Automatic Transcription, it can contain errors.
Alessandro Oppo (Interviewer): Welcome to another episode of Democracy Innovators podcast. Our guest today is Tate Berenbaum.
Tate Berenbaum: Thank you for having me, Alex. It's great to be here.
Alessandro: Your project is quite interesting. How old are you, if I may ask?
Tate: I'm 22 now, but I was 17 when I got into Arweave originally.
Alessandro: That's wonderful. Would you like to tell us something about the projects you're working on? I've seen Community Labs and Arweave.
Tate: Sure. Let me start with a story about how I got into this. I've been interested in crypto since 2017. In 2019, I stumbled across Arweave completely by chance. I use the Brave browser, which gives you the ability to see ads in exchange for crypto, and one of the ads was talking about permanent data storage. That led me to Arweave.
At the time, I didn't have a use case for it, but I realized that out of all the things you could do with blockchain technology, storing data permanently was a pretty interesting one. I set it aside and came back to it when the pandemic started in 2020. I was in school, but my co-op was canceled because the school wasn't sure how to migrate to the online learning environment. I was basically stuck in my basement for six months with nothing to do except build stuff, so I started building developer tools on Arweave.
Arweave is a permanent data storage network designed to enable you to pay a one-time fee to store a piece of data for at least 200 years. The idea is creating a single source of truth where people can upload data and rely on it being available over that period of time.
My company, Community Labs, is a software development company and venture studio building and bootstrapping different types of use cases and applications on top of Arweave. We see our job as similar to explorers in new territory. Permanent data storage introduces new primitives and enables you to do things that weren't really possible with other blockchain or Web2 technologies before. We're building companies around those use cases and helping them spin out and operate on their own. We're sort of like a miniature version of Consensus (like Consensus on Ethereum, but for the Arweave ecosystem).
AO, more recently, is a decentralized supercomputer built on top of Arweave. It's really exciting for us because now we can build on top of permanent data storage and decentralized compute in the same stack without having to rely on some blockchain for compute and Arweave for data storage.
Alessandro: That's very interesting. What kind of technology is used to achieve this 200 years of safety?
Tate: It's super interesting technology. The way it works is you pay a one-time fee proportional to the size of the data you upload. That fee goes into an endowment that is then paid out to nodes over time for being able to prove that they're storing your data.
The reason it works is that if you look at a historical graph of the cost to store data over the last 50 years, on average, the cost has decreased around 30% year over year. The idea behind Arweave is essentially a bet that the cost to store data will continue to decrease. A lot of people are comfortable making that bet because it has for the last 50 years.
Also, as our lives become more digital and more people are using the internet, there are more incentives for companies to make it cheaper to store larger amounts of data. So that's how Arweave works and why it's able to offer that 200-year guarantee.
Alessandro: This is very interesting because humans do things, write things, and people of our time seek the sources that were written in the past. Now we're seeing a sort of change as a lot of sources are digital. I saw some years ago, in relation to AI, that eventually someone could scrape and modify a lot of things forever. Theoretically, someone could remove a word from all books and all images, and for people of the future, that thing might not exist. Of course, this is just an example, but that's why I think this project is interesting.
There's also another project, archive.org, which is completely different but shares some similarities, I think.
Tate: Archive.org is super interesting. I think that's the team behind the Wayback Machine, right?
Alessandro: I think yes, but I'm not completely sure.
Tate: So you're right that the similarities are both trying to make sure that data on the internet remains available over a very long period of time. The core difference, however, is that Arweave acts as a decentralized network where your data is stored in hundreds of places around the world, and hundreds of people are competing to make sure that is the case.
With archive.org and the Wayback Machine, you have to place trust in the team behind archive.org to make sure the data isn't lost. Another problem is that archive.org is carrying the costs to store and maintain that data over the years. What happens if archive.org runs out of money and has to shut down? The risk would be that the data might not be archived, or archive.org could alter or manipulate the data they're storing. I don't think they're doing that, to be clear, but the whole point is that with Arweave, you don't have to worry about that. You don't have to be concerned about someone changing the data because if they did, they wouldn't be able to earn payment for proving that they're storing it.
