Fitting In
Three Things #96: December 3, 2023
It’s pretty clear what the next 6-12 months look like for Spacemesh: improving and fixing all of the things we’ve already shipped, and keeping up with rapid network growth. The number of miners and the total amount of storage space is growing rapidly each epoch. We just exceeded 100,000 miners, and each miner needs to generate and broadcast a proof each epoch. This is a lot of data and we need to continue to improve the protocol to ensure that the nodes and miners can handle it all. We need to make more improvements to the P2P network and we also need to fix the PoST incentives. We have our work cut out for us.
What’s a bit less clear is what comes after this. We released a roadmap recently that lays out the general direction of travel and includes many of our goals, but it’s not terribly specific. We do intend to do all of these things but the roadmap is somewhat aspirational and high level. The timing and order of these things is likely to change.
Putting together the roadmap was nevertheless a good exercise, but at this stage no one knows for sure what Spacemesh will be when it grows up. I previously laid out some wacky, fun ideas, but every day it becomes more important to have a down to earth, concrete, realistic roadmap.
While we don’t know exactly which direction Spacemesh will go, there are several nascent themes or theses in the crypto ecosystem that are growing rapidly and that are relevant to Spacemesh. It’s easy to make fun of investors and their silly obsession with theses, but the best theses are helpful because they speak to broader, longer term, important stories, and because they give people an anchor to understand what we’re working on, why it matters, and how to communicate it, i.e., how it all fits into familiar narratives and worldviews. Here are three themes that come up again and again in conversations with investors, team members, and community members. Spacemesh isn’t perfectly any of these things today but it could become them.
Two quick caveats before we dive in. Firstly, these are all extremely preliminary. Each one, especially the third, would require an enormous amount of work rearchitecting and reengineering the protocol to make it compatible. Secondly, even more importantly, these are just my own ideas, not necessarily those of the team or the community. I’m just one voice and Spacemesh will chart its own course forward as it matures and as the ecosystem and community governance mature.
That being said, if any of these ideas are especially interesting and if you’re interested in working on any of them, please let me know!
Thing #1: Bitcoin Programmability 🤖
Bitcoin is obviously most well known as the first successful, decentralized form of money, but the idea of using it for applications other than money has arisen from time to time. One of those times was around 2013-2014 when a number of projects were attempting to build other applications on top of Bitcoin: Primecoin, Mastercoin/Omni, colored coins, etc.. Later there were other, more sophisticated attempts such as Rootstock and Blockstack (Stacks). Ethereum grew out of this era. Most people today probably aren’t aware that Vitalik first tried to build Ethereum on top of Bitcoin and only decided to build a separate chain after Bitcoiners rejected the idea. After the success of Ethereum, those interested in greater programmability mostly moved to Ethereum and Bitcoin programmability didn’t get much attention for a few years.
Today, however, Bitcoin programmability has become a hot topic again. This is due to a combination of reasons. The most obvious is new features added recently to Bitcoin, especially Taproot. In combination with Segwit, this allowed new applications like Ordinals and BRC-20 tokens to appear; BitVM will allow even more types of applications. The Lightning Network is developing rapidly. There are projects trying to build DeFi on top of Bitcoin. Another reason is the proliferation of the L2 ecosystem on Ethereum and the realization that rollups might be possible on Bitcoin as well. There are a few other attempts to develop different sorts of L2s on top of Bitcoin including Liquid, ARC, and Fedimint.
Bitcoin has many strengths but programmability isn’t one of them. People who are not Bitcoin maximalists, including those in the Ethereum and Solana ecosystems, mostly watch efforts to build things like DeFi on Bitcoin with equal parts amusement and pity. Indeed, Bitcoin maximalism is one of the main drivers behind this phenomenon, and maximalism isn’t a good reason to build a thing. There have been many failed attempts to deploy true programmability on top of Bitcoin in the past that went nowhere, and frankly most of these attempts will probably also go nowhere. Why try to build things on a chain that doesn’t have the required tools and in an unfriendly ecosystem when it’s so much easier to build and deploy on Ethereum and other smart contract platforms?
But these responses miss the point of building on Bitcoin, which has its advantages. For one thing, Bitcoin is still the most decentralized, censorship resistant chain, and the one with the most credibly neutral governance (i.e., without a known founder who can change the protocol by sheer force of will). These features matter a lot for certain types of applications. To the extent that applications built on Bitcoin use BTC or the Lightning Network, there’s much less fragmentation of liquidity than in Ethereum and other ecosystems into hundreds of rollups and thousands of shitcoins. These applications can plug into the permissionless, global payments network of Bitcoin, which is already established and trusted more than any smart contract chain. Regardless of the noise that Ethereans make to the contrary, Bitcoin is the only really established, credible, successful, global, digital money.
