Introduction to D/ACC

Transcript

The Landscape of Technological Optimism

good to see so many people actually interested in DX so I hope you guys end up enjoying this week um so the right so techno optimism so I I think uh last year I'm sure you guys probably remember this was the time when the AI safety and eak Twitter wars were um starting to happen and uh of course uh I mean I can't tell whether that whistle is for Mi safety or E I hope it's forc um but uh basically at the and then at the time of course you know Mark andreason published his Manifesto A lot of other people were uh like saying various things and this made me really

start to think U basically about this question of like what would a philosophy of uh radical technologies that like actually makes a long-term sense to me uh really concretely look like and also like how does some of the work that we're doing I think both in the crypto space and in the in the biospace and Beyond fit into that and like what does that imply about like what kinds of things each of us could be working more on right so the uh I I started the the post with this like three-way diagram right basically uh you know you have the uh anti technology view on one side

which is basically that you know you have dis ahead and safety can come from basically us realizing that industrial civilization is bad and we need to basically go back to living in wholesome cottages and Huts without the benefits of uh modern economies of scale okay sure um and uh then you have the accelerationist view of course which basically middle like says hey let's like turn Humanity into a c kardashev type three hopefully not Kardashian type three that would be different uh civilization as quickly as possible and uh like basically accelerate on everything and especially accelerate on

the AI and so the argument here is like actually the bear is behind us right the bear is uh the 60 million people every year that are dying of Aging the bear is uh all kinds of challenges and risks of the present world and acceleration is actually the the the basically only safe thing and then my view is basically combining this two and accepting that actually behind us is not a wholesome Cottage but it is a bear um and uh at the same time also realizing that like actually we have a choice in uh what kinds of paths forward there are right and some of those paths are great and some of those paths are actually not so great at all right and

uh basically of course I know start off by pointing out that like technology is actually like really really amazing and we have to keep appreciating that right so this is a a chart of life expectancy over the last 100 years and uh like basically you know you can see like all of the worst disasters here so World War I over here World War II over here greatly forward over um over here and uh like you basically see how you know the uh the Long March of uh improving like medical and sanitation technology just over the long time scales just like completely overpowers any of that right and so from that perspective like if you really care about improving uh human

Balancing Technology and Human Flourishing

flourishing and human lives and you really care about the long term and what's going to happen 100 or 200 years from now then the best thing that we could care about is this trend not just continuing but even accelerating right so this is uh one side of the equation the other side of the equation of course is that you know there are ual like bad things that industrial technology causes so global warming is one of those examples though on the other hands there are um environmental problems where one wave of Technology causes them but then the next wave of Technology actually fixes them so this is air pollution uh so air pollution peaked around basically

early 20th century the diagram above here Great Smog of London 1952 um so you know if you guys don't like wearing masks you would definitely not want to go back to 1952 but what happened since then is basically that you know we've uh there's a huge concerted effort to try to move Beyond polluting Technologies and we have non-polluting Technologies and air quality like everywhere is vastly better right I mean I got I remember I got to see the tail end over this in China so 2014 Beijing had to wear a mask always gray in the distance and then you know even by 2019 it had like basically receded to being almost normal right and

I'm sure India is going to go through that exact same transformation over the next decade as well uh so but one important lesson here is that like this part did not happen by Magic right this part happened because lots of people decided is really important and they work hard to make that specific thing happen so getting into diac a bit or a bit right so one of the challenge one of the topics that I think we care a lot about is AI right and people of course have all of their different views about whether AI is going to be soon or not soon whether it'll kill everyone or whether it'll not kill everyone and there's like a lot of disagreement on this and people have their own opinions I personally have a very huge amount of

