Original source: Bankless Podcast
Original translation: 0x711, kaori, BlockBeats
On July 24, Worldcoin released an open letter signed by Sam Altman, announcing the official launch of WLD. The letter states that Worldcoin was founded more than three years ago to create a new identity and financial network that is owned by everyone. Launch of Worldcoin begins today and, if successful, will significantly increase economic opportunity, provide reliable AI-human differentiation solutions while protecting privacy, advance global democratic processes, and ultimately potentially demonstrate a path towards an AI-funded global basic income (AI-funded UBI) potential path.
On the same day, the well-known crypto podcast Bankeless also released an interview video with Worldcoin co-founders Alex Blania and Sam Altman on the Youtube platform on May 14. The interview covered issues such as the vision of the Worldcoin project, the security and privacy of the Orb device, and the future development plan of WorldCoin. BlockBeats translated the interview content as soon as possible and compiled it as follows.
Origin and initial conception of Worldcoin project
Bankless: Worldcoin is one of the most ambitious projects in the crypto space. It has a token, a dedicated Layer 2 network, an identity protocol, and a mobile app. Of course, the most memorable thing is the iconic Worldcoin hardware - the silver Orb. The idea of Worldcoin is that the only and best way to verify that a person is a person is through human biometrics associated with our unique DNA. Sam, can you talk about the origin of the idea of Worldcoin? Where did it come from and why do we need Worldcoin?
Sam Altman: I was originally thinking, if you could have the largest network, like the largest financial and identity network imaginable, then that would be very powerful, truly global. When we first started talking about this idea, I thought about how do we scan the palm of every person in the world, and are there other sophisticated ways to verify identity?
The reason I'm excited about this is that as the world moves toward powerful AI systems, I think it would be really important if we could do something that ultimately allows for the redistribution of wealth - maybe even access to these systems, which will someday be the most important component of wealth - but also to protect privacy with a different perspective, which is going to become increasingly important as AI advances.
So I started thinking about this a number of years ago, mostly driven by my work at OpenAI. I also thought UBI (Universal Basic Income) was a cool thing to work on. I wanted to explore this space, even though my ideas were far from complete. Then I met Max and Alex, who I think is really great, and I'll let Alex talk about how this project got to where it is today, because he's been a big part of it.
Bankless: So Worldcoin was originally a two-birds-with-one-stone project, right? UBI and identity, both of which you had in mind before there was a practical solution?
Sam Altman: Yes, if you look at it from a higher level, it's actually solving the problem of verifying that "people are people". This is the basis of solving the problem, and if there is a non-fraudulent solution that can do this while also protecting privacy, it would be perfect.
Bankless: Let’s talk about how Sam and Alex met, when the idea for Worldcoin might have come about, and how Alex got involved in the project?
Alex Blania: We really started working on it in January 2020. I think Sam and Max had been working on it for 6 months before that, but Max had a full-time job and obviously Sam did too. So when I joined, we formed the founding team and then we really started working on it. Before Worldcoin, I was mostly working on theoretical physics and how to use deep learning to predict AI systems, so it wasn't very crypto related. I actually read about Bitcoin very early, so I was kind of early in the crypto space, but I can't say I knew a lot about crypto before Worldcoin. Then Max sent me an email with the Worldcoin whitepaper at the time and the initial idea, very high-level idea, and explained why it was important and where it could go. I drove to San Francisco and had a few conversations with Max, and then an interview with Sam, and then we spent more time together, and then I became a co-founder of the project.
Bankless: Was Worldcoin a crypto project from the beginning?
Alex Blania: Yeah, I think so.
Bankless: Sam, originally you were conceiving of scanning palm prints, but Worldcoin didn't do that. It chose the iris. Can you talk about that choice? And then we can start to dig into the technology inside the sphere.