Spacemesh isn’t Bitcoin, of course, but it is like Bitcoin in some key ways. It’s already approximately as decentralized as Bitcoin, at least in terms of the number, geographical distribution, and political distribution of nodes and node operators (if not yet in terms of governance—real decentralization takes time). It can’t currently do a lot more than Bitcoin since we haven’t yet rolled out the VM and smart contract support (that’s coming but we want to take our time and get it right). Its economics are modeled on Bitcoin’s. And we intend its governance to function a lot like Bitcoin’s as well.
These things didn’t happen by chance. We’re cypherpunks at heart and we share most of the values that led to the creation of Bitcoin. We intend Spacemesh to be built and operated in the most meaningfully decentralized, credibly neutral fashion possible. We have a long way to go but we’re off to a good start.
And, of course, Spacemesh will be highly programmable. We’ll have a full featured, Turing complete VM. Unlike Bitcoin, this is an intrinsic part of the Spacemesh design from the very beginning, not something bolted on after the fact at a higher layer.
I don’t think it’s unreasonable to think of Spacemesh as an account-based, natively programmable version of Bitcoin. In other words, it’s something like what Bitcoin would look like if it had been designed today.
Thing #2: Proof of Work ⚒️
Over the years I’ve noticed that everything in blockchain tends to happen in cycles—not just Bitcoin programmability. This should be no surprise since all of society works this way: what’s old will eventually be new again; what’s out of fashion will come back into fashion.
The first blockchain ran on proof of work. So did the second blockchain, and the third. But over the past few years proof of work fell out of favor. Pretty much every serious blockchain that launched over the past four years uses proof of stake or some variant thereof. Proponents of proof of stake cite several motivating factors including cheaper security, but I think the biggest one, and the biggest reason proof of stake became so popular, is that it’s much less energy intensive than proof of work.
Proof of stake thus perfectly fit the ESG narrative that took over the corporate world during the same time period. (Nevermind that proof of stake is by definition plutocratic; most people don’t even know what the “G” in ESG stands for anyway.) For a few years companies and countries alike were tripping over themselves to come up with the cleanest, greenest narrative and to convince the world that they’re the most serious about reducing emissions. The Greta generation really cares about this stuff.
As it turned out, however, companies took ESG a bit too far, with some putting environmental or social goals ahead of profitability and competitiveness. Like so much else in modern culture it became more performative than effective. Fortunately, the ESG narrative has begun to fade. There are lots of reasons: greenwashing, the Russia-Ukraine conflict, more pressing concerns like inflation, and the overall silliness of the idea.
It seems like the proof of stake narrative has also begun to fade. One reason is that we’ve seen the dark side of proof of stake including how it leads to massive centralization and even censorship. While it’s true that proof of stake is greener than proof of work, people have begun to wake up to the fact that Bitcoin mining is a lot greener than most people realize and that a lot of the energy-related attacks on proof of work mining were overblown, unsubstantiated FUD.
I think another, less talked about reason is that the novelty factor has begun to wear off. Yes, lots of big chains are running proof of stake now. Yes, they seem reasonably secure and haven’t suffered major attacks. No, proof of stake hasn’t fixed big problems like governance or scalability. And the elephant in the room: blockchains are still struggling to achieve mass adoption, to build apps that aren’t ponzis and that normal people care about, and to repair their tattered reputation. Proof of stake doesn’t fix this either.
One of the big downsides to proof of stake, which I’ve written and spoken about many times, is that it’s not permissionless and that it’s not easy for those who aren’t already crypto savvy. You need to set up a wallet, buy coins, and figure out how to stake those coins safely. Proof of work can be a lot simpler: you just run an app, or in the extreme case plug in an appliance (see next thing), and you start earning coins. No credit card, no bank account, no KYC, and no cryptocurrency exchange needed. Anyone, anywhere with decent hardware and an Internet connection can participate. (This is a big part of our motivation for building proof of spacetime, a greener variant of proof of work, for Spacemesh.)
These are some of the reasons that proof of work seems to be having something of a renaissance. The first proof of work summit took place recently. There are a number of new proof of work blockchains launching, and several that are already running have been doing very well. Maybe it’s nostalgia, maybe it’s the simplicity and elegance, maybe it’s its contrarian, meritocratic nature, maybe it’s the recognition that AI is anyway going to use orders of magnitude more compute power and energy than blockchain—whatever the reason, the appeal of proof of work is clearly growing.
If you asked me to explain this trend I’d say that the world at large feels more and more unfair to everyday people. Bailouts happen all the time, the Cantillon Effect is real, and while the rich continue to get richer life keeps getting harder for ordinary people as healthcare, education, housing, and even food feel increasingly out of reach. Proof of stake is plutocratic through and through and leads to even greater centralization of wealth and power. By contrast, proof of work, when mining is possible using existing hardware as in Spacemesh, is a much more level playing field and is credibly neutral.