Alessandro: Thank you. I totally agree about the centralization in archive.org. I also don't think they are falsifying anything, but the potential is still there.
So you told us about when you had this idea, browsing the internet and seeing an ad. I was wondering how you developed that. You said you started developing some tools - how did it happen? Did you meet someone who inspired you?
Tate: The first project I built on Arweave has a sort of funny story. Both of my first projects were influenced by people changing data or taking it down, and me wanting to make sure that wasn't possible.
One of the first projects was called nest.land. The idea behind it was related to a new JavaScript runtime called Deno that had been released shortly before, created by Ryan Dahl who also created Node.js. Ryan had a pretty big following because of Node.js, and Deno was supposed to be the better version to fix all the mistakes he made building Node.js.
I got into Deno, and one difference between Node.js and Deno is that for Node, there's a package manager where you request a package and it downloads it to your local file system. With Deno, it's simpler - you can just paste a URL to import code.
This was cool because it saved time for developers, but one problem was that code people were sharing could be on links that would change or break. Your code might randomly break because one of those links breaks, and there's nothing you can do except find another link or download the dependency manually. People were sharing code on GitHub links that weren't version-controlled, so they could update, change, or delete it, breaking your code. Worse, someone with malicious intent could change the code to inject something harmful.
We needed a place where developers could share code that they could rely on never changing or going down. So for me, when I found this problem, I thought about Arweave and how permanent data storage could solve it. I spent a weekend hacking with friends on Discord, building a solution that was essentially a website where you could upload your code and get a link to import it, and it would never change or go down.
That project repurposed Arweave as a content delivery network for developers. Unintentionally, it grew from a small weekend project into a vibrant open source community full of developers and engineers who were much smarter than me. I was able to learn a lot and meet many people who are now on our team building things at Community Labs.
Alessandro: That's interesting. It was unexpected to become such a big project!
Tate: Yeah, not at all.
Alessandro: Sometimes maybe you or someone else has the goal of creating something big, but that doesn't happen. I think this is an example to always be open and try things based on your feelings, and something will happen.
It also made me laugh when you said "working with some friends on Discord" - I can imagine people passionate about these topics. In real life, it's sometimes quite hard to find someone interested in these things, while online there are a lot of people ready to stay up late coding.
What was your experience with different online communities? I think it's important because if you're alone, you might be stuck, but if you can find other people, you can create incredible projects.
Tate: That was a winding path for me. I didn't really have a process or goal to meet people on Discord. When I first downloaded Discord in 2017 or 2018, I was actually quite skeptical because I thought it was for gamers. I was a gamer, but I didn't necessarily want to talk to other gamers about games - I wanted to build stuff.
Over the years, I found two truths about Discord communities. First, many people think the best engineers work in big tech companies or FAANG in Silicon Valley. But really, if you know where to look, some of the most talented engineers in the world are 14-year-olds on Discord writing code on Replit with a Chromebook. The amount of talent on Discord is unbelievable, and these people are not only talented but also really friendly and just want to build stuff. They're not profit-driven.
I was nowhere near their skill level, but the second truth I learned was that you have to come in with something to offer. You can't expect a community to give to you - you have to provide something first.
What I did was join communities of people building Discord bots. I was an engineer but still pretty new to it, so I offered to do graphic design for free. All these talented engineers didn't want to make their projects look good - they just wanted to make them powerful. So I learned graphic design using GIMP and Inkscape, then Affinity later. I was making logos for people, and that's how I met the people I worked with on my projects.
Discord is crazy - LinkedIn is not the place to find incredible talent. I would say Discord is.
Alessandro: This is super interesting. I think a lot of people struggle because sometimes they want to do everything themselves, but as you said, there are many talented people who would like to contribute to good projects. I totally agree that on Telegram and Discord, you can find very talented people and interesting projects.
Another thing - during previous interviews, many people complained that finding funding is quite hard. What was your experience?