Sam Altman: Like Alex just said, the idea for the project was centered around a paper. One of the main issues is that you basically need to solve civil resistance in order to distribute tokens to everyone. Why would you want to do that? Because it will incentivize a really large network to grow to all of humanity and really bootstrap this huge financial network. Of course, you can also distribute UBI in this way. So when you want to do this, you want to launch a token to benefit all of humanity, then the big problem you have to solve first is what is often called civil resistance in crypto circles. I'm sure you're very familiar with this. In simple terms, it means that a person is able to prove to the network that the person is actually unique in that network. Why is this important? Because if you don't, what happens? The bad guy just subverts the entire token distribution mechanism and everything falls apart. As you know, David, this still happens frequently today in crypto. If you really want to scale this to all of humanity, it's a bigger problem.
Obviously, this was one of the biggest problems we had to solve first. And we did some pretty deep research. The entire founding team mostly came from Caltech or Max Planck Institute, and we all had a physics background. We looked across the industry and came up with three possible solutions:
One is KYC, which uses the government's citizen identity verification system. But this option is immediately ruled out because less than half of the world's population has a digitally verifiable ID. Considering inclusiveness, this solution is not scalable. It may work very well in Europe and the United States, but it will not work globally.
The second big direction is the so-called Web of Trust. Trying to build a web of trust-based relationships. This is a very idealistic idea, but it has never really been implemented or expanded. But I think this is something that Worldcoin can introduce in the future, and we can discuss it later.
The last option is biometrics. So we actually implemented demos of all three options. After repeated consideration, we decided to go with biometrics. Before we talk about the specific recognition mode, you first need to understand that all the things you use in your daily life, such as iPhone, all it does is re-authenticate you. It will store your facial features, and then when you try to log in to the same phone again, it will calculate roughly the same feature vector, and after the comparison is passed, you can use the phone. This is just a one-to-one comparison, which is relatively easy to implement. But to solve the problem of provable human uniqueness, you need to compare one user with all other users. This requires collecting more information about each user to achieve this, otherwise the fraud rate will increase exponentially and actually hit a bottleneck. This is very important. The fraud rate is not a fixed value, but you will suddenly hit a ceiling. Like facial images, there is not enough information entropy. Fingerprints do not have enough information entropy. In theory, palm prints can, but there is no ready-made commercial solution. Iris is the key.
Bankless: You mean that information such as facial images is not unique enough to be used for reliable verification of natural persons, right?
Sam Altman: Yeah, at least if you're just using a regular cell phone camera.
Bankless: OK, so when you were looking for ways to verify a physical person, you concluded that government IDs wouldn't work because that doesn't fit the idea of cryptocurrency. You also talked about the network trust model, which exists in the cryptocurrency world but is still relatively experimental. So biometric solutions are the most reliable and mature model for establishing a unique human identity. That's why you chose iris scanning. I guess the information entropy of the iris is very high, on this order of magnitude, right?
Alex Blania: Yes, you are absolutely right.
Orb's technical principles
Bankless: So this explains why Worldcoin exists the way it does, and why a hardware device like the Orb is needed. If it only relied on national IDs like the United States, or just used a network trust model, Worldcoin would only work in the crypto space. But since we need to collect actual biometric information and match it to a person's DNA to determine uniqueness, we actually have to build a real-world solution, which is why the Orb is needed. Can you briefly introduce the technical principles in the Orb and how it does its job?
Alex Blania: The way this project is changing the world, as you just summarized, is really important. Digital content (images or videos) and smart algorithms can no longer distinguish what is human. So you actually need to connect to the real world, and I think that's the only way to solve this problem.
We started building Orb very early on. We did a lot of engineering work, even designing a custom camera. Orb has a number of sensors that detect if we are looking at a real person and not a display or an AI trying to trick us. All of this computation is done locally. So the first step is to confirm that the subject is a human. Second, the eye is photographed, a unique encoding of the iris is calculated by a neural network, and then signed by the Orb device, which is the only information that leaves the device. What's very important and very cool is that the uniqueness check is separate from the user's wallet and is implemented through a zero-knowledge proof. That is, we or anyone else can only prove whether a user has been verified before, and no other information.
