Jaan Tallinn - Making Predictions, Biotech, AI, Criticisms and Social Systems - in Hong Kong



you it's extremely important it's it this is what actually makes you an effective predictor you don't do that you're just you can only edit coin clips or we know that we live in an uncertain world and the only thing that that you can make truthful statements about the probabilities you can say that this probability is higher and that probability or this probability you cannot say that this is going to happen nothing yet nothing is going to happen hundred ninety percent probability so thinking in terms of probabilities and figuring out work getting familiar with how probabilities work what up what are what was based here in for example this is extremely important when you want to actually have engaging discussions about the future on a more fundamental level than just anecdotal unfortunately I think the most likely thing to prevent is some huge accident by accident that's that's just excuse to most likely scenario because otherwise we're just heading there obviously they might be situations that the of myself situations like might turn out that the brain is actually much more complex than we currently estimated to be and so yeah like like accidents or or the terms of the target is much higher than night time we think these right now you don't really have evidence that that that it is much higher let me think it right that's why we think it and that is on delivery think it is because that's what we have evidence for hey I don't make any statements where where that I'm 100-percent sure of because i know that's a guarantee to be incorrect nothing is on a percent sure i just say that that the dualistic made dualistic worldview given the evidence that we have about about other things people not being able to predict the future and i'm not being able to communicate reliable over our distances just points that there really doesn't seem to be it is very unlikely that there is there is anything than ten laws of physics it might very well be that we are we are a universe that's one in a bigger universe but this since we cannot tap into the roles of that big of universe at least now i have to consider that laws of physics as we know that though them now r ro there is and just keep them out to keep tweaking them as we get new evidence so yeah we might get today with some evidence it would be obvious a phenomenal if we saw that there is something dualistic about minds but so far we haven't got any evidence there are some evidence as the point of a point to our we living in a simulation or not is called non placement universe Fermi paradox is one of them why don't we see anything else yeah why are we why I belong 11 explanation there might be an hour probable that is is that it's just simpler to assume that the background of stars lifeless starts versus engineered engineered background that will be more complex so it's like conserve the resources of that after Gaia was running us in India statement what being and snowed are relative judgments they can only like to a bacteria like an airplane would look like in concealed the income see on the big so it's like the important thing to realize is that there is nothing special about our scale of space so if we say that this is this I universities like really too big to run in a computer pretty we don't comment what what kind of scared that the underlying universe considered normal or on the line intelligence is consumed all wyotech an AI that's the disorder to do things that my time really sharp shooters shoot ourselves in the foot with so yeah I'm nanotech has an extension of my faculty as they might actually converge to India same strain of technologies I'm not an expert on biotech so so I I might be wrong here but it seems that we might get the ability to engineer biological systems and engineering Timon on a level where they actually start forming shapes that we want them to form so therefore like we just by printing out the DNA and basically engineering the dma we would be able to engineer small biological systems that of 10 form bigger systems that with all or good or it will consider deck nano nano computers at even fat cells that biotech seems seems to be a plausible buff to ability to control systems on a very small scale and not a kiss is all about that well there are good reasons for that and are bad reason so that everyone has different people are different reasons for or doing what I do in general and so in particular we see that technology has may improve our lives and it is that this is a next reasonable thing to do and expect in the near future that there will be more technologies coming out of the technological progress that we make our lives even myself I am cautious about this is basic extra playing that trend to infinity saying that the more powerful machines we get more disruptive technology get better so it i think that that technology that train does not take a 1 4ever Soviet to proceed with caution from some point on eternal I'm a consequentialist in the sense that I got I try although always I'm not always successful I try to whenever I whatever I do I try to figure out what are the first what are the futures that I consider good or bad and see like whether my actions are going to steer the world towards better or worse worse future an example would be the action of posting your seatbelt this is there like a probability pump you have two futures one in one future you are dead in one future you are alive and intact as posting your seat belt takes some probability away from the Sun from the bad future was posted to good future you cannot do there's no current PC you can can easily say that yeah why by the way would be why would you post a seatbelt because you don't know if you're going to going to get him to entice accident there's no point in posting seatbelt our or they might say that done that time you should not post a seatbelt because our accidents were verified by actually detrimental to have you have your seatbelt fastened it actually very similar situation with with a chai development there are people yeah we should today h-hi because like if if you don't it don't do they chart and this thing will happen I mean I think it's equivalent to saying that yeah you shouldn't post on the city but because because you might get in an accident where is Mario can get strangled white seat seat belt so it's really I think the correct way of thinking is that that yeah use you think about the directions and see see what kind of probabilities to take from what kind of futures input put 22 what kind of futures what other kinds of features so working