Tate: I think this shifts and changes with time. Right now, the markets aren't the most green, and that has a trickle-down effect of reducing risk tolerance for capital allocators and investors. It's harder now than it was a couple of years ago when I was getting off the ground, to be frank.
My journey wasn't exactly intentional either. I continued building for a few months after launching that developer tool, and I went on to build other products in Arweave because Arweave had needs and no one else was building solutions. One of those products was actually the first decentralized exchange on a permanent data layer, called Verto. We launched it in June or July 2020.
The process for getting capital was essentially building a product that was useful to people, and investors actually became users because they wanted to exchange things in the ecosystem. We met investors because they were using the platform itself.
I also met the founder of Arweave, Sam, who saw what we were doing. He realized there were about ten people working full-time for fun on this stuff just because they enjoyed it, and he took us under his wing to help us figure out how to make it more sustainable. I was in high school at the time, but the goal was to not have to go to college. Sam said I should meet some people he knew.
I got on the phone with investors, and I treated it like Shark Tank at first - using my manners, saying "yes ma'am, no ma'am, yes sir, no sir" (I'm from the South), giving a pitch, expecting them to try to get my numbers down. But what I found was the exact opposite. One investor told me not to call him "sir" - that it was weird, that this was a partnership between two people. I was giving them tokens and equity; they weren't just doing me a favor.
I learned these things are not like Shark Tank pretty quickly. Once we got our first investor to commit, it branched from there. They introduced me to people who became close advisers and also invested, and those people introduced me to others.
For people looking to raise funding, I think it's important to be able to ship quickly. If you can build quickly and figure out a product that works, that helps. We already had a product - we weren't pitching an idea. Looking back, being honest about the situation also helped. We didn't know the plan to scale to a billion dollars. We were just taking each day as it came, building what customers asked for, and trying to do it quickly. Pre-seed investors don't necessarily want to see a master plan - they want to see honesty, transparency, and people doing things for the right reasons.
Alessandro: I really like how everything happened in a spontaneous way. Some founders are also worried that with venture capitalists, they'll lose control of their project, which can sometimes stop them.
Tate: That can definitely happen if you sell too much. The game is figuring out how much you need to sell to sustain yourself and get to the next milestone. If you can't get by without selling too much, then yes, you don't want to lose control.
Alessandro: Thank you for sharing your experience. I didn't really ask anything about you personally. You said you're from the South, but if you'd like to say something more about yourself, like your hobbies or where you grew up?
Tate: I don't have great answers to the hobbies question. My job is my hobby - I really enjoy what I do, so I spend a lot of time doing it. I grew up in Virginia in the US, in the south of Virginia near a city called Roanoke. It was a pretty small town environment. If anyone's seen the show Friday Night Lights on Netflix, the lifestyle portrayed in Texas was very similar to how I grew up in Virginia.
There weren't many people interested in what I was interested in growing up, which was another reason why I went to Discord. My school wasn't teaching coding or how to build stuff when I was getting into it. Coding became my hobby, and I was fortunate to turn it into a job by starting things and building stuff.
For other hobbies, I recently started trying to get good at tennis, which I really enjoy, but tennis is hard because the courts are very crowded. You also need two people, and I have a weird schedule, so I can't just text someone at 10 PM and say, "Hey, do you want to play tennis right now when I finish my work?" I also enjoy skiing, but there are constraints there too - I live in Texas now, and there aren't exactly mountains nearby to ski. It's a bit warm here.
Alessandro: I was thinking that since you have experience in the Web3 space, are there any generic problems you've identified, like fragmentation or other issues?
Tate: I would say there are two buckets - one could be called problems, the other opportunities. One bucket is opportunities that exist within Web3 to build new products on top of the technology. I spend most of my time thinking about that in the context of Arweave and AO.
The other bucket is the structure, nature, and way that people actually do work in this industry. Tokens are a really incredible invention - the idea of creating reward vehicles to incentivize people to do things with minimal management is brilliant. It's proven itself with Bitcoin.
