Arweave’s Sam Williams on Storing Files On-Chain for Eternity

Sam’s Crypto Journey

Sam first learned about Bitcoin around 2010 or 2011, when the tokens were worth about 3 cents. He was intrigued by the concept of “magic internet money,” but didn’t take it seriously until a few years later when he started mining Bitcoin with a friend from university. At this time, there was a wave of innovation happening in the blockchain space, with people experimenting with different ways to incentivize behavior through decentralized networks.

He invested in the Ethereum ICO and a project called Storage but eventually dropped out of the loop until early 2017 when he became concerned about censorship on the internet. He saw the potential for blockchain technology to enable free speech on the internet and became more interested in the space again.

Experience with Bitcoin Mining

• As a fun experiment, I began mining Bitcoin.
• Became more interested in the potential of blockchain technology.
• Saw the potential for incentivizing behavior through decentralized networks.

Interest in Decentralized Technologies

• Invested in Ethereum and Storage ICOs
• Found the concept of trustlessly storing data to be interesting.
• Realized the potential of smart contracts for trust-based finance
• Continued to research and use decentralized technologies for their potential to combat censorship and support free speech.

Free Speech

Free speech on the internet is a fundamental part of modern democracy.

Free speech is a topic in today’s world. We live in a time of “cancel culture,” where if someone is offended, they want to erase that person from their world. This can be dangerous because we need to be able to debate with people. Even with the Balenciaga ad, I was outraged, but then I realized it was free speech and they hadn’t broken the law. What’s your opinion on those recent events?

Sunlight really is the best disinfectant, and in a democratic system, we have essentially all made the bargain that we are going to have to be exposed to ideas that we don’t like, but we do so that we can correct those wrong ideas. Then we can come to a better consensus as a group. And realistically, the group is also going to get it wrong a lot of the time. But it’s better than dictatorship, which is the only real alternative—or technocratic dictatorship, as you might call it—and which appears to be a significant part of the republic’s suggestion.

Incentives

Incentives and coordination are the two things that we can use to make the world a better place. I would love to understand if, for magic internet money, it was the fact that we could use this technology to change incentives that really made you start to mine and dig in.

In the case of Bitcoin, it was less about the incentives and more about the fact that you could have a trustless mint. What if you put the monetary system outside the control of any individual or group? That’s really an astonishing thing that’s never been done before.

The Ethereum blockchain was the first to experiment with protocol-based mechanisms. Arweave is now a big experiment in the area of permanent data storage, where data is stored permanently and is held in common for all people. The cost of storage declines at a high rate, allowing for a system where there is a good incentive for people to store data. Arweave aims to create a global, permanent, mutual ledger of information that is democratically accessible.

Web Hosting with Ethereum

What do you think the web3 stack will look like to be able to host a website in web3?

Arweave it’s a ledger of speech, but what can that speech be? It can be anything. As it turns out, you can point your web browser to an Arweave gateway stroke transaction ID or even use something called the Arweave name service. Just a name like, yeah, sam.arweave.dev is an example, which is my personal website, and it will be served back from the network for you.

While some pieces of data are the full web application itself, other things are, for example, my tweets. So there’s this thing called “Public Square,” which is a data protocol on Arweave, and it’s basically just a way of saying, “Okay, I uploaded a piece of data, I tagged it the right way, and now it shows up.”

One example is The Graph, which very helpfully indexes all the data so you can add tags to your data. The decentralized nature of the platform allows for fully decentralized web applications that can be queried in a decentralized manner. However, there are concerns about how to make users aware of the decentralized nature of these applications, as they may not always be fully decentralized.

The end user might think that they’re using something decentralized, but the developer disappears or what have you, and boom, the whole thing breaks. And if it breaks when the developer leaves, it was not decentralized because it was centered on the developer.

I believe in the importance of decentralization in permanent applications. I think that central points of failure are a problem, and that people often take shortcuts and centralize parts of their applications for convenience. I’ve been advocating for decentralization for a long time, but many people have not cared about it. However, I believe that decentralization is becoming better, faster, cheaper, and more robust and that people are starting to care about it.

