression is that it has just started influencing policies and that we will see in the future more and more of that and that we now maybe for the first time see

that uh somehow this part of of science is being perceived by by society as important like nuclear energy was 50 years ago when everybody was worrying about it now it's kind of became a routine and now this becomes an interesting thing whether it's a fashion and whether it will go go away i don't know but i would like it not to go away i mean um i mean on the one hand we have like you know politics is influenced by technology a lot otherwise we would not have the nsa scandal so there's a lot of ai and a lot of very smart developers working at the nsa like i thought snowden other people that worked here um so this is part of politics call it power politics and um on the other hand politicians um it's about the government perhaps that's differentiated between government and politics or politicians on the other hand politicians get only elected by the majority of the people when they speak out or represent at least pretend to represent the opinion

of the majority of the people they respect they expect to vote for them so there are a lot of scientific proven facts about reality like homeopathy it doesn't work we know this for decades but our pharmacies make tons of money with it it's betrayal on a mass scale but our politicians don't interfere the police doesn't interfere no pharmacist is you know imprisoned because of selling pharmacy and pretending to be scientifically proven or cosmetic is the same or one another example is that scientists and every expert says that the prohibition of drugs this count is the disaster it doesn't work it doesn't inhibit um the trait or the consumption it creates wars in latin america and afghanistan creates this mafia structures so why doesn't don't we uh legalize it because the majority of people don't understand these complex issues so they think oh it's forbidden it disappears so interesting yeah okay the question was whether the singularity movement is influencing if it has real policy implications right now okay yeah i guess it did and it

does um eminem started usually in academia and plenty of people made decisions toward the singularity or something like that they usually have to use another term for that because it's unpopular and you have to be careful could you maybe mention one or two other terms so people can get an idea how it's called as well or no it doesn't matter really no no they don't have a substitute um they just work on specific aspects that lead to the singularity as they hope like cognitive robotics or something like that they always have some terms further and the next thing are policy makers in incorporate domains i mean usually they produce stuff and they're aware of where technology is leading what implications it has um in 2008 i think the cto of intel mentions the singularity explicitly was it to play it well something like that that we merged with the machines it's only 2050 that was his stand on it and now they try to make inroads into the policy making the links process and so

on um i guess the guys from the singularity university try to to include governmental decision makers as well and we have to see whether it works i don't know yeah i think that was pretty complete um a pretty complete overview of what's going on it's not happening yet but when people like elon musk or you know other people are putting lots of money and making it happen and usually money works so let's see all right um yeah very interesting thank you um so i'm going to try to bring the discussion back to today uh because you know we can we can move we can look into the future we want but we live today so it's interesting to negotiate that a little bit more um so we've seen uh just this week another video by google well it's robotics austin dynamics company and i don't know if you guys seen it but it's the most creepy life like robotics video ever made i would argue and it leaves this kind of uncanny chilly effect on

the viewer so quite an interesting video if you haven't seen it to watch um and so the the feeling creeps up that we are having right in these couple of years um some kind of very interesting dynamics in the short term in terms of ai technology and robotics technology coming together um what what technologies do you find most interesting as of now so they're not really projecting too far you know is it is it uh you're deep learning or are you going on the fringe with more genetic algorithms or what do you actually find most interesting apply you're asking a brain scientist i have to give you a breath please that's fantastic so what what is being developed now with brain science and what we are really excited about are our special type of electrons that can record hundreds of thousands of neurons similar it's basically an entire array of sensors that you can put in the brain and you can actually you don't record anymore electrical activity at some point you record entire electric

field you can get them or very soon we will be able to get entire image in real time how the electric field of the brain evolves with the precision of 20 30 micrometers basically we can see how electric currents flow in a cell or or out and that's that's uh every neuroscientist gets chills when you talk about it i think technology doesn't need to be super super sophisticated to be exciting and interchangeable extremely tremendously yeah when you look at low tech like smartphones with an internet connection those devices are super cheap now and becoming even cheaper so every poor person involved the majority of evolving users and this will change those societies in the next 10 15 years so much more than the last 50 years changed western europe so this is it makes them you know violence decreases etc etc societies become more intolerant and this is a huge step forward in the quality of life of billions of people so on the other hand i mean for me personally what i love about

high tech is to see news about new currents uh to treat hiv forever i mean to solve the problem or alzheimer's demands is going to be curved i mean grandfather's diapers i mean risk has parents or me um i'm sure that we won't have this kind of diseases in the next 20 years and for the wealthy part of the world the high-tech um counts is relevant because they used to have the low tech already and for the majority of the society their global global population low tech is more relevant now because they don't have access and money to buy okay well actually i have no specific technology i'm especially excited about it would be really long story so nobody don't do it please give us an executive summary an executive summary all right okay lab automation i mentioned it already something i'm really excited about like for medical research and for new materials is something i i definitely need for for some stuff i mean you mentioned um earlier that we will fly through space