A rising tide lifts all boats and Spacemesh seems to be benefiting from this trend as well. We built proof of spacetime as a hybrid that has the best elements of both proof of work and proof of stake: it’s permissionless and resilient like proof of work but green like proof of stake. I’m not aware of any other protocol or consensus mechanism that can claim both of these.
It’s gratifying seeing the world begin to wake up to some of the risks and downsides of proof of stake and the upsides of permissionless proof of work. It’s humbling to see that others value the things we do. But this is just the beginning of a trend. We have our work cut out for us, and while it’s wonderful having a tailwind for once, we’re not going to stop or change course even if this changes, as it eventually, inevitably will.
Thing #3: DePIN 📡
I learned the term DePIN recently but it’s an idea that’s been around for a long time. DePIN stands for “decentralized physical infrastructure network”: i.e., allowing people around the world to contribute hardware to a decentralized, permissionless network in exchange for a share of the network’s revenue. The basic idea is that centralized infrastructure providers, including cloud computer service providers and mobile phone operators, have to charge a lot for their services to cover their overhead, whereas decentralized networks governed by open protocols can charge a lot less for access to the same infrastructure.
The first DePIN projects I learned about were OG projects from the Ethereum ecosystem including Livepeer and FOAM Protocol. Recently the DePIN idea has really taken off in the Solana ecosystem, where projects such as Helium, Hivemapper, and Teleport have gotten a lot of attention. There are lots of types of physical infrastructure that can be decentralized. Helium and FOAM protocol are based on specialized hardware that provides Wifi or 5G access, and radio beacons that provide proof of location. Hivemapper and Teleport use ordinary hardware such as automobiles and mobile phones to compete with the likes of Google and Apple Maps, and Uber and Lyft, respectively.
Of course physical infrastructure naturally also includes digital infrastructure: CPUs and GPUs (compute), and hard drives (storage). Livepeer allows anyone to transcode video files, Filecoin makes it possible to monetize spare storage, and projects like Gensyn are working on letting people monetize spare GPUs in a truly decentralized, permissionless fashion. These are not trivial achievements: on top of all of the usual complexity of peer to peer networks, consensus, sybil resistance, etc., these networks additionally need to deal with problems like dispute resolution, preventing and addressing fraud, payments, etc.
There’s something about DePIN that I find particularly appealing—much more appealing than the shitcoin casinos that have dominated the space for too long. For one thing, as I mentioned above, it allows ordinary people to invest in and monetize assets in a permissionless, credibly neutral fashion. For another it solves real problems: e.g., making access to compute resources an order of magnitude cheaper than on centralized platforms controlled by tech giants. As Gensyn points out, we don’t want to live in a dystopian future where only the biggest, wealthiest companies have access to training and inference. We want to live in a future where we can perform both training and inference on our own models using local resources. DePIN could go a long way towards democratizing access to these resources.
For machine learning models to proliferate throughout our world in a fair and equitable manner, it is crucial that we establish the right to build. Not as a legal right enforced by governments, but as a technological right enabled by building with a new perspective - decentralisation. This right is very simple - we should all be able to access the resources required to build new machine learning models. We’re in a race against competitors aiming to dominate our future by controlling an enabling technology that will radically alter our society. - The Machine Learning Compute Protocol and our future
In a nutshell I’m beginning to see these networks as one of the purest, most ideal applications of blockchain infrastructure beyond just money. Web3 is about more than censorship resistance and self-sovereignty. It’s equally about disintermediating unaccountable legacy tech giants that extract maximal rent and make everything shitty in the process. One of the superpowers of Web is its ability to reduce the take rate that these companies and platforms can extract from users through decentralization and better incentive schemes. DePIN is a manifestation of all of this.
What does all of this mean for Spacemesh? As I wrote a few days ago, what is Spacemesh if not a decentralized physical infrastructure network of tens of thousands of GPUs and hundreds of petabytes of storage? That’s a lot of horsepower. And both are currently underutilized on Spacemesh: the GPUs because they’re only needed for the initialization phase and the hard drives because they’re currently storing cryptographic junk that’s meaningless outside of the Spacemesh protocol. If other protocols like Livepeer, Filecoin, and Gensyn have figured out how to put these resources to valuable use there’s no reason we can’t do this, too (although it would require a major change to the protocol).
Today Spacemesh has successfully built a “People’s Coin” that allows ordinary people around the world to mine from home sustainably, profitably, and permissionlessly, and to earn their first cryptocurrency with no need for KYC or a bank account. That’s a huge accomplishment. But I’ve always felt that Spacemesh could be a lot more than this. I’d be really excited to develop Spacemesh into a more useful platform that allows our miners to generate even more value with their spare compute and storage resources. We have tens of thousands of miners that I know would be excited to participate, and while the technical challenge is real we also have a world class R&D team that could make it happen. This is a promising, timely thesis that I think we should seriously consider.