to a similarity I think it's like too broad to rod gold I would say that I definitely see a lot of violent or working towards a better future in the sense that that you pick your actions that take probability from dystopian futures and put those into positive futures and then it seems to me that currently some of the biggest leveraged areas to do that probability pumping or biotech and nai so you figure out how to avoid biotech disasters and AI disasters and try to take direction so so they wouldn't happen try to minimize the probability other people would say here that it's actually important to reform their education and likely principal I would agree with that because education is really outdated and we we have systems that you can dream of when the political system public school system was invented like 250 years ago depending how account but the problem with that is that i don't think we have enough time it ok if you change try to change the world for better through education you need generations and we don't have I don't think we have like many generations left so so I shouldn't need more some other ways of promoting rationality and steer clear of biases trying to and basically what I thought to say that what the world needs is a like an engineer's view of the future so it's like all the systems we create to actually figure out what they're going to do before we went to launch them and add this consequentialist view that I just spoke about and these things unfortunately very hard to very hard to promoting existing society one of the things that I'm thinking that might be beneficial is just use the fact that the world is getting more and more connected as a result of internet and make sure that the pockets of rational people they get connected so actually have a global society on top of society that can can actually see ahead instead of just tech represent inventing stories and sticking to them that might be one approach an analytic is going to work with your member for the borders life right foundations technically yes but I don't haven't done much much work tears in my for years or so so it's like white type what organization is it's a interesting idea but they are not really optimizing towards consequences in the world they seem to be optimizing for that just getting prominent people together and giving them email form discuss things and a website but that's pretty much it so I the problem i had there is that that doesn't seem to be held a forecaster not focused only I forgot the guards as this like the video ready call a prediction that this will be a major issue of the century and digital war about this I think he has some good points there but not actually expect that to happen and the reason is that anything people like that in sort of large society large society is centered around the politics are centered around big racial issues and so it's like that that the issue of species dominance as here with the guys is too abstract to bubble up to politics that's that's my that's my thinking for my parameter I don't know so so therefore I can't remain optimistic that like a witch hunt as it predicts but yeah what one thing is certain that we will ask the technology progresses and if you assume we don't get like a disaster in the next decades which is likely but not too likely unfortunately then we will see weirder and weirder stuff we will start seeing like brain machine interfaces sometimes I see perhaps and so if the technology keeps going whether it's going miniaturization and just gradual replacement of also so far I give if you think about it if you look around in a city like humans are quite often the only Tommy only product a natural evolution are a product of a douche evolutional to see I mean in this room okay there's a planter and there are two products here that's it I think so it's like only natural to assume that technological progress is going to penetrate into into human sand and start reconfiguring them so yeah it's a therefore i think that transhumanist view has quit all the validity however it might be short lived in the same step in with diamond bigger changing you now this is very hard for me because like I come from a totalitarian system had the misfortune to grow up in like so at occupied country and so it's like garlic is strong repulsion towards totalitarianism on the other hand there's a lot of discussion about if we as a society should really be careful about what kind of technology we develop a main argument against status that it's not possible to regulate the technology development somebody somewhere with way to do as they please but that assumes that there's no work government is no like there's no entity that can actually take an enforce and given the new technologies that we will get this century you might get like effective mind-reading technology for example with effective my training click under my fully reliable light detectors for example that seems quite possible to get them if you beat those technologies you could actually architect a totalitarian regime that has the ability to control technology apart then i have the sacchetti for demonic what really hate that the return is mean in principle on the other hand if i had the choice between living in totalitarianism in the human totalitarianism or being dead along with my family i would if it's a totalitarianism I've been there and I it wasn't like it was better than media if you have completely reliable light detectors like and you have like a world government that can be principled look like control control the work by basically saving agents everywhere or whatever whatever to do the multi go to is is to basically create hierarchies of light detection system so you have local community the local people overlords who are responsible for monitoring asking the right questions from that from their local people and then you have like mustard overlords that physical same questions from the overlords is educated higher are carrying them their particular comments it come to always be sure that nobody stood up and injects technologies and even though it immediate doesn't come to mind I give you if you could create a graph structure it actually prevents prevents anyone from from taking dangerous I engaging in dangerous activities it actually seems possible you might get that I've needed to address you




Comments
  1. @10:00 restoring the jury-based system actually solves the problem of public education, because it enables tax-resistance, and "opting out entirely." This then, enables voluntary pursuit of better systems.

Leave a Reply

Your email address will not be published. Required fields are marked *