The tooling is starting to get to the point where it’s as easy to store a piece of data on Arweave as it is on S3 at this point.

Thoughts on the OFAC-sanctioned blocks on Ethereum

I remember the early days of Ethereum and its ethos. However, I’m less excited about its potential now. I don’t think it’s the only project in this space, but the influx of capital over the last decade has caused it to lose its “cyberpunk” roots. I hope Ethereum can find a problem to solve and stick to its principles. A lot of its uses so far have been more like digital gambling than improving the rights of internet and financial system users. I’m hopeful that it can get back on track, but the trend of crypto booms and busts worries me. I’m also concerned about the idea of an “eternal September,” where a sudden influx of new, uneducated users destroys the culture of a community.

Decentralization, open-source, and trustless technology that is permissionless but also censorship-resistant are, in my opinion, the core values and principles of this space. Censorship resistance is the one that’s being relearned right now. Would you agree with those principles? Is there anything else that you would add?

Decentralization is not an end in itself, but it is necessary for censorship resistance. Without censorship resistance, there is no reason to have decentralization. The problem with Ethereum is that it has given up on censorship resistance, which means that someone will eventually ask why it needs to be decentralized in the first place. This person will be right, and they will build a faster, cheaper, and more private version of Ethereum that runs on something like AWS. This is a dark vision of the future, but it is possible.

Decentralization is not a means to an end. Why do we need decentralization? because we need censorship resistance more or less. And that leads me to one of my biggest concerns about Ethereum, which is that if they’ve given up censorship resistance, someone at some point is going to be like, “Wait, why does this thing even need to be decentralized?”

Technology that is trustless and permissionless allows us to build systems that are more open and transparent. It is like an extension of the principles of democracy. It gives us guarantees and rights as users that we cannot get from a company. Protocols can give us these guarantees, whereas companies can only give us promises. This technology has the potential to be pro-social, and it is exciting because it allows us to apply the principles of democracy to the internet. Instead of trusting people’s promises, we can trust the code and the math. Math is more immutable than people.

Arweave’s Founding Story

In early 2017, I had the idea to create a globally replicated, blockchain-based database of newspaper articles. I was on a mountain with my father when I came up with the idea and spent a few minutes jotting down my thoughts. By the time I came down the mountain, I had a plan in mind and called some friends who were working on PhDs in related fields to form a team. We started preparing marketing materials and even built a prototype for our product, Arweave. We wrote a paper and raised funds, but as the market crashed, we shifted our focus and joined a startup accelerator in Berlin. We launched Arweave on the 69th anniversary of the publication of “1984” and continued to build and grow our network.

Arweave is a distributed storage network that provides a permanent, tamper-proof record of data. It was founded in 2016, and its mission is to provide a permanent archive of historical records. Arweave uses a blockchain to store data in a way that it can’t be altered by any party, and it has been used to archive articles, social media signals, and even NFTs. The company’s goal is to make sure that data can’t disappear, ensuring that users can access it indefinitely.

Trust, Biases and Political Situations

This is important for protocol creators to try to ensure the protocol’s integrity and neutrality. So when things were archived out of Ukraine and Russia, we didn’t express an opinion about that. Our opinion is that, first, it’s important to keep a neutral ledger of the past so we can learn from it in the future. Second, in general, free speech is good. The protocol is engineered to pursue those goals, and we try very hard to maintain neutrality as best as possible, even when we have opinions about geopolitics and local politics.

We are building a coalition to push forward these two ideas and nothing more. In the industry today, this is not respected. Satoshi did this best by saying it would be good if we could build a system of international money where no one has to trust any other party apart from math to make it coherent. They said their piece and then stopped speaking, and it still worked.

My hunch is that Satoshi was still active in the early Bitcoin community but under a different pseudonym. It may have been a mistake to use real names and be doxed on crypto Twitter when we started Arweave, but it is what it is. I personally try to keep my speech to those core ideas and keep everything else to myself.

It’s important to keep a neutral ledger of the past. Let’s just store the information about the past so that we can learn from it in the future.

