Where Physics Meets Philosophy: Nanotechnology and Neurotechnology | Intel Technology

Where Physics Meets Philosophy: Nanotechnology and Neurotechnology | Intel Technology

Show Video

welcome to what that means with Camille in this series Camille asks top technical experts to explain in plain English commonly used terms in their field here is Camille morehart hi and welcome to today's episode of what that means nanotechnology neurotechnology and molecular manufacturing very fortunate to have with me today Allison duttman she's CEO of the foresight Institute and in a minute I'll let her describe to you what the vision for that Institute is about she brings together researchers and she's also graduated summa laude from London School of economics in philosophy and public policy Allison welcome to the show thanks so much for having me I'm very excited to be here can you describe what foresight Institute does and it was founded quite a while ago in 86 or 87 but what is the goal we're one of the longest ending Science and Tech non-profits is trying to advance Technologies for the long-term benefit of life and so we were founded on a long-term vision of molecular nanotechnology but from the very early ages it always had kind of a cacophony of other Technologies in the mix including AI implications for longevity and biotech brain computer interfaces and in general neurotechnology it has vast implications for space Technologies if we would ever actually get to Advanced molecular nanotechnology and so forth and so because it had this focus on a variety of different Technologies over the years we've taken the molecular nanotechnology angle and then also spun out individual projects that focus on the other Technologies as well and the way that we usually support projects in that space is through a series of virtual seminars in each of these categories and then a series of in-person workshops where we bring top people in the field together to work on the long-term goals of these Technologies for better Futures that they usually don't have another venue to collaborate on because that's not really what normal institutions are set out to do very much and then we also have prizes in Fellowship in these categories I think we're trying to provide this kind of like extra space where people can collaborate on these more long-term visions of Technologies and how they can actually help Humanity flourish benefit of life or long-term goals or helping Humanity flourish they sound really good but I do want to ask what do you mean by that since you're CEO of an organization that's looking to help guide technology impactful technology in a positive way what is positive for you what is beneficial for you that is a very big question I've studied philosophy before and I can safely say that within thousands of years of philosophy people have not figured out what the good is or at least they have not agreed on what that is and so what we're trying to hold space for is technology development that leads to Worlds that can be assessed as better by many individuals and so we're not trying to drive down really one narrative but we're really trying to kind of like give space where lots of different positive Futures can Thrive and so where really individuals have this ability to achieve the greatest possible lives for themselves and that also includes planetary health and well-being because without that we don't really have a space to be in and so without wanting to narrow it down to one let's say ethical framework that we follow we're just trying to get to Futures through technologies that can be regarded as better by a variety of different entities maybe not only humans in the future you may have to kind of like expand our moral Horizon a little bit really quickly what is a conscious creature what would you define as make looking at conscious we all even as a layperson have an understanding of what at least for us it means to be conscious and I think it's pretty precious preserving this very pedestrian notion of Consciousness is I think something that is inspiring enough whether or not we can figure out all the scientific details behind it I think we get to experience it or at least we think we do that's my best bet so let's dive in just a little bit to some of the technology that your organization focuses on let's start with the molecular manufacturing and nanotechnology which are a little bit interchangeable maybe you can just tell us what they mean what those terms mean and then we can have a conversation about them people have been arguing and debating the correct use of the terminal technology because I think nanotechnology as it is with other terms for example as AI the moment that they can't pulling interest into the field many other things get rebranded as their technology as well along the way and that's also what we've seen a little bit along the path molecular non-technology but maybe the best way to describe it is just like imagine perhaps like for a second that it was really truly possible to build things from scratch so that rather than kind of like taking a default resource like stone or wood and then cutting it into shape and thereby wasting much of the material in the process imagine if you could actually build from the bottom up by assembling precisely only the types of materials necessary to build the object of Desire the question is if resources costs or externalities weren't an issue what could you build Richard Feynman gave the speech in 1959 I think at Caltech where he basically said the principles of physics as far as I can see do not speak against the possibility of maneuvering things Adam by Adam so really building with atomic position following the bottom up the poor vision of molecular nanotechnology is really doing autonomously precise Manufacturing so what kinds of things would be built do they span biological and non-biological and then how do they get built what builds them at that level Eric Drexel for example who co-founded foresight in 1986 he wrote this book engines of Creation in which he envisioned a future that was defined by these engines of creation these molecular machines that could really assemble from the bottom up are very complex objects and they also helped by engines of design and that was AI so basically actually being able to detail this path towards achieving Atomic Precision through molecular engineering the early stages we involved redesigning biology's molecular machinery for example protein molecules to