The Future of AI and Human Agency

just uncertainty on this topic right I think you know the the 10year timeline world is possible the 50-year timeline world is also possible and uh you know there are even longer World worlds as well and I see the arguments for like why any specific version of AI could go very wrong but then of course remember that like the kind of AI that we have now would have looked completely unimaginable to people even 10 years ago not just in terms of its level of power but also in terms of the entire set of approaches that it uses to approach the problem and so we should be ready that 10 years from now the approach is going to be totally different and 10 years from now it's going to be totally

different again and so like we should not put too much Credence into like any one single Galaxy brain logical argument right so this is uh like basically my view on this right and then one other piece one other view that I have on this is basically that like even if AI doesn't actually kill us there's this question of like human agency in a world what does that even look like right and there's a lot of people who basically glamorize worlds where you basically have super intelligent AIS that are making every decision and human beings basically are reduced to just being pets and this is like a world that I think in general like people don't actually want

to have right and uh I think in general like what's the point of developing you know the most amazing technology possible if you're ability to impact the world basically yeah at some point just drops all the way back down to zero right and I think even on top of that there's a good argument that that kind of situation is uh pretty unstable and the reason basically is well if you are a pet if you're a pet then you know like what ultimately those AIS are going to have their own interests they're going to have their own goals they're going to change over time and even if you come up with a way to like keep people in control like that's actually hard to do it's very hard to control something that

is vastly smarter than you right like imagine if like here has a six-year-old child right if you have a six-year-old child imagine if you legally like like were allowed to and you did come to an agreement by which you agreed to be their slave how many minutes do you think it would take for you to convince them to like basically undo um undo those terms and get back to where you were probably not too many right so like basic so I think like human beings like essent staying exactly where they are but at the same time super Intelligence coming to dominate everything like it has risks that even go beyond safety risks and I think you know we want to be cognizant of that as well right also the risk of authoritarianism is uh a pretty

big one right so uh one of the challenges that's been happening in a lot of countries recently is basically that it's in-person protests are just becoming vastly less effective than before uh so uh you know in one of the things that often happens in Russia for example is uh basically you know there's some protests and then the protest happens and then during the protest maybe there's like a little bit of police opposition and it seems fine but then actually like you're being photoed everywhere from every angle and then two weeks later you get a knock on the door at 3:00 a.m. and uh you know then like we know what happens right so it's uh like AI also has like a lot of these

near-term risks that are pretty risky uh so but but at the same time it's like this is all true at the same time as it being true that like all of this stuff is going to massively accelerate science and and cure diseases and do all of those really really amazing things so what is DC right so DC um so you I this was uh drawn with AI that was available last year I guess uh AI That's available this year probably be able to do the same thing but make it a video uh so it's um you know it's and basically know the there's this intentional theme of like solar Punk castles with you know like skyscrapers right and the idea is

D/ACC and Analogies to Switzerland's Defense Model

like yes we can have technology and you know we can have nature and we can be like well defended and safe um all at the same time right and so one of the analogies that I started with actually is you know like a good old Switzerland right and Switzer one is interesting because it's this uh like it for a long time for like basically ever since I became like a very online Internet person it's like the one country that people would refer to at like internet Libertarians would refer to as like a classical liberal Utopia that actually works right and it's like you know how like Peter teal back in a

wa he's like yeah I've decided that freedom and democracy are no longer compatible and then uh you know the what and they're like no yod dog you know we got both right you know we got a strong classical liberal Society we've uh you know very strong individual rights and guess what everything is freaking direct Democratic all the way down to the level where in some cantons if you want to become a citizen like even that gets put up to a vote so what basically the question is like unless the Swiss have like some super Gene that makes them like morally more Enlighten than everyone else which seems unlik according to my current understanding of biology then basically you know what like what might uniquely enable this

environment right and I think one of the answers is basically that's what someone gets defense for free right like it gets such a high level of defense for free that it was even able to survive like freaking World War II with its entire self surrounded by hostile territory and uh well basically you know it it's uh know it it has the mountains um you know um and of course no what's happening and of course the flag is a big plus um yeah so basically right the and then of course what's the opposite of Switzerland it's like the fraking Siberian steps right that just like get overrun by incredibly big armies and

