 At Protocol Labs, we drive breakthroughs in computing to push humanity forward. Today, we'll talk about vision. We'll talk about our impact. We'll talk about the network. And in this vision section, I want to describe what we mean by push humanity forward. Whenever we say that, what are we talking about concretely? I also want to talk about the what behind that push humanity forward. What are the kinds of projects, the kinds of ways that our network is having an impact in the world? And I want to finish with a discussion about how we approach this. How is it that we're building that kind of impact? What is unique about our system? What is unique about our network? Why we think that we can cause some impact in this set of projects that other groups can't? To set the context, over the last few hundred years, we've seen massive global improvement across most features of the human condition. Most ways that we measure it, humanity is just getting drastically better off. Now, at the same time, we have encountered a series of existential risks. We now have the capability of destroying ourselves and other forms of life on the planet. There's a massive set of extinction events happening. And we are encountering a range of problems that may mean the end of our species. At the same time, this century is a critical moment in our species history and potentially the history of life. We discovered computing, and we've been integrating computing with our work, with our lives, with ourselves, with our species. It's unclear exactly where the future will go, but we know, based on the timelines, that most of these extremely impactful moments, the things that sci-fi has long foreseen, will likely happen sometime in this century. At the same time, despite these massive challenges, despite this incredibly difficult century, we have inadequate macro systems that were designed a few centuries ago and are not well equipped to handle problems of this magnitude. Something like climate change is beyond what those systems were designed to address. It's no wonder that our nations haven't succeeded in defeating this problem, because they're not built with the structures to do it. So this is an extremely critical century and is going to require all of our work, all of our focus, and all of our attention on making sure that that amazing global improvement graph continues, that we improve the quality of life for all of humanity, that we stop and hopefully unwind that massive extinction set, and climate change, and so on, and that we steward a good transition during this century to try and make sure that that ends up with a much better future for all beings. Now, humanity has enormous long-term potential. When you think about our cosmic position, we've been wearing the 4 billion year mark of a long story to get here, and that's only if you count life. If you think about the entire span of our universe, 14 billion years. We spent a very long time together getting to this moment. And what we do today, what we do in this small century, is going to have massive impact for the future of our species and for the future of our galaxy. And who knows, maybe in the future, we'll figure out how to go around the speed of light and maybe it'll be an impact to galaxies beyond. So your life and your time period is extremely special. So use it wisely. For many of us in the Prog Lab's network, what this means is that we must do every year, every decade, every moment what we can to try and aim for really good outcomes across a variety of systems. Some of them are much more mundane. I just talked about galactic level impact. However, things like securing the internet and establishing digital human rights today, this decade, can have that level of impact. These things tend to compound. Think about the importance of the printing press. Think about the importance of writing, of language. Those structures radically altered the future of our species, of life on this planet, of our entire history. So in the last 80 years, computing has transformed our species. Superpowers now sit on our pockets. They sit on our desktops. They sit integrated in a vast array around the world. Decade over decade, there's been this tremendous relentless improvement. Billions of humans and trillions of computers now work deeply integrated. Most human activity in some way is mediated by the internet and these superpowers in our pockets. The internet is constantly granting us new superpowers. And the key piece of that is that it's an open and permissionless network where any human, any group can get together, create, envision, and foresee a superpower, translate that into some code, ship it out to everyone else in the world, and the entire species patches up and upgrades. It's an amazing, amazing invention. And a lot of the benefits that we see today in the last 20 years are thanks to this open permissionless network. Without this kind of openness, it probably wouldn't have worked this way. Now, at the same time, the infrastructure that we're building, this computing system, is giving a lot of power to a few groups. We've talked a lot about corporate surveillance and so on in the last 10 years. But what I'm particularly worried about is state surveillance. There's already mass surveillance going on at the same time. However, we have not yet seen what's possible. Orwell famously described this tremendously dark world in 1984. But I think that was child's play. The kind of 1984 vision of the future does not at all factor in the kind of control systems that you can put in place today. Not in the future, not in 10, 20 years today, using today's technology. Already, we're seeing these kinds of systems being played around with across the world. This is not just in a far off place. It's probably in your country. And if they work really well and they'll start spreading and exporting to other nations. And this is a very, very pernicious system that over time will