really position reactive groups with atomic precision and thereby turning them into these machines that are capable of constructing more complex materials and so you build up from the bottom up almost like 3D printing but on a very very tiny level and for that very long term you could really build a variety of different and either structures or eventually even other factories that could produce new factories I think the promise there is almost Unbound to really reach this really abundant future for example if you just imagine devices and materials that could be designed for example really improve lighting efficiency promote local power generation if you have it and could be like help us meet the energy demands that we have as a planet which are like a major big deal for the long-term future and very sustainably we could have for example new water purification methods with these very very precise materials and could really alleviate much of the water scarcity that we have today we could also for example go into the medical domain and develop new methods new Diagnostics new personalized medicine and new personalized therapy tools that could combat diseases fight aging I'm sure that some people have heard of that kind of term nanobot of very tiny robot that could clean up the inside of your bodies and remove the things that aren't working very well and like repair the bids that need repairing and so you have a lot of individual different promising Focus areas in there if you kind of go all the way sci-fi but I think even we talk about space Technologies it's very difficult to make progress on long-term space development exploration but with new types of materials that could be designed to be for example incredibly lightweight you could also make progress on long-term space travel in a way that we is currently really beyond our imagination so you're envisioning something where in conjunction with AI for example Maybe it exists in an area where like there's water scarcity to use the example you provided and it looks at the surrounding materials it has to work with or that the atoms that are near it and it decides how it could then maybe construct something to create purification in that area given the elements that are around it versus it may create something else to purify water in a different location if you could just create a Machinery that was able to clean up the mess that we made like any kind of molecular Machinery should be able to also clean up the mess that other molecular machineries have made if you just look at biodiversity and resource prices that we have all of this could really be helped a lot by just being able to kind of like produce better materials from the get-go that don't even cost that much of waste and externalities but then also by cleaning up some of the mess that we coming to the clear level of civilization have left behind yeah and some of the medical advancements just to Define it in more detail would be something like chromosome replacement therapy I mean you're alluding to maybe repairing certain organs maybe if there's tissue damage or removing certain toxins but you could even I mean I think in the longevity space aren't we looking at something like using proteins to recreate what the DNA blueprint is telling it to create as opposed to like the sort of damaged molecules that occur over time we're actually doing replacement therapy at the chromosomal level to sort of keep our biological naturally deteriorating bodies repaired and free of deterioration one problem in longevity is that even if we make all the progress that we can make and this field has accelerated drastically it may still not be enough to get people currently alive over this threshold of Lev which is longevity escape velocity and ground X is basically preserving your entire body or just your brain so the hope that later on we will have the types of technologies that allow us to rethought the biggest problem in cryonics is that currently most of the techniques that are being used to preserve human bodies they're incredibly toxic and so even with vitrification methods that we currently have we're still causing massive damage to most of the cellular structures that we have in the human body and so molecular noun technology is usually very much almost a necessary condition if we want to make sure that we can a preserve folks better but then also B be able to potentially go in there eventually later and repair some of the damage that current existing Technologies are leaving and so that is a very very far out case and there's some criticism I guess in the space but help me understand this better that AI in conjunction with nanotechnology or molecular manufacturing can then decide what to construct and autonomously go and construct it and it can happen at such a small scale obviously that we may not know it's there as humans are you referencing the great scenario I'm assuming is that correct yeah yeah so it's basically that you'd have these self-replicating Nano machines that could self-replicate and kind of like could eat up all the resources that we have as a human civilization including humans and including our biosphere and it could happen so fast because they are self-replicating that we may not have time to intervene or even notice it before it's delayed a pretty out there worry based on misunderstanding of the types of molecular non-technology systems that are actually being promoted by the community but instead what most people in the field are arguing for and what is in fact much more pragmatic is these Nano scale factories which where we could really control the input and where we could exactly know what we want to manufacture those are much easier to build arguably than it's like free roaming and entirely autonomous Bots this kind of like worry of a great Google has really been kind of resting on this misunderstanding of what the field of nanotechnology is trying to do and there is even a case to be made that by the time that this view became very prevalent it was potentially being used to redirect funding from the more ambitious goals of molecular noun technology to the more near-term staff then obviously like anything that sounds like a new threat is also incredibly exciting for the media and so that definitely played a role I think as well it's pretty unfortunate we can't rule it out but I don't see it like becoming a near-to-ground Zone anytime soon especially when it gets a name like gray goo yeah of course that definitely