like that's where all of the all of the craziest and all of the most undemocratic stuff in the world happens so basically the question is like well what if we take this as a metaphor and we try to like make a world where the underlying kind of base level rules of the world look as much as possible like this and as little as possible like the opposite of that and I would argue that this is good for safety and at the same time it can be good for preserving a a free and Democratic Society right and so the question is like well what does this actually look like right uh so I uh basically yeah the question is well like what are the rules of the world that we

can change so we cannot change the laws of physics unfortunately I mean maybe we live in a simulation and like maybe if we you know just like put up the right words um on the beach in in the right font in the right language like it's going to unlock a cheat code and then we'll be able to just like uh know like change the cosmological constant maybe we can like set the speed of light to only 100 m/s and then just like see what social changes that'll result but unfortunately we don't have that ability right uh so what do we can't change the laws of physics and at the same time um look the obviously you can change the the laws of politics but there's lots

and lots of people that wants to change the laws of politics in all kinds of directions already um and so what if we think about the question of like what are the benefits that we can get by changing the laws of technology and so here I have uh this uh and so the D in DC it is meant to stand for defense but also standing for Democratic also standing for decentralization right so it's a type of Defense where instead of Defense being done the way that like actual medieval castles did it where the King Controls it and then that becomes like a choke point that the king uses to extract even more even more value from the economy then instead and also unlike the vision of defense that sometimes

Changing the Laws of Technology: Defense, Democratic, Decentralized

gets propagated by large country governments where the vision of defense is basically things are defended because you because you can trust me this code is safe because this code is made by trustworthy entities the hardware is is safe because it's made by people in the right countries why like actually have a model of Defense that's like more inherently Democratic and and inherently distributed So Def I basically have this like four-sided typology right right and so top level is of course the world of atoms and the world of bits right and within the world of atams you I split it into micro and macro right and so macro is like generally when people talk about

defense like okay so there's this whole other separate topic where like when people talk about def well defense like about a third of the time that's actually a euphemism for offense so you know not doing that but then you know the other thing is that like to me this also includes things like resilience physical infrastructure right so this includes like solving the problem for if Global Supply Supply chains break how do we like reduce the the number of people who get hurt even in that scenario by like a factor of 10 right so resilient fiscal infrastructure is one and then world of atoms micro basically means biod defense right and in biod defense there's like actually this huge amount of uh un unclaimed

The World of Atoms: Bio-Defense

opportunities to make uh Society more well- defended and these unclaimed opportunities actually uh like most of these they do not look like what happens during Co right what happens during Co is centralized bi defense right like first you know you had centralized epistemology and you had the World Health Organization like basically publicly say that like you know you should not wear a mask and of course that decision was like originally a result of like trying to overfit to this one specific situation where they wanted healthcare workers to get masks first but then of course the world is a chaotic place and like this is why you're supposed to just like pursue truth stubbornly instead of being Galaxy brains about everything and so instead

uh like basically that ended up creating this whole Meme and then even in 2021 and 2022 you had people just like wiping surfaces with alcohol all the time and like not actually putting proper resources into making sure the air is clean right and uh then what happened second right basically you know we had travel restrictions quarantines lots of things that were I mean realistically like that's like a massive social social experiment on people's uh like on people's mental and psychological Health that like if you put it through like the FDA or an IRB they'd say hell no right but we just did it on the whole population so a lot of that a lot of the stuff that happened there was like

actually high cost things that I propose for biod defense are all not high cost in that way at all most of them are invisible right so very basic one is just air filtering right basically yeah so if you think about this room right so when you breathe out um a you know your your breath contains uh you know like roughly it's like 96% regular oxygen and nitrogen about 4% carbon dioxide and then like there's like tiny fractions of stuff that also includes some of whatever viruses you have right hi virus okay that's okay um it's okay maybe I'm the virus maybe maybe we're all the virus right so right so basically uh it as it turns