But the slippery slope is people conflating the equity of a company with the business model of a company and applying that idea to a token. The problem is that when you don't separate those things, the risk is that people turn the business model into selling the token - the business model becomes convincing other people to buy the thing.
In some cases, that might make sense with things like NFTs or meme coins. But for others, you enter risky territory where people market themselves as building meaningful technology, but the technology isn't actually the product - the product is the token. People just want you to buy the token, claiming it's because of the technology, even though the technology may not exist or work.
Many projects have gone remarkably far in this industry practicing this approach. It's hard for me, as someone who cares about technology and likes to build stuff, to see the industry so captivated by people who aren't always honest about what their token represents. There are risks to that which people have unfortunately had to deal with the consequences of. I don't think those risks make the entire industry invalid or not worth entering, but with great power comes great responsibility, and we're seeing that play out.
Alessandro: I've seen many projects that just want to sell tokens and describe a big product, but nothing is made, and they just raise money and sometimes seem like scams.
When we talk about the future of Web3 and building decentralized applications, how do you see it?
Tate: It's a big question. I think there are a few possible scenarios that might play out over the next 10-20 years. One with a significant chance of happening is crypto and Web3 as payment rails for the next version of the web. It's taken stable coins close to 10 years to become cool, but there's truth to the idea of making stable coin payments ubiquitous across the world because it's so much more efficient than existing systems. I think crypto as rails for the next version of finance makes sense and will happen over the next 10 years. It's already in motion and too big to be killed at this point.
What's more interesting for me is beyond finance - what areas in crypto will take off that aren't directly related to finance? Arweave has an interesting position there. One way to predict the future is to figure out what will change in 10 years, but people forget to think about what won't change (this is Amazon's strategy).
Some things we value will not change over the next 10 years. One is valuing a source of truth for data. As the world becomes more digital and globalization continues to evolve, the importance of having a source of truth will not change. That will always be a need, regardless of how much AI is doing things for humans or what humans are doing in general.
I think Arweave makes sense as that source of truth, and the next version of the web will need that type of technology both as an accountability mechanism and philosophically to give future generations the ability to learn from our mistakes. Books are one source for that, but books can be burned. We saw this in Germany when people were manipulated by Hitler because he could erase knowledge and prevent people from understanding the consequences of his actions. Arweave makes it possible to avoid that.
So in 10-15 years, I think crypto will be the way to transmit value, and Arweave specifically will be the way for knowledge to be shared - any knowledge that has value in persisting.
Alessandro: I'm thinking about governance, which is also quite political in the Web3 space. Besides the technological infrastructure, there's also a political aspect. What are your thoughts on governance based on what you've seen in DAOs and organizations?
Tate: Governance is harder for me to predict 10 years out. One thing that captivated me when I got into crypto in 2017 was the idea of creating supposedly perfect governance structures powered by imperfect people. That makes sense in theory - DAOs create a structure to incentivize people to do things.
In practice, I think it has some bugs we still need to work out. I don't know how many DAOs have actually been successful at surviving without any form of centralized governance. The core reason seems to be the tragedy of the commons - people on paper have a universal incentive to contribute value to the DAO and everyone benefits, but in practice, no one contributes because everyone assumes others are contributing. So nothing happens without someone standing up, coordinating people, holding them accountable, and getting them to do things.
That's the Achilles' heel of the DAOs I've witnessed - what prevents them from working and sustaining themselves over time. This comes from my own experience trying to launch a DAO to decentralize Verto development. We realized the exact same thing in the process.
I think there's promise in the theory of a DAO if we can perfect the model and avoid the tragedy of the commons while also avoiding centralization. But it's not perfect yet. There are many DAOs that market themselves as decentralized that I don't know if they actually are.
One of the products we're building at Community Labs aims to create a potential solution to this problem. Ironically, one of the more decentralized things is Bitcoin itself. It's headless since Satoshi doesn't exist anymore. But one problem is how Bitcoin's codebase is actually governed. If you look at GitHub, there's a specific number of maintainers allowed to make changes to what people recognize as the official Bitcoin codebase. That's crazy - Bitcoin's been around for all these years, Satoshi is gone, yet there are like five people deciding whether Bitcoin's code should change.