helps so can we talk about neurotechnology a little bit sure I think early on I would confuse the terms nanotechnology neurotechnology yeah well again it's a very very broad field and spans all the way from Neuroscience to specific devices that are currently being created what we are focusing on at least at foresight are the types of new technology that we're interested in is looking Beyond what's currently available to the types of neurotechnology that could actually help us cure many of the diseases that we're really worried about but then also allow us to expand our capabilities and potentially even have something like brain computer interfaces that would allow us to make much better I guess interact with the real world that is already very computer mediated but then also potentially eventually do things that are much more kind of like out there such as like whole brain emulations which obviously as a term that many people if they do know it probably they know it from sci-fi so can you describe what whole brain emulation is whole by nomination is basically a potential future of state of a technology where you can really emulate an entire brain in the wake of like concerns around these artificial intelligences that don't share our Hardware don't share our evolutionary environment in which they were able to gradually learn the values that we hold it would be very difficult to align these AIS with human values because the idea is that if it was possible to emulate a human brain in a computer and that is a very very very big goal and then at least we have some reason to believe that it's more easy to align those human brain emulations with our current human values than it would be for an entirely kind of like external AI we would also have better ways to kind of like interface with AIS that we create and actually get smarter ourselves potentially we'll be able to kind of like run on the incredible speeds that many of the Computing programs are currently running and so forth so there has definitely been like a recent uptake in interest in the field nevertheless I should also say that human brain emulations I would say we won't get there anytime very soon so I think without significant funding and significant interdisciplinary collaboration it is still very much like a high hanging fruit let's say within the newer technology space but it would be running on circuits not biological material is that correct well so here you're getting into the kind of like long-term possibilities of having biological Computing there are a few companies and Labs that are working on biological Computing particle labs for example is an interesting one one could imagine that potentially the hardware could even be biological in the future so like yeah it's possible we're talking about with things like neurolink right is another kind of common reference in neurotech space and looking at stimulating the brain stimulating brain function reading the brain and then allowing for communication externally that wasn't possible otherwise what kinds of things are being looked at in the space Caltech is working on one and there's a few others as well at the same time though I think it would be difficult to get the entire readout that you can get from the invasive ones they just have a better reach locally and so there are I think a few really interesting kind of projects going out there you know I'm assuming that in the future may even be like not necessarily a patchwork but I'm just hoping that we can just develop also better technologies that can be more more non-invasive while still being invasive um and I think that there is a lot of progress in the space but I think we still need to work on durability we need to work on reach and we need to really work on readouts I think it would be great if there was more collaboration between individual come companies and what they can do Bernie in terms of readouts and really like trying to have a bit more of an open access to pushing some of the progress along so what about this concept of reading versus writing when you're talking about brain computer interface I want strawberry juice right now or something right as opposed to grape juice or whatever and you're able to actually read that when a person can't speak or communicate any other way that's sort of the inside out you're trying to understand from the outside what's happening on the inside what about the reverse of that what about the outside in where you're sort of implanting information or potentially implanting thoughts it is written about it's not like this so far out there concept right now so help me understand where are we with that and what kinds of things are we able to implant a recent paper used generative AI models to recreate images from human brain activity that matched with what the human subjects were actually kind of like looking at in that scenario there's like a really fantastic paper out there it's rather terrifying if you look at the individual depictions and then they were able to kind of like use generative AI to actually really give a pretty good understanding of what the person was actually looking at and so it was this kind of reverse engineering a fear for that at least is that this technology in the hands of the wrong people could really to like massive infringements on our privacy like I think the place where we really most have privacy in currently stores is our mind on the long run people worry about something like mind crimes and there's been some research at fhi that was looking at these digital minds and like the types of kind of crimes that you can intrude on mines if you actually had an ability to change thoughts to change kind of like the environment in which they occur and to really like kind of like yeah engineer experiences for people is I think pretty massive because at the end kind of like the mind is the filter to which we perceive everything else in the world and if wanting us with that filter I think it's that is probably kind of like one of the worst crimes we used to say data is the new oil but at this point we've got the data we have a lot of data and AI is working hard on the ability to be able to process that data and then generate predictions and insights from that data one of the things we have a lot of data on is biological data that people myself included provide willingly for the benefit of a dashboard back to myself on how well I'm doing and I know that there's data anonymization and other techniques to