David Wood: Exploring AGI, AI singularities, and the future of human aging

guys welcome to another future wisdom pod we have David wood with us he is a futurist Catalyst author and singularitarian which I is a word I had a tough time to pronounce uh basically means that he's probably 5 to 10x as obsessed as I am with Singularity which is a topic that I write probably five to seven percent of the time David welcome to the Pod I

think we're gonna have a cool convo today can you looking forward to it and uh tell us a bit more about yourself my background was in the early days of the mobile Computing and smartphone Industries I spent 25 years writing software coordinating the writing of software developing Consulting teams working with our clients so they could build what became the world's first successful smartphones that's companies like Nokia

Sony rxn Samsung Motorola and others so I was somebody who envisioned the future in part as early as 1995 some of my colleagues and I were writing about a world in which smartphones would become more important although we didn't anticipate quite how important they would be let's be clear so they took longer to arrive than we hoped that when they did come my goodness they have changed

the world more profoundly and the reason I'm talking to you now is I have the conviction that what happened to smartphones is happening to other Technologies such as reprogramming our biology rather than reprogramming our phones and what enables all of it is the growth of artificial intelligence right let's start with this yesterday I wrote that actually I don't know if it's been published yet I think it

will be tomorrow but I wrote a pulse saying that smartphones cannot be compared with AI because AI is what invented uh smartphone wealth intelligence that is and when we create intelligence well we create the tool that actually creates other tool do you agree with that statement in part yes intelligence is the secret source of humans it's what we have used to conquer the planet it's what we

have used to go all the way to the moon it's what we have used to understand what's happening in the forest regions of the universe it is the thing that enables all the other progress and the more intelligence we have the more tools we can create to allow us to become even more intelligent smartphones were a bit of the same you know as people have got more

smartphones they could do more things as well smartphones where an augmentation to people's intelligence I used to say a phone is a smartphone if it makes the individual using it smart.com so smartphones haven't augmented our memory they have put more information at our fingertips they have often helped us sometimes even listen to us sometimes they give us feedback sometimes when you're about to click on a link

your smartphone or software running on the smartphone will say hey don't click on that link you think it's from your bank it's not from your bank probably it's probably from somebody who's trying to steal your bank account details so smartphones have been part of what's made us potentially more intelligent but AI is going to change things even more fundamentally than smartphones have done yeah I do think

so as well um first question which is somewhat philosophical is there other forms of intelligence like I've discovered for example going to meditation Retreats and so forth that sometimes letting my gut and my heart decide for me although there's brain cells in there um well in the gut precisely and in the heart too they've found some but are there other forms of intelligence which we don't know

about I'm sure there are in fact this is why it's quite exciting and also scary about the new kinds of artificial intelligence Jeffrey Hinton who is sometimes called the Godfather of deep learning said recently that for a long time he and others were trying to develop AI to understand human intelligence better so they had some ideas what was happening inside the brain their layers of neurons and

they tried to model it but they now realize that the way artificial intelligence reaches its conclusions is significantly different from probably anything that's happening inside of our brain there are processes of thought there which we could describe as alien Minds now our own minds have got aspects which are hidden and unknown we are often intuitive we have the parts of our thinking happening subconsciously our brains tell

us things which have been calculated inside the brains but outside of our conscious introspection sometimes if we introspect vague thoughtfully very carefully we can Glimpse what our subconscious is doing but there is a lot happening outside of our conscious awareness right and there is intuition of various thoughts happening in these AI systems as well but it's a different kinds of intuition almost certainly and that what makes

that's what makes things scary but also exciting you know when these systems get things wrong often we're puzzled how could he get this wrong they can be very good at recognizing pictures of different breeds of dogs but if you change just a few pixels suddenly it will confidently assert that's not a dog that's a giraffe you say what it doesn't look anything like a giraffe but somehow

