# Decentralization Is Really Really Hard | Lane Rettig - Spacemesh

- Speakers: Lane Rettig
- Channel: [Ethereum Denver](https://streameth.org/ethereum-denver)
- Date: 2024-10-09
- Duration: 24:49
- Topics: blockchain, web3, ethereum, Entertainment
- Watch: https://streameth.org/watch/yt-qP902Bdg7KQ
- YouTube: https://www.youtube.com/watch?v=qP902Bdg7KQ

## Description

APPLICATIONS FOR ETHDenver 2024: #YearOfTheSpork ARE OPEN


APPLY NOW
https://ethdenver.com/apply

## Transcript

intro um seconding everything she just said there's a lot going on at e Denver it can be a little overwhelming but please do take the time to explore and make sure you get over to the Biddle Hub on the other side make sure you see the Zen Zone upstairs food trucks lots and lots of of stuff going on um I could not be more excited to be here I've been at most of the eth denvers they've gotten bigger and better every year really excited and honored to be here and just have an opportunity to share the space with you guys um my name is Lane I am a developer working on a project called space mesh uh we are going to talk for the next 20 minutes or so about a topic that is near and dear to my heart hopefully to yours as well decentralization it is a term that we tend to throw around all the time without always being very thoughtful about what we mean by decentralization and so my goal here is to talk about things like how we Define decentralization how we should think about it right um why it matters so much and also how we measure it and then I'll also talk a bit about um decentralization in the context of this Maes project okay good so tiny bit of background on myself I've been in the ethereum community since 2017 I uh worked as an ethereum core developer with the ethereum foundation for a couple of years I was initially attracted to the ethereum space and I should mention as well the first eth Denver is one of the main events one of the main reasons that I decided to join this community full-time um the the thing about ethereum for me and other blockchain platforms and just kind of the bundle of ideas and Technologies is that it's a opportunity for democratization right I have a background in traditional Finance um and I've always thought of ethereum as an operating system for building better human institutions and as a platform for building let's say value creation networks to allow people around the world to participate in a incredibly neutral decentralized fashion to create value for themselves their families their communities um that's very much what I've been focused on over the years um I joined a fledgling layer one project called space mesh in 2019 that's what I've been working on since then for the past five years uh the biggest and most important piece of news I have to share about space mesh although it's not exactly new news at this point is that after five years of grueling R&D we finally launched the the main net in July of last year and uh I'm really excited this is the first eth Denver when I've been able to say space mesh is now alive um really really exciting if you have not had a chance yet to kind of check out space mesh to download the software to join the network to mine Etc from home um now is a great time to do it because the network is much more mature and more usable than it was eight months ago when we first launched um yeah it's been an exciting few few few months and few years um space mesh I should mention so we have two primary goals with the project the first is to become the most decentralized blockchain in history I think we are well on our way on that journey and I'll share some updates in a moment on how it's going um and the second goal is to enable able ordinary people around the world to join the network as miners right not as a light client not as a full client but actually as a full mining Pier a validator on the network using existing Hardware whatever computer you have whatever internet connection you have um and to do that sustainably and profitably and actually those two goals are the same right in order to have millions of miners on the network it needs to be very easy for home miners to join the network so that's what we're working on at space mesh um so in the next few minutes we'll talk about first some remedial 101 on decentralization itself I'll talk about how space mesh is doing towards that goal um talk about some of the challenges we're facing on the road map to decentralization uh how we're addressing those challenges uh and kind of looking forward to the future um of space mesh as well and how we're going to improve so first of all what is decentralization again just to reiterate it's a very important idea it's kind of the reason we're all here in the first place right remember why you started but again it's something that we don't always um communicate about very thoughtfully or very rigorously so our our dear friend vitalic actually has uh what I think is still the best answer to this question he posted a post on his medium seven years ago it's a lot of years in crypto years right um called the meaning of decentralization um I could easily spend an hour just talking about this um obviously we don't have that much time today I strongly encourage you to read this article if you haven't uh I also published a substack article I have a substack I'll have the link at the end where I did some deeper analysis and reflection on this as well more than we can get into today but we'll just do a very very high level overview so um the first three here and just to be clear in this section most of these ideas are vitalic ideas I'm just going to kind