ratchet up into something potentially really bad. And the concern is one of these. It's at every point in time, the people that are building these things are trying to do a really good thing. They're trying to help secure systems and they're trying to help secure society. Because they know the damage and the dangers that come when safety's not carefully considered. However, never about the current administration, it's always about who's going to inherit in the future that set of powerful capabilities. Don't build extremely powerful technology that you wouldn't want the worst groups on the planet taking control over. So it's extremely critical that we build the structures to prevent this kind of mass social control in the future. Now this is where Web3 comes in. Web3 brings verifiability to interactions across the internet. It makes sure that we embed cryptographic requirements in most interactions across the network. And I'm not talking about kind of like the ownership thing, like that's been, I think that's a kind of passing moment thing about Web3 because property rights are pretty interesting and pretty cool to have on the network. However, what's much more important, much more deep, is to bring verifiability to make these systems trustable for everybody because you have hard mathematical and cryptographic and economic guarantees about their operation. You don't have to trust that some contract is going to be respected. You need a communications infrastructure that respects you a communications infrastructure that puts the data in your control and that isn't going to just take everything you've ever said and later on in 10, 20 years in the future punish you for it because of some like massive shift in what has become acceptable. That's what we're talking about. If you think this is like super crazy, just go back to read about what happened in 1930 and 1940 and 1950 and 1960. That entire period has many concrete examples of what happens when different administrations take control of people's social communications. A number of us wrote up these Web3 values a while back and we think that they're potentially extremely important. We're not sure whether these will actually yield a really good digital society, but it's a good start. And a lot of these values can be enshrined as digital rights by building secure systems. If we can create secure systems that enable these properties and then later get large groups across the world to ratify these as actual rights, we can ratchet up the security of the network. We can build and improve our communication systems to get to a much better spot. Now a lot of our projects across the network work on a variety of these values and they're trying to establish that kind of operation and that kind of system. Actually, in reality, hundreds of logos would probably appear on this page. So lots of the faces around this room are working on specific systems in these areas. Now it's going to take all of us. It's going to take all of our projects, all of our systems integrating well and delivering an extremely good user experience. Computing doesn't upgrade because something is more important to do. Computing tends to upgrade when it's more convenient, easier, funner, and so on. Remember that early on, the internet was for cat videos, right? So if you want the world to ratchet up into a better system, you have to make it drastically easier to use. You have to make it more convenient. Users will not select privacy. Users will not select security. Users will select convenience and fun. Oftentimes, people dismiss the requirements, but it's important that you think about your superpowers as writing on top of a massive scale international computing infrastructure that's run by a set of corporations. So the application stack and the personal computing devices, those end up being manufactured and controlled by various different groups. There's probably a problem there in terms of the number of manufacturers for computing devices. We probably need a lot more groups to do that. However, it's extremely difficult. Now, the layers below are actually pretty good. When you think about the openness of the internet and the openness of the networks, that's working quite well. The problem sits below. The way that the internet is steered today is through a series of contracts, terms of services, and the action of various corporations and courts around the world. Those contracts can change very easily. Those courts can decide things in the future that are different from the past. So what we need to really upgrade our systems is to create a layer that won't change, that has hard requirements for change, that has a much stronger foundation. We want math, cryptography, and economics to be to underlie the structure of the internet, to underlie the security of our communications network, and to guarantee your future access to these communication systems. I think most of you have not been in a country that suddenly overnight disabled the internet, but more and more people every day are encountering that kind of experience, and it could happen anywhere. So these kinds of infrastructure improvements can lead to public, international, digital utilities. Don't think of them as corporations. Think of them as a utility that is there for the benefit of all, for the benefit of all of humanity, not run by any one group, but run by the math, by the cryptography, and by the economics behind them. Through that, we can establish these digital human rights. And from there, we should get them ratified. But I think we're in a moment where we should build them first and ratify them second. So that's why a lot of us are working on Web3. Now, there's another very interesting path in Web3 which is around upgrading our economies and governance systems. I won't touch on this for very long, but I'll give you a glimpse. So along those inadequate government systems, we also have inadequate macroeconomics. Our