that thinking has gone down a wrong track and that shows that we're not aware of the processes so if we can make these mistakes that's also chilling because it may be that we think we're getting on very fine with these AIS and one day it's going to look at us and say hey you're just a pest I'm going to get rid of you which is another topic but I you open like five to ten doors there let's start with um let's let's start with uh the the last one that you you just mentioned about uh basically AI hallucinating um can we Define the term hallucination like meat obviously hallucination well you get high sometimes and on LSD and you start seeing things um that's hallucinating yesterday I read a post on Twitter of a guy

saying that his child his two or three year old child was hallucinating and inventing stories now is there like examples of us humans uh hallucinating stuff like in my work day for example am I like completely wrong about something is that the definition of hallucination and can we compare that with uh gpt's hallucinations there is a broad similarity but just as it was saying before the similarity

Isn't So exact often when we hallucinate we're not seeing something completely uh wrong we're seeing a few points in the real world and we're jumping to a conclusion you know we humans we need to be quite good in the past and recognizing faces so sometimes when you see on a bit of toast a few marks you think hey there's a face there and you can think hey

this is a face of Jesus Christ on a piece of toast or we sometimes see faces in the clouds so we are seeing part of a face and the brain is filling in the rest and that filling and the rest is an hallucination it's like imagining these AI systems are doing too but the the question is like how are we qualified as humans to really know that they hallucinate is my question so we can test this is what science is about you know people come up with opinions they say I see this somebody else says I see that well then you investigate you measure things more carefully so these AI systems they're coming up with uh claims sometimes they are claiming there was a notorious case recently there was a claimed that a particular Professor

a real Professor has been involved in a allegation of sexual misconduct and the article that this came up with had a reference in it it says here's the online source so people say gosh this professor's terrible he's evil but then you looked and that reference didn't exist there was no such article so that's how we can say sometimes when things are hallucinating you go and you look

for yourself you weigh things up You observe carefully being aware that sometimes your own observations might be misleading too but that's what scientific process is about trying to eliminate that bias trying to have multiple people independently verifying it so that's how we can say that some of these systems are currently sometimes hallucinating right okay then I want to touch into various points I don't know why my

brain kind of lost tracks I'm taking notes here because we're getting deep kind of real quick but um okay so intuition intuition what is your definition of intuition because the brain is this interesting tool right like I sleep sometimes and I wake up with business ideas uh which are pretty good actually they're getting better and better so it's that's quite interesting then intuition isn't this like Consulting

your hardware and pulling out memories that United together um isn't that intuition well intuition is a pattern matching that happens subconsciously or beyond the introspection so there are logical things happening all the time indeed we can sometimes wake up with the answer sometimes the ideas will pop into our head and it's clear that there's quite a lot of thought going on behind them sometimes it's simply that

there's a pattern that our subconscious mind is recognized and is tugging to us so I wouldn't say it's just one thing and precisely this is intuition I would say most of the processing that happens in our brain is outside of our Consciousness so there's many kinds of intuition and we shouldn't uh we are afraid of intuition intuition is often wonderful but nor should we put it on

a pedestal and say well I have this powerful intuitive feeling therefore it must be true our intuitions developed for particular reasons historically over evolutionary time skills our intuitions say made sense often in the environments that we were living in there were wise intuitions but in our modern world we have different challenges so one of our subconscious intuitions is that where there's a sweet things available we'd better

eat them because you know there may not be so many things in the near future nowadays our environment's changed our environment is full of supermarkets with far too many sweet things and our intuition is saying is hey eat this is misleading us so we have to use our rational brain to say hang on this is no longer something we should indulge in we also tend to operate

tribally in much the same way was very useful to be able to identify with groups of people because we could not survive by ourselves as individuals we relied on support from tribes so we have instincts to help us identify tribes and to cheer along tribe members accept what's said by the tribe Elders even though partly we may not agree with them well these intuitions again are very

dangerous in their more complicated more risk uh Laden World in which we operate today right in some scenarios use your intuitions in most I would say 80 percent of the time use data uh there was this Brian Johnson um guy I'm not sure if you heard about the enemy biohacks massively but he's 100 data and he's zero percent intuition which is interesting but I think we have