I think Worldcoin offers an alternative that makes it private, open source, and decentralized. So, I know it sounds counter-intuitive at first because it involves biometrics and so on, but if you really understand the engineering behind it, I think there are actually no privacy issues that people are worried about.
Bankless: Absolutely. I understand that iris scanning is an established technology. From what I understand, most of the technology in the Orb is actually to prevent humans from cheating the Orb. It scans your iris, that's what it does. But most of the technology in the Orb is to make sure that the human in the process of getting the iris scan doesn't cheat. Can you walk us through the various layers of technology that go into the Orb?
Alex Blania: Orb actually solves two major problems, which is, as you said, spoofing. That can be broken down into two things: one is showing the device something that's not from a real human, and two is physically attacking the device itself. So you can go directly into the processor and try to insert data streams that are not from imaging. That's the first thing. The second thing is imaging resolution. All the biometric devices that we've tested, their resolution is not high enough to scale to a full human. They don't have high enough resolution, which is why we had to make our own lenses and imaging systems and so on. That took a lot of work.
So I think a similar amount of engineering went into both of those. But to your question, there's a multispectral sensor on the front of the Orb that captures images at multiple wavelengths, like infrared, 3D time-of-flight, etc. to make sure that we're seeing a human. You can show a display or more complex optics to fool it. It takes quite a bit of engineering to make it happen. And then, the cool thing is that all of the processing is done locally on the device, so there are several neural networks doing all of these checks in real time. That's important for privacy.
How WorldCoin solves privacy issues
Bankless: One of the concerns about the WorldCoin project is privacy. Since you can prove that a person is human by scanning their iris, how does WorldCoin solve the privacy issue?
Alex Blania: First of all, we agree that privacy issues are serious. This is why we were very careful when designing WorldCoin. Indeed, biometric technology is disturbing at first glance. But if you really understand the engineering principles behind it, you will find that these concerns are actually unfounded. WorldCoin mainly addresses privacy issues from three aspects:
First, biometric data is not stored. Image scanning is done locally on the device, calculations are also done locally inside the device, and only the iris code signed by Orb leaves the device. These codes are then used to match all users.
Second, and most importantly, through zero-knowledge proof, identity uniqueness verification is separated from the user's wallet, which provides great privacy protection.
Third, the entire setup is self-hosted. When a user registers, they will have a non-hosted wallet and subsequently generate a zero-knowledge proof to prove that they are a member of the user group.
I don't think there is any other solution that solves the same problem at this scale with almost zero privacy risk. In addition, all the code will be open source. Most of the hardware is already open source, and the firmware code is slowly being open sourced. The whole setup will also be completely decentralized, which is of course difficult, but we will do it. We have invested a lot of time and resources to build WorldCoin in this way.
Bankless: Understanding the system at a high level, it seems like there are a lot of moving parts. There's the hardware inside the sphere, there's the zero-knowledge cryptography, and there's the phone app. Can you detail how a person authenticates through an iris scan, which is then converted into data, and then through a series of steps that turns it into a verified login that someone can do in their WorldCoin app? Can you talk about the transmission of data, the processing of data in the iris scanner, and the zero-knowledge proofs? Can you walk through an end-to-end process from an iris scan to someone's phone?
Alex Blania: OK, let me try to explain how this system works. A user is interested in WorldCoin, learns about it, and downloads the World App (currently the only client for WorldCoin, but other wallets will support it in the future). The user looks for a nearby Orb device on a map, and then shows up in front of the device in person. When verifying, the user clicks "Verify Now" and two key pairs are generated, one for the Ethereum wallet key and the second for the identity wallet key. Both key pairs are stored on the user's phone.