out one of the things that we learned is that if uh this event was happening like literally outside right now like imagine this roof disappeared right then like actually virus risk for everyone here would Reduce by like a factor of more than 20 right it's like pretty massive right basically the majority of spread happens through aerosols relatively small particles that just stay in the air for a relatively long time right so what we can do instead is one thing we could do is better ventilation right but then the problem is that in Shang the temperature during the daytime is like around 30 degrees and so if we if we actually like removed this and then like we literally replaced it with glass well if we replac it with glass glass would literally have a greenhouse it would be

worse than 30° right but we don't want that right instead of like showing to overturn the world one simple thing you can do is just air filtering right basically with any room you can put in a couple of boxes and that by itself can just massively reduce the risk another option ultraviolet light ultraviolet light is like much more convenient and uh like we've actually done some experiments of just like putting it up in like very small scale events and it's just like completely unnoticeable to people I there is it is more convenient and

Bio-Defense Continued

arguably it's also a higher risk technology because uh like ultra light does all kinds of different things and so I like right now it's looking like filtering will be the primary and like ultraviolet is is more the backup option uh but basically yeah if you have things like that then you could plausibly reduce Transmission in indoor environments by like you know factor of five or 10 or potentially more then on top of that if you do um all kinds of other things right so early detection of pandemics um if you do much better testing uh so you know 2021 to get a PCR test you have to go to a lab 2024 if you want a PCR quality test you could do it with an at home machine that's like

basically about like maybe twice as big as this mouse and then maybe in a year or two later you'll have like a test that you can just like breathe into right it's like oh hey hey what what's what up officer oh oh were you concerned that I'm drunk oh no no sir I just want to see if you have virus nope no virus carry on oh wait wait but you are drunk no so you know it's uh pretty basically like these are things that become convenient enough that you're able to just like have them at home and just like add a a couple of things that if if we were to do them every day the the inconvenience level would literally be less than toothbrushing but alone like the combination of this stuff alone

would basically transform us from being vulnerable to pandemics like to even existing Co potentially just being a a disease that just does not manage to spread right and so if you want to ask what like a lot of the time like we if you want to take this back to AI safety right one of the questions that gets asked is well okay let's let's like not talk like get into this like Galaxy brain mode of like oh the AI is obviously going to have an IQ that's over 9,000 and so it's to figure out some magic way of killing everyone and we're not going to be able to stop it right it's like well no let like get into the specifics and try to identify like what are actual threats right and when you actually try to identify the threats there's basically two things that tend to come up one of them is bio

AI Safety

and the other is um cyber attacks right and with bio basically and so and well cyber attacks and also persuasion right um so convincing the people to kill each other and that just happens to be like you know these three planks and then the actual killing step also the fourth plank right so with bio um basically the the worst thing that you can do is you basically make a virus and you make a super virus and then the virus like stays in your body does nothing for about like two years and then it spreads to everyone and then suddenly it kills everyone and it's too late to do anything right and so arguably actually shortterm anti-id Research In addition to being useful

Direct L is like actually very useful even against the science fiction threat but why the reason basically is that like what is the one mechanism that we know today about how to make a virus that lasts for a long time and then is unnoticeable for a long time and then after half a decade kills you well it's called the human immuno deficiency virus you know people have probably heard of it and like actually over the last couple of decades we roughly are have gotten vastly better at being able to treat it right and like actually people people make arguments that long Co actually has some like has a lot of similar mechanics and that's why it ends up resisting right so Research into treatment of existing viruses also helps us against like realistic science

fiction threats right so basically yeah like it's not an either or between you know like worry about the science fiction threat versus worry about the threat that's right here in front of us like actually one is uh the things that work against one are also very good things to work work against the other so that's bio right basically there's like an entire spectrum of things that we can do where if they were deployed the level of convenience uh like loss of convenience the uh level of like any kind of centralized power that would be required is like actually quite low but it would make a huge difference and this is like basically an unclaimed gain that like we really could gain with a surprisingly low level of resources so that's bio um then world of bits right