to listen to our intuition but then we have to back it up right ideas come from our intuition but then let's verify we shouldn't say original it's like fast fast thinking and when you're CEO there's a lot of of that so and how to get better at intuition I would say like more data more quality data in your brain and taking care of the hardware in there

so that um you can make better decisions which kind of leads to the next the next topic but before we do the next topic is longevity um I want to talk about one comment you made about the similarities between the brain and neural net or the lack thereof kind of um my understanding when we look at the brain is kind of MRIs you know with these brain scans and then we see the zones in which people think a kind of heat map although it's not heat that they measure its uh electricity which I think could generate heat but um and then the neural Nets which is kind of similar uh just that the brain is in 3D and the neural Nets to me it looks somewhat similar um yeah but there are differences too for

one critical difference is that the calculation in the brain typically only uses uh each calculation only uses a small subset of the brain it's a so-called sparse Network when I say a small subset I don't mean localized do I mean there's a set of neurons throughout the brain that are involved in that thought whereas typically in today's artificial neural networks it's a decision typically involves every single

neuron in the brain in that system so many people are looking at creating sparse networks so they will still have large numbers of neurons hundreds of millions billions hundreds of billions even but typical calculations will become more efficient by using just a subset so that's one way in which people are trying to learn from the differences between artificial neural networks and a real biological neural networks right

because when I ask you a question you kind of scan your brain right it goes in various areas in various boxes and then it Returns the answer with the highest weight kind of the one that you would be more most certain of which is like the one with the that weights the moles that is the most solid that is the most insulated which is is that the right way to think about it that is a very good way to think about it yes I'm sure when you ask me a question different parts of the my brain are saying well I could answer it like this but another part of the brain is saying hey be funny another part of the brain is saying why don't we hurry up and talk about longevity let's do a

bridge and uh well indeed there's another part of the brain which sort of decides which of these is the right approach so you just did both you made me laugh and you made it segue to longevity which is the book at your uh well the left on the screen the abolition of Aging do you want to live forever what are you doing to live forever and else white are you interested by the topic three questions in one I'm not sure I want to live forever but I do know I'm not in any hurry to be extinguished there's so much more I want to do so many things I want to experience I want to see my family members grow up became I became a grandfather two and a half years ago and it's really wonderful

seeing my grandson learning and so I'm not in any hurry to stop existing existing is Good Wife most of the time is very good right now eventually I might get bored don't expect anytime soon eventually I might say right switch me off even then I might say Wake Me Up Again by the way if there's something really new interesting happening so life is good life can be

wonderful and how am I ensuring that people like me and by the way I'm very interested in everybody having access to this it's establishing of every Aging for everybody not compulsory and if people have a strong desire to age and die that should be their freedom but I think when people see the possibilities of people being eternally youthful they're gonna recalibrate their thinking and many of them

will so you know what maybe this Rejuvenation stuff is better after all so I want to make that available to as many people as possible as soon as possible right and part of it is with individual things we can do now like restricted uh calorie consumption that's one way which might add a few years of Health to somebody's life there's various other supplements you can take but

they can only extend your life a bit there's other things like making sure you get good quality sleep there's things like having a purpose in your life there's things like having really good friends that you care about all of these are statistically correlated with typically living longer and by the way avoiding very dangerous activities and so on but that's only going to make a bit of a

difference if we really want to significantly extend people's lives we've got to hurry up with uh biological Research into the mechanisms which already have a lot of suggestions behind them it could undo the damage of biological aging you know as we live we accumulate damage in all parts of our body at multiple systems and when we're young these are pretty good at repairing them most of the

time you know we fall down we skin our knees we get up and within a few days the things are being healed when we're older similar things happen well you can die from a fracture to your hip when you're old enough so the intrinsic repair mechanisms eventually start working so what we should do is we should repair the repair mechanisms we should graft back in more stem