current systems and structures are prioritizing things that are very short-term oriented and not long-term oriented. Most of the valuations of systems tend to be in a very short time span and they're undervaluing the importance of the future. Today, things like Tesla and SpaceX were completely irrational to start back in the early 2000s. Those are now some of the most valuable companies on the planet and they're creating an extreme transformation in humanity's future. Think of the electric car transition is in great way due to Tesla and the fact that we're actually quite close to getting to being multi-planetary is thanks to SpaceX. But those two companies seem completely irrational in the beginning and that's because they're extremely risky. They're extremely hard. They're not valuable and when you think about their economic success, it's very clear that they were going to be super profitable in the future. So what's going on? The problem is in how our macro systems are valuing these kinds of risks and in how they're valuing those kinds of payoffs. So the problem is in the core. There's another story like that. Early in the story of nuclear, there was this kind of set of assessments around fusion and how much it was going to cost and all of those projections were never delivered. So the reason we don't have fusion today is because governments couldn't pony up the massive capital expenses that it required. Now in reality, that expense was smaller than many other things governments were spending on. So it was extremely short-sighted. It was again one of these problems that after the Cold War kind of waned to some extent, people were no longer investing in this kind of far-reaching thinking. Now when you think about all of the budget for fusion and you compare it to modern day digital home entertainment, we as consumers are paying dramatically more for home movies than fusion. And so we're being extremely short-sighted. But what's happening is that we lack the structures to be able to invest in our future. We don't have a way to ask a species, as humans, as individuals and groups, be able to invest in that long-term future. That's a problem in our economics and we need to fix it. The good news is we have crypto economics today. You don't have to go and create new types of structures and so on in an extremely difficult way. You can build them out of smart contracts, deploy them to these networks and scale them. And that's extremely, extremely promising. It means that you can build totally new economic primitives, totally new ways of organizing large groups and organizing industries out of smart contracts. That's an extremely interesting way of upgrading these kinds of problems. So I'm a strong believer in the potential here. I think that this is one of those things that is early today. I think my sense around all of the economic primitives that are being explored is that it is as impactful as crypto itself, meaning when Web3 was first starting, when the first blockchains were appearing, it feels a lot like the new primitives that are being talked about today. So I'm really hopeful that in the next five, 10 years, we're going to be able to experiment with a lot of new structures and scale. And we design a lot of the structures around protocol labs itself with this in mind. A couple of small plugs. One is there's a very interesting set of financial primitives for doing climate change corrections. Being able to decarbonize the planet will likely come from the legibility that Web3 gives us. It gives us the ability to create many more and more interesting and verifiable financial instruments for causing improvement. And there's also groups experimenting with full retroactive impact markets where you're able to create impact markets for actually having significant positive impact on certain goals. So all of this is super interesting. Now I'll briefly mention something that is becoming more and more of a topic across our network, and that more and more is capturing a lot of the attention of many individuals in our community. We need to get to a spot where we can develop the new interfaces of computing much more safely. Things like augmented reality and virtual reality are going to come sometime soon. And billions of humans are going to be spending a huge fraction of their life in those worlds. So it's important that those are open and permissionless networks, that they're not locked down and siloed the way that the current social networks are. It's also important that we transition to things with great computer interfaces with an eye towards the best outcomes in the future. Robotics is around the corner as well. For many decades, the hardware was well ahead of the software. And nowadays, the software is improving so fast that most of those problems will likely be solved in this decade or the one after. And so that means that we might start seeing larger and larger scale of robotics being deployed. This is one of those cases where people mis-underestimated how difficult software was. People talked about robotics in the 50s and 60s. The hardware and software were both bad. People focused on improving the hardware. The software was way too hard. And it took us up until now to get those improvements. But we'll likely see more and more and more of this. A kind of nightmare fuel type of thing is go watch the really fun and happy videos about all the robots dancing and so on. It's both really promising because those things could be extremely useful to save lives around the world. But any kind of dual use technology worry about what misuses might happen in the future and build those technologies with safety mechanisms to prevent those kinds of problems. Now the big ones, the timelines are extremely unclear. And what some of us do is to try and update based on latest evidence and so on. In the past, I used to have the two diagrams flipped here. I used to have BCI coming sooner than AGI. That