as not as okay i mean um we're already building the beta for that so um you you have fields like cool brain emulation where you have um increasingly better brain imaging um you have an increasingly better understanding of the human brain it's also true for the for the human body um we will start with simple things like brain machine interfaces neuroprosthetic tissue replacement and so on and all the stuff that you can apply ai to to amplify the research so this together with things like 3d printing um for new materials that are more robust more more light or anything better um yeah it's making inroads i mean those are things we will see on on earth as well it's not only for space applications of course so yeah that's something i'm excited to summary very interesting so i'm going to have to limit myself to apply the ai because there's just so much awesome technology coming up but in the field applied i happen to be interested in ai in manufacturing um but i just

sort of aware that nobody shares that taste so um for the more one of the things that's going to happen in the next couple years and that will be super exciting is the disappearance of user experience uh in human language interfaces um you know that's that's just going to be huge because it means interacting with technology across the board wherever we do that it would be so much easier it would be so much more intuitive than you know the crude approximations to something that you actually can use that we are now doing with very good units um yeah i think from my personal view i agree with all of you in a sense and especially i'm very much interested in the human aspect in the next you know five years that said i mean the five-term near term view it's an interesting article came out last week and one before a year ago it seems that the semiconductor industry has concluded that moore's law is now finally slowing down officially uh there's an article by

intel there's another one by the semiconductor corporation the the group that encompasses all the companies in this space and it's indecisive there's a lot of chats online about this but obviously the number one trend that singularity inclined people usually tend to quote is moore's law so that's slowing down what are the implications here are we more in a i think somebody mentioned the sigmoid instead of the the lock curve are we uh are we on s curve or are we on the locker what's the deal well i thought that every exponential has to at the end if you wait long enough end of the sigmoid because the universe is not infinite time is not infinite so sooner or later you have to slow down uh if it's happening now it makes sense on my computer i have a computer at 170 years old and i think a really neat one which means that you know computers they did not improve much as they did before uh but your your other aspect of the question was

how is it related to ai the prospect like how does how does this inform this exponential thinking if apparently one of the trends is slowed down well i actually i don't think it really matters a lot because you know what what kurzweil like to compare is the computational power of a computer versus computational power of the brain and and finding a point when they intersect and then presuming that at that point the computer is as intelligent as the brain but computational power and being intelligent are two different things right it's like saying you know we two three hundred years ago well we need just enough paper and ink to make it cheap so everyone could write and then everyone will get power everyone will be we'll be you'll have shakespeare all around because paper link was some time ago really expensive really luxury but today now everybody has a paper running and there's no shakespeare not a single one i guess so having enough computational power if we had you know even accuracy of claims

that he accelerates even faster than this one than exponential and doesn't matter if it started if we have really really really powerful computers it doesn't matter you have to feed them up with knowledge and that's what i was talking about all this evolution millions of years of extracting and compressing knowledge into something really powerful that actually can be in thousands that's an exponential that started four and a half billion years ago when earth was uh formed as a planet and it continues and will continue with the same pace and that's that's the exponential we have to look for intelligence and we cannot accelerate that one and this means that we can right now bring machines only to our own intelligence and then we all together can continue this much much much lower exponential than the warsaw and we there's no way wrong and before i answer a question i have to speak against what you said you said you don't have more pictures today i am convinced we have much more people living up to

their creative potential is higher than shakespeare and mozart because back in those centuries you know 99 of the most talented people even among the most hunted people they had to work for 20 hours in the field and they died when they turned 30 and they married when they were 10 years old and 15 and then they died 10-15 years later and today we have so much more people that can live up the the reason is why we don't we see mozart and goods and beethoven all the time in the radio and how many in berlin is because the brand they have a strong brand in classical music and it's hard to break them and a lot of groupies worldwide okay coming back original question um moore's law yeah most of all is just one segment of the reality and of the progress technological progress it's an important one but it's one segment and we have also to look at the times time span so we are reading about that slowing down but at the same

time there are some crucial developments on other fields like quantum computers and quantum computers kill everything and there are more and more articles about the development and breakthroughs in quantum computing and some computing will look i will let's make moose law look like a fault so

Rise of AI 2027 · Berlin · May 2027

Be in the room for the next one.

Ten years of talks started as conversations in one curated room in Berlin. The next edition brings 300 selected leaders together in May 2027.