cells because stem cells are what it looks after other things we should figure out where the genetic reprogramming can be applied or if not genetic reprogramming epigenetic reprogramming which is a bit more technical it means how the different genes in each cell are expressed or not expressed and that's to do with the accumulation of various tags or methyl groups that can interfere with how their genes in

each cell are expressed so if we undo that damage it makes the cells younger again and it kicks back into the life these natural repair mechanisms so that's the kinds of things that are possible we can see some of these mechanisms working well in various other species some species the early age at all of course they get chronologically older but even if they're older they don't become

weaker whereas we humans every eight years when we live from about the age of 35 onwards doubles our mortality I mean it doubles the chance that we will die in the next year and so when we're pretty young is a very low chance maybe one in a thousand chance that we'll die in the next year when we're 30. by the time we're 65 it's a something like 100 whatever the time we're 80.

I can't remember the exact numbers but it's growing it's growing a lot so but other animals aren't like that which ones tortoise or turtle is one there's the naked mole rat is another various species of fish Ferris whales now they might have a slight increase in mortality but it's nothing like the exponential increase that we humans have so there are things in their biology we don't have

and we are quite good compared to many other animals in terms of living much longer we have a pretty good lifespan for our size of animals but we're by no means the champion and we could borrow some of the techniques that are in some of these other animals that's part of the theory so if we are interested in living a lot longer we need to campaign to

get that research funded better and we need to work out what's holding back that research and to deal with these blockages and that's a lot of what I do yeah and I could get into a bunch of longevity questions but we don't we don't have much time because these species some of them when you talked about the fishes uh to meet rang like prehistoric and also some

of these they don't get exposure to light I think they live like really deep downstairs so there's pressure uh tortoise you know well they're just very very slow they take their time and so forth I think their whole metabolism is is slower compared to mice for example um but I want to talk about one thing you mentioned and you said if I'm ever bored then yes like

freeze me and so I can restart again but I think that's like biology um being bored is kind of a mental phenomenon I'm a biohacker and I very much believe that yeah it's like the the health of the brain and the blood flowing there and all the biochemicals to make you happy and so forth which is related to a polls that I wrote well I didn't post it yet but it was that I see somewhat of a political problem and older folks that they do have wisdom right but most they just stop growing mentally uh what whether it's biochemically or whether it's just that they're comfortable um they have lots of wisdom but they they stop wanting to learn more for example uh Biden with the AI thing the other day well he said oh

I had someone try the tool or when it came to signing up new contracts he had piles of books next to him and like dude do you ever like play with technology is is that interesting to you and I feel that a lot of people in power they just stop learning and they they don't play with tools like artificial intelligence and so forth and what so my question for you is what is the equilibrium between wisdom and um learning new things when you're at Power and we might also transition with the role of AI into baltics and what's going to meet for us in the next couple years so I'm not an expert on Joe Biden so I won't comment on him but I will say there are some 90 year olds with very Nimble

Minds you know some people do become more fossilized in their thinking but there are others whose minds are endlessly curious even into their 90s and Beyond so there's no inevitability of Minds becoming fossilized you were saying earlier you know we can meditate we can look after our minds if we cultivate curiosity cultivate openness then we can retain these qualities and capabilities so I don't accept that aging

will necessarily mean our minds slow down I'm very interested in the Rejuvenation of our brains as well as the Rejuvenation of our bodies and indeed the Rejuvenation of our minds and our spirits and I want to find out which factors make people become close-minded in the sense that you described and I want to ensure that we don't become damaged in that way just as I don't want

my body to be damaged with the kinds of things I spoke about earlier so there is possibilities for political leaders to remain Nimble even in their 90s there was a prime minister of Malaysia recently who was prime minister many years ago and he came back for a while in his 90s and his mind was still Nimble so there are examples like that in the UK yeah 96

had a very active mind and sadly her body was no longer capable and so as we know she died last year in terms of the more general questions about politics actually I see this as incredibly important if I ask myself what might prevent us reaching this better World in which AI is solving many of our problems in which we are living much longer the thing that is