When verifying, the user shows the public key to the device. Orb then first performs a natural person check, followed by a uniqueness check, which takes an eye image, calculates a unique signature using a neural network, and signs it with a private key. The signed signature is the only data that leaves the device and is sent to a backend server for verification (which will also be decentralized).
Later, when the user uses World ID, a zero-knowledge proof is used to prove that the public key has been included in the user set. This allows users to use this proof in different ecosystems, not only in Ethereum, but also bridged to other chains, or used when logging in on a web page without revealing any personal information.
In addition, based on this "natural person proof" concept, WorldID is also an identity protocol that allows developers to attach other verifiable credentials, etc. So its function is not just human verification. In short, this is the entire process from eye scanning to mobile phone end.
Bankless: Worldcoin does not ask for your name, date of birth or address during the verification process?
Alex Blania: No, the only time this would be an issue is when you are using a deposit and withdrawal solution, because within the app, also within WorldApp, there is a non-custodial wallet where you can withdraw funds through a deposit and withdrawal provider, and they will ask for your name, but that is not us. In short, within the WorldCoin system, your name, date of birth, address, or anything like that is not required.
The intersection of Worldcoin and AI
Bankless: So Worldcoin provides public addresses for humans, right? And the way this address is derived is verifiably human. Sam, you are one of the co-founders of WorldCoin, and from the perspective of OpenAI, you would find that having a list of verifiably human addresses would be very useful in the age of AI. So why do we need this kind of address in the next decade or century?
Sam Altman: I think there will be other reasons in the future in addition to the reasons I mentioned now. I think it's very important that in an era where AI is being developed as a tool for humans, I think there will be a lot of benefits from them, access rights, governance rights, etc., which will belong to real humans. We like many of the advantages of the WorldCoin solution. As Alex just explained, although the privacy perspective is different from what people have previously imagined, I think it's a very attractive perspective. I also think it's a very fair system. It's probably the most inclusive system we can imagine.
Bankless: Alex, have you seen other use cases in the crypto space? Airdrops are another example, but since we don’t have proof of a natural person, the distribution of tokens is watered down.
Alex Blania: I'm going to start with crypto first, and then we can broaden it a little bit. I've been to a lot of different countries and talked to a lot of users in Latin America, Europe, and Africa. I think there are a lot of financial primitives that the crypto space could technically offer, but they're not being implemented right now for a couple of different reasons. One reason is user experience, but a bigger reason is that you need reputation to enable these features. The concept of credit scores, which is basically an on-chain credit score so that you can take out low-collateralized loans, I think is a big concept that's hopefully coming. And proof of person is the foundation of identity. It's not identity, the two are different because identity means I'm a unique person, this is my name, that's my date of birth, or that's my Github account. So proof of person is like the foundational building block that makes all of these things happen.
Bankless: Sam, most people know you through OpenAI and your work with ChatGPT. What are some specific ways you see bringing OpenAI and WorldCoin together?
Sam Altman: I think it's too early to talk about the specifics right now, but as we watch this process over the next decade, hopefully it will become apparent.
Bankless: How does the actual distribution of Worldcoin work? How does it get to people? And what is the distribution plan like? How do I get WLD? How much can I get? What are the details?
Alex Blania: There are a lot of details we can't talk about here because there's regulatory uncertainty in the U.S. So we won't go into the details here, but the overall mechanism is very simple. You download an app, you get verified through an orb, and then at some point you get your physical person certificate, and then you receive WLD every week after that. The whole system is built to scale to billions of people. This will continue for the next decade or so.
Bankless: Will there be incentives for people who get your Worldcoin ID early?
Alex Blania: This whole thing is about more than just Worldcoin ID. At launch, Worldcoin consists of the three parts we talked about: Worldcoin, World ID, and World Dapp. Worldcoin is a protocol designed to scale to all of humanity, which is something we have never seen before because there has never been civil resistance before. What will be more important in the next ten years? The token or the World ID? That is unknown because we have never seen a token held by 3 billion people.