of add some commentary and a few kind of tweaks so he splits decentralization up into the first three categories I've I've added the fourth one um logical decentralization is kind of like is this thing a monolithic entity or is it more like a a morphous swarm that's the term he uses blockchains are only useful if they are logically centralized right so this is why it's so important to understand decentralization because actually there are aspects in which ways in which we need blockchains to be centralized they need to be logically centralized to give us something like a single Ledger a single view on Truth a single state machine in the case of ethereum um architectural decentralization this is going to be the focus of this talk this is mostly what we mean in this community when we talk about decentralization uh a very reductionist view of this is how many nodes are there on your network we'll analyze it a bit more deeply later um and then there's political and I added economic to this list I added economics specifically because in the context of proof of stake I think it's very important to understand things like the Nakamoto coefficient which is an imperfect metric but it tells you the number of validators that would have to collude to Halt a proof of stake Network um but basically it's like how many decision makers are there how many doors do you have to knock on if you want to like change something or stop something um but again we're going to focus on the architectural primarily here um and the second piece of his article talks about why decentralization matters also a very important question again important very nuanced we don't have time to go deeply into this this right now um but fall tolerance is uh non-controversial fairly obvious it's Sound Engineering practice it's been a thing since way before blockchain um attack resistance I think of as a specialized case of fault tolerance where the faults that we're trying to prevent against are intentional rather than accidental and collusion resistance I'm not even going to go into this right now it's again like super super nuanced and super complicated but it's another goal of blockchains in general and definitely a goal of space mesh cool so let's jump into the kind of cre piece here on on Space mesh and how we're doing um you know every talk I've given about space mesh at this event previously has always been like a theoretical talk and I'm just so thrilled that this is more of a in practice talk because the network is now live and we'll see some live numbers so we have this thing called the network dashboard if you go to dash. spaces. in the web or just Google spacemesh dashboard you'll you'll find this um I don't know if this has a laser pointer yeah so we have some some some exciting metrics here you'll see a lot of these kind of like like hockey stick shaped graphs which is really really exciting so we have now more than 1.3 million Smashers on the network um I'll talk about what this means in a moment um you know transaction volume is growing steadily circulation in coins security I'll talk about this in a moment as well uh there's no feeling more satisfying as a builder and as an entrepreneur than to be able to stand on a stage and put up a slide like this I mean this validates this is our hockey stick graph right this validates all of the hard work The Blood Sweat and Tears that we've been putting into this project for six years um so let me explain what this means right so we are uh obviously time on the x-axis here the y axis is what we call minor identities so we have over 1 million I think we're actually what did this last thing say 1.3 3 million these do not map directly to miners as we think of them in the sense that these are not like people or organizations or computers on the network the way the space mesh protocol works right now it's very common for a single minor to have a dozen or in extreme cases even hundreds or thousands of identities which just refers to like the individual plots of hard drive space that they committed to the protocol um so you know I wish I could say we have a million minors uh I won't make that claim we have millions of minor identities it's probably a couple orders of magnitude less than this we don't know the actual number of exact minors but really Healthy Growth uh another hockey stick graph right so this is the security space mes uses hard drive space for mining unlike proof of work unlike proof of stake um we are at I believe 1.3 exabytes with an e that's a crazy number like just I can try to give you some sense of it it is 250 million songs which is twice the size of the entire Spotify catalog right it is also 10 times the entire storage of the internet archive Foundation they've got 100 pedabytes of storage we're 13 times that right now um we're about 60% of the size of the storage the actual data being stored on filecoin and I think this is insane given that we only launched this network 8 months ago um and if you were to 51% attack space mesh you would need to have at least this much space so that's why this is our security metric this is actually the most interesting of the three slides I'm going to put up um in the sense that this m this does map more or less directly to like a real world thing and that thing is unique peers or nodes on the network this is kind of imperfect the data here is kind of like smooth and aggregated so like don't pay too much attention to the dips and the bumps um but you know we're at around 20,000 unique peers on the network uh public peers right the number is actually higher because and I we don't know how much higher