most likely to stop us getting there is our decay in our political systems in which we no longer have an open society in which we no longer have a respect for science but instead science is determined by whatever is good for the guy in Power you know and if the president says such and such is the cure for covid everybody will say yes that must be right

or the opponents will say no that mustn't appearance so they decide scientific questions on political grounds rather than scientific grounds the Soviet Union they fell under the spell of biologist called lysenko for many years lysenko had a sort of Marxist theory of biology which is based on the earlier lamarckian idea that if there was a struggle the crops would grow stronger in reality there was terrible fireman's

course that was not science by science that was science by politics so we mustn't get in that state and there are parts of the world now where there is more populism and populism I'm not against people being interested in large scale and politics I encourage it but what we're getting instead is uh system in which people imagine hey we are the people and we are being misled

by an elite these scientists these technocrats we're not going to listen to them we know better and it sort of have people trumpeting their own ignorance and imagining it sending on an equality with a careful thought of people who've been studying this all their lives so we might end up in our bad politics which could stop us making the progress so we must improve politics and indeed

as you said AI can be part of the solution as politicians are speaking in real time the AI can annotate their comments on the screen hey this contradicts what the politician said last week hey this is true we should trust what this politician is saying because and it's gone and has checked it and so partly the AI will help us to know what's reliable partly the AI

may go beyond that it may suggest that very interesting compromises or a synthesis so you've got two sides that are squaring off and unable to agree on a very controversial topic I don't know gun control or the war on drugs or whatever and this AI will come up with something that will make people and all sides stop in their track and say you know what this is

a better solution than anything we've come up with so far so I think AI could help in that grounds too so I look forward in the UK to having something called the house of A.I rather than the House of Representatives or the House of Lords and we might still have a House of Lords but maybe there's a third chamber which have got ai's reviewing their legislation and

suggesting new ones humans should still be in charge but the AIS can become very much more involved in helping us to have better politics and with better politics we'll have better allocation of public funds we'll have better sets of incentives better regulations after all the politicians are really struggling to know how to regulate some of these fast changing Technologies so AI can help they can make suggestions

as to how these regulations and incentives should be put in place right um I I thought about that AI chamber making decisions just like we're making decisions to conserve some animal species which is weird and then I had the thought oh they might be smarter than us so they might not do it that way but anyway that that was a thought another one was populists are dangerous

if they don't have any back end right if you're all front end and have no back end just like in business then that's dangerous so it may be fine to to use marketing and sales and populism to get elected but you need to have a back end um and then yes the the system of right or wrong or or truths the truths GPT we could call it the danger is just that some people don't want to know the truth you know or will constantly deny the truth as we're seeing it with uh conspiracy theorists you know like the movie don't look up uh so that that could be interesting but um David how much more minutes do you have because I have one last question about singularity in your predictions let's talk a little bit

longer and uh one thing an AI can do is address this issue that people don't want to hear the truth often you need to find a different way to talk to them so instead of just brow beating them with things that make sense to you which are alien to the person receiving it they are unable to think it you need to find ways to express your opinion that makes sense to them too you must find a common point and then build on it and then help people to see uh from a different perspective that requires a different kind of intelligence and people who are high IQ are not always good at that social intelligence so that's what I think AI may help as well wonderful point I mean while you were uh talking that the

thought I had well yeah it's hard for me to convince people sometimes because we're so abrupt you know like this is the cold hard truth EQ is incredibly hard especially as your IQ scales it becomes like incrementally harder and yeah you're right I could because I could picture um a GPT saying given your background I do understand that you think that way but even that is not

good EQ right it's like super it's and we were talking about other forms of intelligence I think that's one right there you know to really have the capacity to have empathy for someone's really hard especially when that person is messing with your ego you know and I think us high IQ folks were always trying to see the truth when we try to seek the truth and someone