separate my personal identity from the biological data that I'm submitting but a lot of that data ends up in the cloud and there's really only you know I can count on my two hands sort of the number of hyperscaler clouds that probably contain vast quantities of information on human you know biometric data I don't know that I have a great question to formulate I'm just wondering what your thoughts are on that when at the same time we're approaching things like molecular manufacturing and we're approaching things like neurotechnology I think that part of the reason by foresight has so many different Focus areas including biotechnology American and Technology decentralized Computing space neurotech is because we think that progress will occur at the intersection and I think we see that a lot right now at least in terms of potential in this long memory space that you just mentioned there has been an incredible progress recently in the field of decentralizing decentralized privacy preserving machine learning and there's a few there's wonderful work at openmind and we had a fellow Georgia's cases um who was looking at privacy preserving and machine learning for kind of like medical and Health Data there is such a big opportunity for kind of like us to learn from the type of data that is the most sensitive to us and there's usually financial data or Healthcare data but at the same time it also has crazy privacy restrictions for good reasons because it is extremely sensitive and could also expose our family if not dealt with correctly but we do see in the Computing realm now a lot of really wonderful tools like these privacy preserving machine learning tools that could really be used to leverage on these contact data syncs and so you could imagine for example like encrypting the data at the source and only kind of like letting metadata through that is still significant enough for specific machine learning algorithms to make sense of and to kind of give you back useful information and so without you necessarily having to share their private data out it is definitely people are making progress on this it still cost ineffective as compared to large normal algorithmic approaches but the problem is that usually these more decentralized privacy preserving machine learning approaches don't really have a chance go against the existing Big Data trained AI models but in a space like healthcare where we literally cannot share our data they have a comparative advantage because they're the only ones who can actually properly kind of like make use of this data and so I think it's a really exciting space to be looking at what about the problem where we have data from a vast swath of Humanity on exactly how our biology functions that's now sitting in the hands of essentially the few this sort of giant clouds that can actually hold the literal quantity sheer quantity of data that exists on this front so who has access to that data because I know you work on having the benefits not only rest in the hands of the few so how do we tackle that kind of thing well for many of the type of healthcare data actually we have pretty strict regulations on this and that's also the kind of data that would be very useful for us to actually make progress on longevity because many of the things that people currently are doing if we were able to kind of like cross-compare across what different people's people are already doing really make big progress a lot like I think that yes in general we have lots of our data in the cloud but in specific Healthcare and like fields that are like really sensitive we don't but what about like Fitbit and watch data and you know Garmin Apple all that stuff I mean that's all going to a cloud we can talk about kind of the risks associated with that but I think if we want to talk about new types of opportunities that could be unlocked if we actually were able to share relatively sensitive data or like not share it but actually have it homomorphically encrypted use a friendship privacy Federate learning all of these approaches that are currently being explored I think we could be doing much more than we currently do with our data we are already arguably sharing way too much and I think if we made progress by funding a few of these more privacy preserving Solutions because they're not far out from being actually applicable then I think we could really take everything home again and we wouldn't really have a need anymore to share everything to the cloud the cryptography unfortunately is still a really underfunded field so mostly being practiced in Academia there's a few individual labs and so forth but I think with more investment in these spaces we would have really amazing tools at hand to actually be making progress on this and I should say that this not only holds for longevity but the type of kind of like promise that we see in privacy preserving machine learning Technologies cryptography Technologies security Technologies and auxiliary approaches they would also hold for AI if we worried about centralized AI is becoming too powerful by having access to this big data then I think we should be focusing much more on these decentralizing solutions that come out of secure the security and cryptography space because they have an ability to allow us to make sense of the data and allow individual and Rec organizations to collaborate and actually reap the benefits of the data without actually sharing it around and so we could perhaps compensate a bit for the normally centralizing dynamics that exist in AI oh that's an interesting perspective so disseminate it more so that we have broader access to it and therefore sort of diminish that centralized potential power that could exist it all relies on cryptography Technologies they need a lot of love these things just so people know we did do a podcast on homomorphic encryption and also on data anonymization so we talked with some technical experts who are working on those things that people want to get a little bit more and I think the data anonymization one talks about differential privacy as well I've really enjoyed the conversation thank you so much for having me never miss an episode of what that means with Camille by following us here on YouTube or search for in technology wherever you get your podcasts the views and opinions expressed are those of the guests and author and do not necessarily reflect the official policy or position of Intel Corporation

2023-06-10 13:13

Show Video

Other news