because not all of the nodes or minors or peers in the network are visible to to us to our nodes many people are running private nodes nodes and you know private internet connections Etc behind routers and Nat and things like that um but this is significantly more nodes than our live today on bitcoin or ethereum right which are both in the order of thousands not tens of thousands um again I'm just I'm very proud of this uh still some Healthy Growth here we want to get this to a million that's our goal um okay so back to the dashboard again I didn't bring this up before but we have this weird thing down here called a decentralization ratio right I talked about measuring decentralization right what does that mean it says 81 out of 100 so I guess we have like a B minus right now um this is a fairly arbitrary metric that we threw together it in no way shape or form is like perfect or you know doesn't capture all the dimensions of decentralization by any means um it's it's this kind of like subjective metric we came up with for now which is a weighted average of two things the first is the number of those minor identities and the second is a genie coefficient which I think is very interesting right you'll recall I mentioned economic decentralization so it looks at the genie coefficient which is a measure of how uh evenly or or unevenly distributed wealth or income is in a population so in our population of miners who have produced blocks in theast P 100 epochs um yeah what's the genie coefficient of that number uh there's a bunch of kind of arbitrary I keep using the word arbitrary arbitrary parameters here that we set um I don't I wouldn't say we're 80% of the way on our goal to decentralization I'd say we're 10% um so we need to improve this but this is all open source and this is just like a conversation starter I wanted to share this is I'll do another talk on this just this topic in the future all right so let's talk a little bit about why decentralization is hard right that was the subject title of this talk first reason home miners every other blockchain network layer two layer one whatever that I'm aware of has one thing in common and that is that they have professional node operators okay they're experienced people people who understand devops they're running Linux right in in real data centers space mesh has none of this okay our whole goal and the only way we're going to get to a million nodes a million miners in the network is by making it really easy for everyday people to download an app and just minimize it run it set it and forget it um but that presents a whole bunch of obstacles and challenges it you know you need to keep your computer running 24 hours a day that's not that easy if again it's not a Linux VM running in a data center it's your actual gaming computer right um there's High resource requirements which we'll talk about in a moment there's frequent upgrades things like this home users unsophisticated users are not great node operators um our mitigation here is that we need to improve the ux of running a node a lot we're a bunch of backend systems Engineers we're not product people but we are hiring product Talent onto the team and we're going to make this like a really fun experience for our users peer-to peer let's talk about this um again every blockchain Network that I know of runs on peer-to-peer but peer-to-peer is much easier when those nodes again are you know things like Linux VMS and data centers rather than Windows machines sitting on people's home connections um unfortunately I know a lot more about routing protocols on the internet than I ever wanted to know about this is one of the biggest challenges we have um but one of the fun facts I've learned in the past year or two is that the internet the modern internet is not designed to make peer-to-peer easy right there's all sorts of like dark design patterns and hostile design patterns the way that isps run uh the way they put in some cases tens hundreds of thousands of of of nodes and home internet connections behind a single IP address which kind of fundamentally breaks things like TCP IP the majority of nodes in our Network are not routable they're not dialable um it's it's challenging but the good news is that we have some really wicked low-level P2P Network Engineers on our team and I we know how to improve these things there's amazing Technologies we use Li PTP lib ptp's got some fantastic protocols to do whole punching things like this so this will get better resource requirements again we want it to be as easy as possible for any ordinary person to just run our software on their computer without worrying about their CPU their memory their bandwidth Etc um unfortunately the requirements are pretty high right now um the way the protocol is designed right now every time a new minor joins the network the complexity goes up right so it grows linearly in the number of miners each minor submits a proof each Epoch every node in the network has to download validate those proofs that's very bandwidth heavy it's very computation heavy and this is one of the reasons it's become very hard to join the space mesh Network as a unsophisticated home minor we expected best case scenario 10,000 Miners and as I said we have more than two orders of magnitude more than that right now this is our biggest challenge right now we know how to fix this as well we are working on protocol improvements like um distributed verification which means that each node will only be responsible for probabilistically verifying a tiny fraction of those proofs rather than all of them but that's core protocol research