is trying to blur the picture we get kind of angry right so super good point there when I talk about your predictions um for 2025 let's start with that one two years from now then uh 2035 which is in a lot of years from now decade plus from now and um make sure to talk about Singularity because Singularity is this at first it was clear in my mind but then as I got more information and thought about the laws of physics that stuff cannot accelerate forever if we if we can even use the laws of physics in that case it went a bit weird you know is this Singularity thing like related with time I think the brain really has a hard time picturing time space time kind of is it like accelerating super fast

do I get in slow motion in that new world do we we have invent on do we think differently so I think if you could give projections as per what will happen according uh to to David uh in the next couple years the next decade that would be nice and if you could also Define your precise visions of singular RT and that of others that would be very helpful Singularity is quite simple it's when AI is more intelligent than humans in every way so today AI is more intelligent than ours in lots of narrow ways it plays chess and go better than ours plays poker better than us he's able to navigate from A to B better than us but there are many other things in which it falls down if it goes outside of

its programming it doesn't have the ability to generalize but when it is more capable than ours in every way that changes everything it's a singularly disruptive moment because then the fate of the world is no longer on our hands the fate of the world is in the AI science as you pointed out the fate of other animals on this planet is dependent on us you know whether

or not the chimpanzees survive is not in the chimpanzees hands anymore it's on what we humans do and so after the singularity we humans may still gesticulate we may pray to the AIS we may bag with them but ultimately they will be in control and that changes everything so it's not about moving faster than the speed of light or moving backwards in time or whatever there are

physical analogies the main physical analogy is that you can't see clearly what's going to happen next that's when our extrapolations break down why can't we see because we don't know what these super intelligent beings are going to do I'm not we don't know what's going to motivate them it would be a bit like you know if you and I were chimpanzees 5 million years ago having this

conversation and speculating what bigger brain day humans might do in the future what could we imagine we might say hey we chimpanzees we like bananas maybe these creatures will have better bananas we chimpanzees hey we like sex maybe they might invent cleverer kinds of sex right but the chimpanzees would never have said hey there's going to be Pythagoras's Theorem there's going to be creatures that can see

the moons of Jupiter there's going to be creatures that compose Beethoven's Fifth Symphony it's beyond the hour their conception and similarly what these AIS will do is beyond our conception so that's the sense in which there's an analogy with a black hole right where when's it going to happen if anybody tells you categorically they know when it's going to happen don't believe them the reason we can't

say is we don't know fully when AI is going to reach that point of being smarter than us because we still only understand dimly what our own general intelligence is that's our discussion we had earlier you know there's a lot going on in our brain that we're on the dimly understand despite everything we don't understand what's going on in these AI systems either so the best you

can do is give a statistical spread of dates and I would use I used to say that I thought maybe 2045 is a reasonable average time for when we might reach this now I've brought my time scales a lot further forwards maybe 2030 is a 50 chance that we will have smarter than AI smarter than humans in every way but it could be even sooner you know if it's about 25 probability you want from me then maybe 2026 maybe in a few years time why because the recent breakthroughs with large language models sort of got even larger that are showing some emergent features which the designers did not expect if you couple that with combinations from other AI so you couple there's neural network systems built by Transformers with other layers built out of classical

symbolic reasoning or evolutionary algorithms and so on you may have breakthroughs much more quickly so that's why I can't categorically say such and such a thing will happen in 2025 2035 or 2045 I say we must watch carefully and we must put much more resources on to figuring out what's really going on in these systems and how we could constrain if possible I said we couldn't anticipate

exactly what these things do but maybe there are ways to constrain them this is the so-called control problem of the alignment problem they're slightly different I write about both of them in my book The Singularity principles you can see up there on the corner and we need much more people studied in this and discussing it thankfully there are more people in the last few months who are

attending to this because they are also getting shocked at the capabilities of the new systems but they're still nothing like enough so my um learn more about this because whatever I may like to do in 2035 it may not be possible if we fail to manage wisely the growth of AI right because what we're seeing and you weren't lagging the last couple of secs Dave but um