Defense in the the World of Bits

so I think uh basically everything that we do in the crypto space very relevant to the of bits so this like cyber versus info is kind of an artificial distinction that's to me and the way that I Define the distinction is basically that cyber defense is defense against things where all reasonable people agree who the attacker is right so computer hacking is like one basic example right like the code obviously doesn't agree who is the attacker because if the code agreed who's the attacker the code would just not let them in but all reasonable people do right info defense is like is much more challenging because of course one man's misinformation is another man's unjustly suppressed valid point and it's like

inherently much more subjective right so basically cyber defense a lot of this is uh already happening in crypto there's a lot of this happening our traditional web too there's a lot of work that can be done with encryption signatures zero knowledge proofs the hard the whole Hardware side certain types of blockchain applications um secure operating systems there's a lot of points of intersection between keeping your coin safe uh today and keeping Humanity safe from uh potentially like worst case AIS if they come if they do crazy things 10 or 20 years from now right and you know arguably of course the two are actually

very connected because if an AI starts doing crazy things what's the first thing it's going to do it's going to get a bunch of money by stealing it from people's unsecure wallets right uh so and in trfi it'll be able to do the same thing as well uh so that's uh cyber defense and then info defense is like anti-cam tools tools for Collective truth finding tools for identifying misinformation but particularly no this is a d it's not uh you know centralized I'm the king trust me act and so the goal is to try to find mechanisms that do this in a way that do not involve like enshrining one particular centralized Delite that decides for everyone else what's true and what's

false right because we've seen how those kinds of things are like one they're just flawed and two is that they've like in the current climate they're just completely incapable of like actually getting cons getting consensus and being listened to by more than half the population so and as we've seen you know Community notes like people every like people from both sides of the spectrum really you know like follow it and generally tends to appreciate it and the it is that like it freaking goes after everyone right you know it goes after the Reds it goes after the blues it goes after um like authoritarian governments it uh it goes after corporations um oh the the most fun one is like I'm sure some of you guys have seen this right

where sometimes like you get this advertisement and the advertisement says like hey this is an amazing product buy it for $24 at this link and then you have a community note that just says hey this is just a reseller here's a link where you can buy the same thing for $5 like who here is seeing this right so like basically like this is I think one of the best things that El does he like actually allows you know the community notes tocy to just like totally destroy his advertisers so the like mechanisms that actually are like kind of democratic and that are open a community notes say fully open source you can verify the outputs and like inspect them yourself like actually we do stand a chance of getting like past this uh know kind of single tribe barrier and getting

Democracy, Individual Rights, and Future Governance

Buy in from everyone uh so these are like roughly the four planks right and uh I mean obviously I think they all feed into each other in in various ways and there's definitely like some uh uh some points that are kind of somewhere um in between them right but the core idea is basically that if you just ask the question of sort of how do we develop technology so that the underlying at least kind of like middle level laws of that the world operates under are laws that favor defense over offense and that favor individuals and communities over like very powerful centralized actors then like these are

some of the natural answers that comes out of that and there's a lot of value in working on each of those individual things so thank you [Applause] it seems that you guys have absorbed the mic can you raise your hand if you are holding it over there okay um are you up for answering a few questions amazing uh we'll pass the uh can you guys pass the mic forward for me awesome thank you so