The solution we're building is creating a model where you can store your code on Arweave and create incentives for people to fork those codebases if they believe changes should be made, then creating a fork resolution system that algorithmically figures out which version of the codebase should be considered the main version, defined by what everyone else is using at the time.
Alessandro: That's interesting because I've also seen this centralization in the people who can decide if code can change. I'm thinking about civic participation software that mediates between different positions - like conflict resolution. Like an AI agent that acts as a mediator. If I fork your code because I have a different idea, and then we want to merge the code again, we need to talk and understand different positions.
Tate: The thesis of that product is that maybe the best governance is creating a function for anyone to do anything and creating rules for the best version to win, rather than enabling a council or committee to decide what's best. Setting the rules of the game and letting people play to figure out what version is the winner. Maybe the problem with DAOs is too much focus on empowering the group to act together rather than empowering individuals to act on their own.
Alessandro: I was talking about AI agents. You're working on decentralized agents, right?
Tate: One of the products we're working on, still in the R&D phase, is AO as a supercomputer. It gives us abilities that are different from other crypto ecosystems. One feature called "cron" enables processes (like smart contracts) to wake up and go to sleep without human interaction.
Another core mechanic is trusted execution environments, which creates an environment where you could store private data or run private computations that aren't accessible to the world while still being able to prove they're happening and validate they're correct - similar to zero-knowledge proofs.
When you combine these primitives, you can create cross-chain agents that operate from inside AO but can make transactions on other blockchains because these processes are holding keys to make transactions independently. We're excited about the use cases for DeFi, bridges, exchanges, cross-chain infrastructure, and increasing composability between AO and other ecosystems.
You could also see AO as a coprocessor. If people are building trading algorithms on Ethereum, it might make more sense to put the logic in AO and have it execute on Ethereum because computation is so much cheaper on AO. You can even run large language models inside AO, which would be very expensive on Ethereum. We've been doing this for the last six months.
Alessandro: Maybe a stupid question - if an agent is decentralized, does that mean it cannot be stopped? If it's like a smart contract that's deployed?
Tate: It's a good question with AI safety implications. Technically, there is a threat there, but you could convince nodes on the network not to run a process if it were malicious - you could blacklist it. Also, these things still cost money. If you're funding this agent and the computation behind it but run out of money, no one will run the computation because they won't have the incentive. So there are some guardrails, but it's definitely an interesting angle to consider. We don't want to create the Ultron of this world and have it live on AO!
Alessandro: About the projects you're working on, are there any problems you're struggling with that you can't find a solution to?
Tate: We're fortunate to have a great team that loves finding solutions to theoretically unsolvable problems. One of the more interesting philosophical questions we've been considering is: does the best technology always win? As we've discovered what AO can do, it's becoming evident that it's objectively the best technology - the most efficient, enabling things that no one else can compete with. But looking back in history, does the best technology always win?
The reason I think about this is because AO has many strengths, but we're still trying to figure out how to talk about it since it's so different from everything else. We haven't figured out how to pitch it to the world in a way that makes people understand its implications and value.
Personally, regardless of the answer, I think we should plan for the best technology not winning so that we can make it win even if it weren't the best. But it's an interesting philosophical debate we've been having recently. What do you think?
Alessandro: Actually, if we say "best technology" means there's one technology better than others, I think decentralized search networks would technically be better than centralized ones, but there's a cultural problem. Most people don't know decentralized search networks exist, and when they discover them, they're alone because all their friends are on mainstream platforms.
I think most of the time it's a cultural problem - you don't know how to describe your platform to people. There's still a digital divide between people who stay up coding and people who've never tried ChatGPT or other AI tools. I think the same about technology as people say about politics - if you're not interested in politics, politics will be interested in you.