Use Cases

• “Songs of April”
• Pro-democracy newspaper in Hong Kong
• The first provably neutral newspaper in the world

I find it pretty amazing that sometimes we wake up to news from China that we had no knowledge of before. For example, one time we woke up to find that people had been archiving a pro-democracy newspaper in Hong Kong onto the network. Another time, there was a video called “Songs of April” that collected audio recordings of emergency services calls in Shanghai during the lockdown. The video was shared widely and then censored on social media, but it ended up on our drive and caused a huge spike in traffic from China.

I also find it meaningful that we’re building provably neutral web applications on top of the Arweave network. This allows us to create web applications that are provably neutral, trustworthy, and out of the control of the people who upload them. For example, we’ve created the first provably neutral newspaper in the world. It uses bonding curves to estimate the crowd’s belief in the truthiness or non-truthiness of a statement and ranks new articles based on that data. Even the person who uploaded the app has no privilege to change how it works. I think it’s going to be amazing.

One of the main use cases of Arweave is immutability with NFT metadata. Can you talk a little bit about that? How does it work? Which NFTs are stored on Arweave today, and how much would it cost for someone to store their NFT?

All permanent data storage can only exist within a risk model. Otherwise, it’s magic beans. It doesn’t exist.

I think the interesting thing is that even though permanent storage is permanent, it still exists within a risk model. In our case, we’re saying that the cost of storing data should decline at a rate above 0.5%, or we’ll only store it for 200 years or something like that. All permanent data storage can only exist within a risk model; otherwise, it’s like magic beans and doesn’t exist.

You might think that 200 years sounds expensive, but it turns out that storage is actually very cheap. For something like an NFT image, it typically costs only a few cents to store it. In general, it ranges from half a percent to a cent per megabyte. So if you have an NFT image, it will normally cost one or two cents to upload it permanently to the network. That’s pretty cool because it’s very cheap compared to the cost of generating a smart contract on Ethereum. During the boom’s hype last year, it cost $30 to generate an NFT.

Even something like Solana has relatively low storage fees compared to the cost of small contract interactions. So the price is pretty low compared to other systems, which is great. And you mentioned how it works and how you can access it. The answer is that you can use something called the bundler network to upload data to Arweave by paying in tokens on other networks. It covers all of the base networks that people use for smart contracts, as well as the L2s on Ethereum.

You can pay in the native tokens on those systems to the bundler network, which will then use the Arweave token to store the data on the Arweave network. It’s a pretty neat system that makes it accessible to anyone who wants to store data.

What would you say to those people who maybe think the idea of storing it permanently isn’t the right idea here and that maybe they just want to store it for a short period of time?

It depends on the use case.

I think it depends on the use case. Arweave is not suitable for storing short-term data, but it is great for distributing records of history or financialized data that needs to be permanently publicly available. It is also possible to create a protocol for ephemeral data, but it is a harder problem to solve.

The product’s market fit seems to be its permanence.

Absolutely. That is what the system is, and it’s highly specified and built for that use case.

I think that Arweave can scale particularly well because it offers arbitrarily large bundles of transactions. This is possible because it has permanent data storage. Temporary, decentralized data storage has not reached the same level of product-market fit that Arweave has. Decentralization is a means to an end for Arweave, as it is necessary for a resilient network and a long-term endowment.

Controversial Use Cases

• There have not been many negative use cases of the network.
• There are mechanisms in place for dealing with negative use cases when they arise.
• Minority speech is protected.
• Node operators are responsible and accountable for the data they store.
• This leads to a decentralized system of content moderation.

Arweave Token and Mining

• Arweave tokens are used to buy storage on the system.
• 84% of the tokens are put into an endowment, and 16% goes to the miner who mines the transaction into a block.
• The endowment accumulates tokens over time, which can be used for rainy days when the cost of storage declines.
• The endowment is a network-controlled zone that provides a buffer for future storage costs.

I am excited about Arweave 2.6, which will hopefully enable users to mine with just a hard drive. This will make it possible for regular people to take part in the network and increase decentralization. The hard fork is planned for mid-January.

How are you defiant?

I don’t like to define myself as anything. I let other people do that.