Q&A

much uh so I believe that in order to defend ourselves we need to know that what we should be defending for MH and in the current world we're kind of in accelerating phase of developing technology so the complexity of the society kind of grow exponentially but uh for ourselves and for our institutions and for our states um the pace of learning goes kind of sublinear or at most like yeah linear Pace MH so there's going to be a bigger and bigger discrepancy MH so uh if there the world is getting more and more complex and it creates more and more attractors um Runway scenarios and uh

feedback loops that we are not aware of how could we actually defend ourselves um in in this situation yeah so I think uh the thing that's hard and the thing that's probably realistic is like getting the entire world to continue to like continually update their knowledge and habits to be able to like each individual personally defend against everything right the thing that is more realistic is solutions that can be worked on by a like relatively large and like open but still compared to the world relatively small uh group of people and then those Technologies can

be developed and they're developed in a format they can be released they're open and it becomes easier for like institutions and uh like basically tools that and places where people already are that already exist to adopt them right and so like you try to increase the like basically try to develop things where the path to adoption is a path that requires as as a little effort from almost everyone as possible awesome we'll go over to Star Wars here Audi um so I was thinking about this kind of binary You' set up between the world of bits and the world of atoms

and and I was curious if there's maybe room for another adjacency here what I'm thinking about in particular is things which happen that are in like kind of a psychological realm or a sociological realm and even to a certain extent the anthropological realm when we think about the interactions between humans and the problem of like wet wear in the individual and then Collective wet wear and and so I was curious we have to reduce right so we can pass on information that I'm curious about just how much broader does this go or do you are you maintaining the binary on purpose yeah I mean I think like to me I would say anything sociological I would generally consider it Parts part of the bits because these days most of society happens through bits um and uh and I think like the everything in the uh info

defense bucket I would definitely call already in that category one other thing that I talk about a little bit later right is that like I think this post is meant to be not just about DC but also about uh like techno optimism in general right and so one of the things I talk about is uh basically uh the value in like accelerating development of social coordination tools and that has value not just in the context of like of Defending against attackers but also even sort of in peace time right or at least within the context of communities that are already reasonably aligned with each other how do you actually like come to consensus quicker right and I think like that Al also is part of the answer to the question of like how do we

actually accelerate the adoption of these Technologies and uh avoid like both you know like people not adopting them because they're lazy and people not adopting them because they're like concerned about you know like is this some like weird thing that's being pushed on me by like World economic people that want me to eat bugs right and uh like I think you know there there is a techn logical parts to that which is basically that all of these Technologies they're open-source things where you like lots of people can look at and and verify them right but uh at the same time there is actually this uh like sociological part of uh like being able to much more quickly identify like points of uh of common agreement and

actually execute on them one of the reason one of the things that I think we can do both uh you know like as the ethereum community and as the social Community is like really actually help the dog food some of these things and also I think another important social technology to dog food is like the technology of like actually funding the development of all this stuff right so I think there's definitely a lot that we can do for that um I have a question about the sense of urgency of uh especially on these defense offense part cuz if uh if we're not even talking about AI but just AI that is used for you know like drones with facial recognition or or or or

systems surveillance systems or or or or better control mechanisms it might be just too late if these Technologies if if the like d Technologies are not uh adopted and scaled fast enough like do do you agree with that and if if yes like watch yeah I mean I think there's like like two different versions of it it's to late right one of those versions is basically like it's too late we're screwed nothing we can do about it in which case like we should probably just you know retire and just like go and enjoy the time that we have left but maybe for a lot of us actually the best way to enjoy time is to ner nerd out on

really cool problems in which case there's really cool problems over here so you might as well join the other camp and then the second uh possibility is basically that uh it's uh it's too too late in the sense that like only using sort of relatively soft methods of like hey let's like you know like get a bunch of crypto people to donate a couple hundred million here like make a meme coin and like donate a couple hundred million here appeal to a government donate a couple hundred million here and like maybe get a few billion here like that's not going to work and uh you know we need something that's sort of more muscular or controversial right and this uh you know like we saw some of this play out in the AI safety debates right where you have people who literally

think that we're in a 10-year timeline world and so there just is a tail risk that literally at any time nii could just like go totally crazy right and that's why things like know like sb147 was pushed though ended up failing in California and then there's a couple of uh relatively similar bills that got uh are being pushed in other contexts right I think also aside from SB 1047 the other thing that's going to come on the horizon is Hardware regulation right because uh if you have uh AI like basically it does look like AI for the AI paradigms that'll be around for the next while are Hardware dominated in the sense that the version that you can do