Bankless: The Worldcoin project has a lot of ambitious components. There's the orb, which is the hardware component, which is itself the cutting edge of hardware. There's the token, which has its own cutting edge distribution mechanism. There's the mobile app. There's the actual scanning work. And then we haven't even talked about the second layer that's built on top of the OP stack, there's a lot of moving parts here. So what is it like to manage a startup that needs to be at an extremely high level in so many areas?
Alex Blania: It's been a really wild ride. Because I think people always forget that at the beginning, there were really only four of us. We were all fresh out of college, so we'd never worked anywhere. Now it's a massive team of 180 people in different areas working together to achieve this. And as you said, we have a pretty mature hardware team, a protocol team, an economic team, and an AI team. There's a lot of stuff, a lot of moving parts. So it's definitely been a pretty exciting journey. And then over time, it's moved from a hardware focus early on to a product focus and now an operations focus. So it's going to be a massive operations machine.
Sam Altman: It’s been really cool watching the Worldcoin team take on new challenges. One of my favorite things about startups is watching people learn how much they can get done and how much they can accomplish. Especially with small teams. I can’t remember who I first heard this from, but there’s a somewhat famous saying in startups that if you knew how hard it was going to be, you’d never start. So it’s important to be a little naive in the beginning. Small teams of smart, driven, and aligned people can accomplish a lot more than they realize. But even with that, the Worldcoin team has certainly exceeded expectations.
Bankless: Sam, are you more involved in the WorldCoin project on a day-to-day basis? Maybe not on a day-to-day basis, but where are your touchpoints with WorldCoin?
Sam Altman: Whatever Alex needs, I'll try to help. But it looks a lot different, and I'll try to help wherever I can.
WorldCoin and Optimism Collaboration
Bankless: WorldCoin recently announced that it will issue tokens on the Optimism mainnet. Please explain in detail the reasons and significance of this decision.
Sam Altman: Okay, actually the thing that we announced last week is that the token itself will be issued on the Optimism mainnet. So we didn't announce a second layer of the OP stack. We will certainly do our own Layer2 at some point, but not at launch, which will come later this year. Why Optimism? In fact, they've given us a lot of feedback on the entire technology stack over the last few months. Just two months ago, we decided that we had to strengthen this and really start collaborating. Of course, it starts with migration, but it's much more than that. We're going to collaborate on identity primitives and of course Base. So yes, there will be many things happening. So the simple implementation is just putting WorldCoin on Optimism. That's the beginning.
Bankless: So WorldCoin will also use the Optimism chain for other infrastructure?
Alex Blania: Yeah, the infrastructure for WorldCoin will also use the Optimism chain. So WorldID is completely independent because WorldID is essentially a Merkle tree that lives on the Ethereum mainnet and then users can do inclusion proofs on it. And then this will bridge to different ecosystems. So it's not just the Ethereum case. WLD will be issued on the OP mainnet, and then later this year, we will have our own Layer2.
WorldCoin’s Expansion Strategy
Bankless: So how many irises have been scanned so far? How many people have been issued a WorldCoin ID? What has been the strategy so far?