has changed in the last five years. The improvements in AI are extremely significant. And most experts tracking the requirements for building something like an AGI have now updated their assessments. The kind of scary thing there is most groups think about something like AGI being possible between five and 15 years. That is an extremely short timeline. We haven't yet figured out AI alignment. So that's a problem for us to figure out. These later things like the economic and governance systems and these other interfaces are a small fraction of what our community works in today. But that's growing and growing over time. I've seen many groups starting to shift their attention. But it's very important that we continue the mission on what I think is critical as soon as we can, which is to secure the internet and establish digital human rights. Locking that in place first is going to mean a tremendous set of improvements for the world. Now that said, if you do pay attention to these other interfaces, getting a significant good result in the AI alignment problem might potentially be one of the most impactful things that you or anybody else in the world could do right now. So definitely want to strongly encourage people to check that problem out specifically. The other interfaces will take longer. But I think definitely AI alignment is highly critical. Now, I want to talk about how we drive breakthroughs in computing in our network. At the end of the day, the way that innovation happens is by mixing science and technology. Science is the way that we expand what we know. Think of it as a process by which we as a species expand the knowledge that we have and we increase what we know individually and as a group. Now, the scientific method is often seen as the hallmark of science. And that is, of course, how an individual and a small group gets to discover a new scientific concept. But I think the larger scale impactful thing is the structure around the scientific method that enables that structure to scale to the entire species. Being able to have a process that integrates our thinking around the world and gets to drastically better improvements along the way is what has given us that 2 to 300 year improvement rate. Most of our success comes from important scientific discoveries that were then translated into technology. So the other side of the coin is capability expansion. Technology expands what we can do, as individuals, as groups, as a species. And technology development is long and expensive and it involves what is described as science to technology translation. But in reality, I think these two are two sides of the same thing. And they should be thought of as an integrated system. In the last few years, I've been talking about this innovation custom that sits between these, where you have two incentive fields, one from reputation in the academic world and one from technology development through corporations and broader markets. And they don't quite intersect. Maybe when interest rates are low, they tend to touch. But certainly, in many cases, in many fields, it is just not profitable to work on translating improvements. There's tons of important technologies or things that have been conceived that sit in scientific papers. We have an abundance of great results. But the bottleneck is in between science and technology building. That's the key part. And it's scaling that that is going to let us cure more diseases faster and so on. Now Bell Labs is a huge inspiration to a lot of us in our network. It was an amazing hive of invention. It was responsible for so many of the technologies that we now take for granted. And they understood the translation process extremely well. The way they modeled it was as a full R&D pipeline, starting with research with small groups. But where the bulk of the work happened was afterwards, in what they called fundamental development, productionization, production, application development, and then later on operating the business. Although most of the academic credit happened on the important discoveries like the laser and the transistor and the solar cells and so on, and usually most of the attention went to the few people that had that conceptual breakthrough, the vast majority of Bell Labs work sat downstream of that moment. They understood that innovation actually happened later. They described innovation as actually getting a conceptual result into being developed into proof of concept, into devices, being put into a product that was delivered to the market. And they said that if you hadn't sold something at massive scales, you had not yet innovated. So I think it's really important that if you work on these far-reaching projects, these important scientific breakthroughs, these technological developments, you have in mind this whole pipeline. If you don't get it all the way out into the market, it doesn't matter. If you don't, or maybe it matters, but it doesn't matter as much. If you don't make sure that people can actually use and benefit from what you're building, it won't have the impact that you foresee. Most of the staff at Bell Labs focused on the broader downstream impact. And it was only a small fraction that worked on the early stuff. Now at PL, we think about this as the broader R&D pipeline. And we think of these five stages, research, development, productionizing, production, and scaling. Research is where we think about new ideas, new targets, new possibilities. Development is where we refine those into particular projects. We test out many things. We build proofs of concept. Once those proofs of concept can be ready to be put into a system, we tend to productionize those. We tend to reduce the cost of operating them, the cost of building them. And we start fitting them to applications, fitting them to products. Once we're able to turn that into a project that people can use, we get to making a thing that finally people can download