with 100x Hardware is like a very significant step up from then from what you could do with one exra hardware and so one other plausible regulatory thing you could do is like basically get you know us China Taiwan like major actors to try to say like actually come together and like agree on a bill that just that says you know like like all at least like large scale industrial training Hardware gets shreked and uh then like you maintain the capability that uh like basically both like both at the same time can order a global shutdown of like at least large scale Industrial Hardware and like lots of things are going to slip through the cracks and like individual gpus that people have for their own purposes are

going to slip through the cracks but if you can if we have the capability of like reducing Hardware by a uh compute Hardware by a factor of 100 for like one or two years then like this is something that a like a lot of AI safety people who are like of the very worried Camp have responded positively too right so these are the kinds of things that you would talk about if uh you were in the war worried camp on AI in particular right and then if we do that then your arguments would be that like we need these more muscular things in order to make the hits a fan timeline a little bit more than 10 years instead of 10 years and uh but then at the same time there's the question of like well ultimately if you're buying time you

have to buy time for something right and like a world where like basically AI keeps getting easier and easier to make and like crazy stuff keep getting easier and easier to make but then like if basically things keep becoming more and more unstable then the like at some point you need to do like something else right and so you need a DC anyway and then of course you know on the whole AI side there's uh of like you know there's definitely a lot of people who oppose like both 1047 and I think what would realistically oppose Hardware regulation and other things as well right and uh the uh like the two places where that comes from is like one is just like strongly believing that this uh like very risky 10year timeline world is just

not the world that we live in and two is uh like basically seeing downsides of that in terms of you like centralized control and potentially in terms of just like slowing developments down when like actually AI devel development could do a lot of good right but then I think if you believe if you're the type of person who believes those things then I think for you it's like you would want to support DX stuff and then uh like basically yeah you know like and the DX stuff is valuable like in part as a hedge right and in part uh to reduce the chance that something crazy is going like something medium scale crazy will happen that will just push the world to start like actually doing the more

muscular stuff whether it makes sense or not right so I think yeah like in the intermediate world like basically yeah like you would need you know like more whether like regulation whether like much on like specific Technologies or potentially uh potentially other kinds of things and uh like that's one of the things that you would actually start start looking into but then you know like you're I think the specific combination of these things just ends up depending a lot on like what which combination of risks you're worried about at different time scales and this is something that's like like lots of people have very madeup opinions about in all kinds of directions awesome um we have two more

questions and then we'll leave it at that so I'm going to pass the mic to the next person um so going back to H Switzerland as a model for democracy um how do we prevent uh democracy from eroding individual rights so they say that democracy is to woles and a lamb deciding what to have for dinner right so how how do we prevent that outcome yeah I mean I think uh you know there are lots of countries where democracy has turned into that and Switzerland seems to be the ones where like empirically relatively less of that has actually happened right so maybe come to what whatz so it next year and this will

be maybe one of the topics of one of the salons and you'll have a great time great uh so next question for you uh two quick things one um on the pets note uh it's been very prevalent in a lot of our discussion in the future here uh my cat leads a pretty banging life and also is influencing my agency pretty heavily right influencing my travel schedule this kind of thing and so when I think about this playing forward uh you know the inverse case of letting the chimpanzees align the humans right is like also suboptimal so in a lot of ways I think about this is like how can we gracefully pass the power off towards

positive uh super intelligent entities down the road so I think uh from a chimpanze perspective the better alternative might be like instead of uh passing things on and becoming what a chimpanzee is today which is a zoo animal you would you could why not become the human and so this is where the whole BCI road map ends up coming in all righty um I think that we're going to have to cic because we do have to be mindful that tension game is Happ happening after this and I'd like to give everybody that's been standing in this hot room um a little bit of a break so um let's give falic a huge round of

Applause thank you so much for coming