what we're seeing is well it's not like don't look up the the movie by dicapriochus which which was quite good I recommend it to everyone but it's people saying just another fat right um well mostly in the business Community is talking with CEOs and so forth and it's just like back to normal right uh I'm just gonna go back to flipping buildings and AI is just this

other thing it's just another smartphone right so and when I talk to people like you obviously spoke with other smart folks um and yet your opinion is very much valued on the topic I think there needs to be like way more conversations When you mention like monkeys eating banana well made me feel really dumb you know it's like if we're limited with or current brainware to to

think more but we can't because I think we have like Hardware limitations and so forth but I still thought that these monkeys they didn't really know that they were conscious or not then I had um another thought which was basically like these things once they um understand and uh can rapidly progress to to space and time well yes they might completely modify the laws you know and

another plausible scenario could be that they could just put us in some simulation you know just like if nothing happened then the other question I had is do we put monkeys and simulations not for now um and also like why because we don't identify them as Consciousness and there's so much so many debate about Consciousness and and like how how it happens and so forth so what

what do you think about these things because me when I think about Singularity I see like time slowing down or it would probably be seamless in another world but it's just so hard to picture and that kind of gives us headaches so at one point of the the Spectrum we're like oh let's talk about it let's have discussions let's grow neuron cells but at the other we're

kind of swimming in the total darkness so how do you play with that well you're right that we actually can become smarter with the help of AI so just like the monkeys couldn't conceive of things but humans can make their environment a bit nicer AIS can help us to understand more right if we are using the AI as well it's like I pointed out the AIS might

help us in real time to know which politicians were telling the truth and which were telling horrendous falsehoods similarly the AIS could enhance our brains so they may also help us to know what to do to be better meditators to clean out our brains from junk to detox them and maybe even to augment our brains and the kinds of way that Elon musk's neurolink company is experimenting

with in which we could have part of our brain connected to some external brain cells so that's possible it's one other point I wanted to pick up in fact there's certainly you made so many interesting points the analogy with don't look up which I think is actually quite profound because spoiler alert if you haven't watched don't look up and you don't want to know what happened then

skip the next part of the recording but then don't look up after a while people say hey maybe this is a problem and we maybe need to do something with this Michel that's coming in meteorites coming in and they have a plan to deal with it and guess what then some other people say hey we can make a lot of money by mining this meteorite you know so they say let's not destroy the meteorite yet let's get some incredible trillions of dollars of value from it and they mess it up and the world is destroyed it's the same with AI many people are saying well you know there might be an issue here but my goodness we can make trillions of dollars in the meantime so let's fix the alignment problem later on so in

many ways it's a profound story I saw it and thought it's about climate change but now I realize it turns out that it's applicable to right whether or not we are seriously looking at this and then finally about what these AIS might do to space and time so I fought against your suggestion that the singularity meant itself a change in space and time but are you all

right these more powerful creatures than us might be able to do some space-time Engineering in the way that we can't they might be able to reach back through time and copy Minds from the past into the future they might therefore be able to resurrect the Dead and they might therefore allow frb who has ever lived to be reborn which is an astonishing religious vision this is what

the Russian cosmists aspire to do and their Nikolai theodorov in the 19th century you didn't imagine AI is helping but they imagined at some stage there would be technologies that could bring back their memories and maybe even bring back the consciousness of those who had lived before so that is a possible future you and I might die in the next few years because of a nuclear war

or whatever but enough of humanity Will Survive and they will eventually build these agis artificial general intelligences which will eventually do space-time engineering and copy our Consciousness forward from just before we die to our Paradise in the future I don't write that probability very highly maybe one percent or so but it is something that I'd like to study more why why do drugs when you have futurists

you know you're blowing my mind it's like it's taking left and right and I I guess some people could have headaches from from that uh disclaimer here but one last question um just to challenge a singularity Vision are there limits to the speed and quantity and skill at which these AI good scale um yeah there's chips right uh but I at some point maybe like if they