Maybe in ten years, we'll see a new generation that's more technologically savvy. The best technology doesn't always win - throughout history, often the most deadly technology won, which is sad. Many times we developed technology thinking we'd be free from certain work, but then humans ended up working in different, sometimes less human ways.
Tate: It's a tough one. You can point to cases where sometimes the best technology won and other times it didn't. I think power is certainly a driver for things to win - what empowers people. Another driver is experience - what provides the best experience.
I was having this debate with an old friend who brought up QWERTY as a keyboard layout. It didn't win because it was the most efficient compared to the previous layout (I think it started with a D, maybe Dvorak). QWERTY won because it prevented typewriter keys from jamming. The pain of increased key jamming was greater than the pain of being slightly slower at typing. So QWERTY won on typewriters, and now it's on keyboards everywhere even though the jamming problem doesn't exist anymore.
That suggests the solution offering the least friction probably has the greatest chance of winning. But maybe what takes it home is the network effect - after everyone uses QWERTY, no one has a reason to move away from it because the friction of changing would be too significant.
Alessandro: When people see an advantage for themselves, they want to use the product. Sometimes in our complex society, what seems effective for me personally might not be the best overall solution. For example, if I save 20 minutes on my commute with a certain solution, maybe the real solution would be moving my office closer to my home or changing where I live.
Going back to people who want to start something or contribute - we said they can go on Discord or Telegram or Reddit and find a community. As you said, if you approach a community, it's better to have something to offer before asking for something. This is understandable, but sometimes it's not easy for most people. It can be hard to approach a community, and many people think, "I have this idea and don't want to share it because people will steal it."
I'm fascinated by how we're chatting now, and maybe someone will hear what we're saying and get inspired or contact us. This is part of "knowing that we don't know" - you're saying something philosophical, and philosophy sometimes is seen as boring, but it really affects what we do in the real world.
Tate: The last thing I'll say is that I have another friend who's more of a YouTuber than an engineer. He explained to me that while I found like-minded people on Discord by offering to do things, he found people by publishing YouTube videos that resonated with similar people. There's a term he created that I think about a lot: "energetic echolocation."
The idea is that if you're putting things out there - videos, content, products, code, logos - you're putting something out that will automatically attract the types of people who would resonate with it. People don't see the world that way - no one really thinks that they can find anyone they want as long as they know what to put out into the world.
YouTube is a good example because anyone can build a following if they know what to put out. In most cases, they're just creating things they like themselves, and naturally, people like them also like that content and gravitate toward them.
For us building a company and products, the game is creating a business around solving problems for other people. But how do you find the people you're solving a problem for? My friend Max's idea is to stop trying to find them - start putting stuff out, and for the people who do resonate with it, start small, have conversations, build it up grassroots and organically. Over time, if you're actually solving a problem, energetic echolocation enables you to find other people who resonate with that.
Alessandro: It's also related to finding people with the same intention who want to create the same thing. As you said, if you produce content, you attract people and create a center of power to change reality.
I have just a few more minutes. If you have a message for people in the Web3 community or civic tech space who are working on mediation agents or organization?
Tate: I've been a founder for five years now, and it's been an interesting journey I wouldn't trade for anything. If I had to describe what was most important for me, I would say doing things for the right reasons matters more than anything else. If you're not doing it for the right reasons, when things get hard, you'll lose the reason to keep going.
I really love Arweave. I care a lot about permanent data storage. I know it's a boring thing to think about, but it's powerful - this could meaningfully impact the lives of generations beyond us. Every day, regardless of how hard things are, I can't think of anything better to spend my time on than building on top of Arweave because of that potential impact.
For people looking to get started, if you're building things and putting them out into the world, that's the best shot at getting funding and finding like-minded people. Don't be afraid to socialize your ideas for fear of someone stealing them. If you're afraid of that, maybe you should just build it. A few years ago, it was more important to learn how to code, but now you can log into Cursor and have AI build it for you.
Alessandro: I agree, it's incredible what AI can do now. This was a great conversation. I really appreciated it. Thank you a lot.
Tate: Thanks, Alex.