Sam Altman: Absolutely. So far (May 14th) it's 1.7 million. All of this is just the beginning for us. So it's 210 devices right now. Most of the time it's only been about 50 devices, and that's mainly because we had to iterate on the product. Scale only matters after everything launches. So we're only a few weeks away from launch, which is why we're talking about this now. What's the strategy? In the early stages of the project, the question was, when you go out in public with an Orb, how many people will sign up? Is this actually effective? Can you generate enough throughput with one of these devices to actually scale to billions? That was our initial definition of product market fit, what scale had to be. So we built a prototype, and that was me running around Berlin, signing up people with these devices. And I remember at the time, one of my colleagues, Sandro, he did 70 signups a day. We closed our Series A at that time with those 70 signups a day. Because if you take 70 signups a day, multiply that by a week, then you get three times, I mean, it was 300 to 500 signups a week at the time. That means tens of thousands of devices are enough to scale to billions. That was one of the things that was very surprising at the time. And from that, it was like an early proof of concept in phase one. And then phase two we took it global. We had 15 devices, and we tried to deploy them all over the world and in many different markets to see if there were any fundamental barriers that we didn't know about. So we went from Tromsø, Norway, where the capacitors were exploding because they were too cold, to Nairobi, Kenya, to Latin America, all during COVID. So it was quite a tortuous journey. We had some team members who went through five different countries to finally get to the one country that we really wanted to go to because of COVID. More recently, we now have O-Pacific, a team that's more specialized in this. We're focusing on four markets, Buenos Aires in Argentina, Lisbon in Portugal, Nairobi in Kenya, and Bangalore and Delhi in India.These four markets, because they are perfect launch pads for larger regions. So if you want to build a product, these are the four regions we think you should start with. So we are basically sending product teams to all four locations. I try to go to them as much as possible, and we will expand from there.
Bankless: You can’t hire every Orb operator to cover the world’s population, right? So how do you spread the Orb around the world so that operators can use it? Do you hire every Orb operator? What’s the strategy here?
Alex Blania: No, the bigger part actually has to be a decentralized protocol. So it's not a strategy. The way it works is that when you operate an Orb, you get revenue from the protocol for every sign-up. So that's pretty much it, and there are a lot of incentives.
Bankless: Is Orb itself an affiliate program?
Alex Blania: In a sense, yes. There are a lot of incentives behind it to actually make it work, but that's where it's going to go. And then the Worldcoin Foundation is basically just going to set the standard, and then other people will be able to produce their own hardware devices and connect to the protocol. This is just the first stage of this big process, and it's estimated to take 10 years to be decentralized. Eventually everyone can produce Orbs, and at the end of the day everyone can operate them. The protocol will basically pay people.
Bankless: Interesting. Orb actually became a commercial product that anyone who wanted to participate could build. There was also some kind of financial incentive from the protocol to encourage Orb manufacturers to build Orbs. So what were the learning lessons from the grassroots experience of Orb operators? What was it like to manage these people since they were not part of the actual company?
Sam Altman: Now we have a team that has worked at Airbnb, Uber, and all these companies. They bring a lot of experience in how to operate these decentralized models. There's a lot of questions about quality control and fraud detection and so on that we don't need to go into. But the other thing is, give the user as much power as possible, so put as many features as possible into the app, educate the user as much as possible before they sign up.
Bankless: Yes. I don’t think humanity has ever seen a project this scalable. Alex, where do you think we are now in the WorldCoin roadmap? How much work is left to do?
Alex Blania: Sam said that WorldCoin's progress and professionalization exceeded expectations, so we may be a little ahead on the roadmap. We are still in the early stages of this big project, and there is still a lot of work to complete global coverage. But so far, we have made good progress and laid the foundation for subsequent work. The key is to maintain momentum and continue to iterate and expand. By building valuable products, attract more users and developers to join the ecosystem. The whole decentralization and incentive mechanism also takes time to develop. So there is still a long way to go, but the outlook is positive. With continued focus on users and staying agile, we can achieve the ultimate vision.
Sam Altman: We have three and a half orders of magnitude of growth ahead of us. It's not super early. I remember when Worldcoin first debuted, the crypto community totally derided it. I think that was to be expected, like, put your iris in this orb and you get some tokens. It was a knee-jerk reaction, and I think that's understandable. But one thing I've noticed is that it has attracted talent in a way that I haven't seen with other projects. There have been a number of people, in fact, it's happened to me twice, where I've been in some kind of social setting, some kind of social occasion, and someone from Worldcoin would pull out an Orb. People within Worldcoin loved the Orbs. So the Worldcoin team slowly continued to build and iterate and attracted some talent, attracted some builders and thinkers and developers that I respected who were not on the Worldcoin team before.