grow that fast they would literally become particles in the air or atoms so is there a hard limit of a physical limit or yeah the laws of physic would they actually stop the growth of that thing in space-time or is that far-fetched there are limits people have calculated them and we are very very far from reaching these limits I think it's Seth Lloyd if I remember right

who has actually calculated this and there are many many many orders of magnitude of growth of intelligence Beyond where we are now right now how quickly could AI self improve that's a slightly different question okay and that depends on whether we are constrained by Hardware or by software if it turns out that we need to wait until this much better Hardware that may take a while but

if it turns out that we already have good enough Hardware to to support super intelligence then we might get GPT 5 created say and gpd5 which can already write code gpd4 can write code gpd5 might write better code gpd5 might write the code for gpt6 overnight so we switch on gpt5 and then we wake up the next morning and GPT six is here and taking over uh

taking over the world and that will be possible if the constraint is software rather than the constraint being hardware and I believe it's very likely that we already have enough Hardware to run much more efficient algorithms than we've already got and some of the investigation of what's going on inside of these neural networks suggests that there's an emerging discipline called artificial neuroscience real Neuroscience figures out what's

going on inside the neurons in our brain from their inputs and outputs it's quite difficult because we don't like to cut people's brains open when they're still alive artificial Neuroscience is looking at what's happening in these hundreds of billions of neurons and it's quite complicated to start off with it's just a massive numbers but largely people are making sense of some of it and what they can

see suggest that these systems are very inefficient in some aspects and so we can make them much more efficient that's why it's possible that when we get a bit more advanced with our AI systems they will debug themselves and they will rapidly improve it's not inevitable it might be a slower thing but there are scenarios at least 50 percent that when we reach a certain level we'll

suddenly go boom in the word Sevilla udkowski and but because that probability is there I think we've got to think hard before we get there because I'd like that foom to result in uh sustainable super abundance for humans with their AIS looking after us and they might feel that we are wonderful creatures despite our flaws that they will fix some of our terrible flaws they will guide

us to grow to our true potential that's what I want to see a world of no aging a world of enhancement a world of unimaginable improvements in Consciousness and flourishing and we could get there if the AI is configured right but if you mess up that configuration all bets are off well if I think if it's creative and intelligence it would configure itself in any pants which

is to me alignment seems impossible and and then the question is well I think it will almost immediately create super abundance and if so why would it need to kill us in the process is another thing and when I listened to you it was like a god this AI I do I can't picture that but yeah when I heard you say uh the human nature that's always a complex question it's qualitative it's a quantitative question of who is bad and for how much and also it's really subjective what is bad you know in the mea regions well women's have less right than it's totally normal there while our Western sieves have total different values um but let me ask you this this final question Dave and thank you for today because you you added so

much value I'm really grateful for that what are the probabilities that uh one Singularity happens where proof gone vs um we live in this super abundant world with a a benefactor a nice glad that wants or good can you pronounce yourself on that or try yourself the best the best we can do is just a vague hand waving I think it's more likely than not that we

will have a positive singularity maybe 60 percent positive Singularity 30 negative Singularity ten percent that somehow nothing like this will happen and would just be muddling along slightly more in a slightly faster slightly uh more advanced in some way slightly backwards in other ways I don't think that's very likely because I think the technology is so Advanced and we're not far away from it reaching that Tipping

Point and so we've got to push harder for the positive outcomes and push harder to avoid the bad outcomes push harder to make sure the positive outcomes are more likely that's what I meant to say can we align it can we control it well there are ideas and the idea may be that we don't allow the super intelligence to have its own motivation its own agency we

design it instead as a tool and we avoided having any sense of Consciousness or volition is that possible perhaps and that's what deserves a lot more attention thank you very much for your questions it's been a fascinating conversation thank you Dave where can people find out more about you so you can google David wood futurist and you'll find my details the podcast that I do London futurist

podcast amongst other things or you can find me on Twitter as at dw2 that's to the numerical digit dw2