and use or use over the internet or something. And that's kind of production. Once you have released it to the world, you start an adoption loop and you get to scaling. Now when you sort of overimpose the innovation chasm, think of the realm of startups as kind of in the last part of productionizing and getting to production and scaling. When you think about an early stage startup, they're kind of in the productionizing phase. You will hear many VCs talk about how they don't fund research projects. And what they mean is they don't fund research or like that early fundamental development. The word development is tricky here because it could apply to everything, right? So we're really talking about this kind of in-between spot where it's very difficult to go from conceptual breakthrough to something that's gonna fit into a product. That's where there's very few projects and very few teams working and where there's the lowest incentives to work. One other important piece here is that it is not a linear thing where one project turns into scale. In reality, one single idea will yield many possible paths, many different projects, many approaches, and of those, only one of them often will end up working out. From there, you try to productionize in a variety of ways and only one of those will tend to work or a few. And then once you have many attempts, there will usually be something like five to 20 startups working around the same idea, and eventually one or two will win. So this is a pretty large pipeline that involves tens to hundreds of teams working on the same idea across a 10 to 15 year time span. So when you think about large scale R&D, what the world will benefit from is a large open network that enables as many groups as possible to give it a shot. What we benefit from is going from ideas to production as fast as we can. And the current IP structures, the current ways in which groups and companies tend to operate, they tend to try and hold back other groups from trying the same thing. And that's a mistake. What we want as a group is to, we all benefit, game theoretically, from speeding up the pipeline. So we should be thinking about new structures and new ways of accelerating this. New forms of IP that are more open source and embeddable, new funding structures that can come in and into the pipeline here. So the last thing is how do we accelerate this R&D with an innovation network? I tend to describe PL as a large scale innovation network that relies on open source. And though open source has been with us for a long time, we've always been kind of struggling between this kind of like alphabet model, which is kind of like one company directing things, versus the YC model, which is a more open network built on investing. And of course, the crypto network model where you have a fully open structure. A couple of years ago, I put this slide together comparing these and these are back of the envelope figure, so most of them were probably wrong now and shifted. But what's extremely salient is that to me, was that YC has had a dramatically larger impact on research and development, on extremely hard tech and bio and so on, that alphabet and X and so on, hasn't. YC started as a SaaS accelerator and eventually opened up to support and help grow tons of other organizations and companies. But the fundamental difference was the approach in the structure. The way in which it helped and supported groups made the whole difference. Enabling many teams to take on their projects and being okay, funding, groups doing similar things over years, being okay, supporting everyone equally, being able to be supportive of all groups, ended up with a dramatically higher impact over 10 years than Afflevated. So this is extremely promising and has been super inspiring to us in how we think about broader R&D. So that's the R&D pipeline. Now one final thought about VC. The VC model is extremely profitable. And the key thing here is that it's able to invest in early stage startups that have 100 to 1,000 X returns as those startups during the companies. Now everyone famously talks about how this is portfolio approach is extremely risky, right? Like you have to invest in hundreds or thousands of companies for a few of them to really have those big results. If you can do this well, you can scale this and make it an extremely profitable structure that you can then bring in to fund the early stage R&D. Now, as I talked about before, the most successful and valuable long-term companies actually start way earlier than most VCs invest. So the earlier stage stuff, things like SpaceX and Tesla, are much more valuable long-term. Things like say Bitcoin or Ethereum early on were like that. But those systems are way riskier and much more complex. So you have to change the approach. The current VCs model doesn't actually work for that early R&D. It's way too risky and way too long-term. The VC model thinks about 10 years. This is really like a 20 or 30 year time horizon. So the way that we're thinking about PL in the future is to be able to take this innovative loop around VC with startups and be able to flow the cash model into those earlier stage projects to be able to reduce the risk of those areas and make those kinds of things highly profitable. Now, of course, we want to do this in a highly open source environment where as many groups as possible can help each other. And one of the super exciting things here is that with crypto, we might be able to think of new innovative IP structures to help get everybody to open source, even when they're working on super early stage R&D stuff that is extremely expensive. Concluding PL is this innovation network with a full R&D pipeline with a world-class community. We're very open source and we have this crypto and VC business model that enables us to create, support and grow projects so that we can drive breakthroughs in computing and push humanity forward. Thank you. Thank you.