Bankless: So not only has the Worldcoin project progressed in its own trajectory, but the perception around Worldcoin has also changed. Can you talk about that experience?
Alex Blania: We didn't really want to announce it at the time. There was a leak from a really nasty journalist that forced us to deal with it. So I remember very clearly, we were a super small team. Because of the coronavirus, I lost my student visa. We were stuck in a small town in Germany called Erlangen. I remember getting an email from a Bloomberg reporter at about 11 o'clock on a Friday night. It said, if you don't respond, we're going to write about this. So all hell broke loose. Everyone was talking about Sam Altman launching this. So a dozen people stuck in a small village in Germany trying to deal with all this. It was the earliest stages. But then, as you said, we attracted incredible talent, one of the best teams in the space. And then a flywheel effect started happening, and people outside the company got excited about it.
Bankless: It’s been incredible in the last few weeks because there’s so much excitement and attention going on right now. What do you think, Sam?
Sam Altman: It wasn't as big of a shock to me because I've been through this kind of ups and downs before. But the level of personal attacks and hatred on this project was really high. I remember one day, I don't really want to check Twitter too much, I don't think it's healthy to check it too much, but one day I logged on and spent a long time scrolling through what they were saying. I was like, this is really very personal.
Bankless: Sam, do you face this situation with OpenAI as well?
Sam Altman: Yeah, OpenAI has very fervent supporters and very fervent opponents. You have to learn to live with both. I think no matter how you explain the technology in the sphere and zero-knowledge proofs, some people simply can't accept the privacy aspect. If they don't like it, they don't have to sign up. I think this is an important point to make all the time.
Bankless: Exactly. I talked about this interview in my last Bankless Weekly, saying that I had breakfast with the Worldcoin guys and there was an Orb ornament on the table, which was interesting. Some people were like, wow, Bankless is talking to the Worldcoin guys, that's shocking. So some people have very extreme views...
Alex Blania: Yeah, I've had a lot of those conversations, as you can imagine. I think 99% of the time, the answer is usually that the other person didn't actually read the documentation or didn't try to understand what was going on. If you really go through the details, you can usually calm the other person down. The bigger issue is that you don't actually need to trust us, and you won't need to trust us in the future, because a lot of things are open source. We get that feedback a lot, and I think it's completely understandable that not everything is open source yet. Obviously, we're going to move forward with that as quickly as possible. But the reality is that we have to make trade-offs. So, don't trust us. Read the documentation, read the code. If you find something you don't like, literally, talk to us, and we'll find a way to fix it.
Bankless: If this interview gets people interested in Worldcoin, what does Worldcoin need most to move forward? What do you need most to change the situation?
Alex Blania: We're hiring in basically every area. From product to AI, everything. In our case, it's not AGI, it's edge neural networks for detecting all kinds of fraud attacks. So it requires different skills, for sure. But it's really all over the place. If you're strong in operations, if you're strong in technology, please reach out to us.
A brief talk on AI
Bankless: Can you briefly talk about your stance on the AI debate?
Sam Altman: Number one, we need to make more technical advances. We've laid out a roadmap of how we think we're going to solve this problem. But it's indisputable that we need new techniques beyond LHF (language model fine-tuning). I don't understand why anyone would say that because we have LHF, the fine-tuning problem is solved. That's absolutely not the case, and I want to be very, very clear. So there's a lot of work to be done there.
Second, I think once we have the technical ability to regulate superintelligence, we will need a complex set of international regulatory agreements, cooperation among leading groups. This will require a very complex set of negotiations and agreements, and we are laying the groundwork for that now.
Third, how do we limit the abuse of AI? Human abuse of AI will cause great damage or harm to society. So how to formulate these restrictive rules is something we need to discuss.
Bankless: Great, Sam and Alex, thank you so much for joining us today.
Original link