and development and that will take time so it will be months until I think this burden uh gets easier uh okay everyone's favorite topic pools again I could easily talk for an hour about this it is very complicated very nuanced uh I think I'm thinking I need to kind of do a talk on this as well it's a very controversial Topic in the space mesh Community um another fun fact I learned recently the first Bitcoin pool launched in anybody first Bitcoin mining pool what year any guesses 2010 crazy early okay pools have always been a reality for every protocol that I'm aware of every permissionless decentralized protocol Bitcoin ethereum proof of work proof of stake it's a thing in in space mesh as well the unique thing about space mesh is that I said many times our goal is to make mining easy for home miners in proof of work if you have one Bitcoin Asic one modern Bitcoin mining machine you will produce a block on average every 30 years okay that means you have to join a pool in order to earn any rewards in proof of stake in AET work like ethereum you need 32 e to run your own validator that's a lot of money that that's Out Of Reach for 99.9% of people on Earth so again you have to join a staking pool our goal with space mesh is decentralization democratization um every minor even the smallest minor receives at least one reward every Epoch every two weeks right let me say that again this is kind of the most important fact about space mesh the smallest minor on the network receives a reward every two weeks at least every two weeks uh that's guaranteed by the protocol it's deterministic it's not probabilistic and that's why space mesh took so long to build um but there are non-economic reasons to join pools as well this is something else we kind of didn't forecast such as the things I was explaining the fact that it's just a little hard to run a node from home right now our mitigation here I mean pools are not going to go away um we need to educate people about the trade-offs the risks of using pools such as giving away your private keys not a great idea um we need to make it easier to operate pools so there are many of them rather than a single pool and we need to reduce those non-economic incentives to join pools so we're working on this I think this will get better over time and finally economics um so that small home user the the minimum storage space right now is 256 gbt right if you have that minimum committed to the network you're earning about $1 us per Epoch that's every two weeks $2 a month if you have one terabyte you're earning about $8 a month right now of course this there's a lot of variables here as more miners join it goes down the coin price obviously moves around a lot I highlighted this this is a screenshot from a mining pool called H9 um per petabyte so per unit space it's as you can see like Chia is in the order of $9 per pedite space m is 358 this is after that crazy exponential growth you saw so it's it's come down a lot but it has at least another order magnitude to fall so that we can expect economically again depending on things like how the market does uh the storage size of the network to at least 10x if not 15 20 30X Etc before we expect the market to find equilibrium which means those rewards will continue to fall um there's nothing we can do about this there's no mitigation this is just kind of Economics um but uh it's going to be exciting to see where we go from here okay um let's go into kind of the final section which is how we do better so I've already covered all of these things just reiterating we have kind of these lwh hanging fruit changes things we know we need to improve to make space m more decentralized we need to improve the product experience improve P2P make it easier to run a node from home to sync a node that's a bit hard right now um we have this set of protocol upgrades that will have the effect of reducing those resource requirements making it much much much easier to run a node and we think then also reducing the impact of pools in the network um we want to again encourage to we know pools are not going to go away right space mesh is a permissionless protocol we can't shut down pools even if we want to but I think a better approach I I kind of like the uh approach the ethereum ecosystem is taking to me right with things like meev boost um we know we can't shut down meev so the best thing to do is democratize access to it increase education Etc I think we can do this with with um with pools in in space mesh and as I said before education education education people need to understand the risks and the trade-offs of things like joining a pool the next step is the VM again could spend an hour here talking about this I'm a VM engineer this is the piece of space mesh I'm the most excited about we have been working on this for years for 5 years we built a full-featured web assembly based VM called space mesh VM or svm we threw it away because the goal posts have changed and the landscape has just evolved so much over the past few years this is just a kind of a random snapshot of like the things you need to be thinking about right now if you want to design a smart contract engine or a VM in 2014 uh there's a lot of stuff to figure out there's different VM choices there's modular versus integrated there's layer ones and layer twos and rollups and different types of rollups and different types of sequencers um we're there's there's ZK considerations where we want to be future proof we are having a research Summit in a few days um and we expect to I can't promise we're going to make a final decision but I hope we will we're going to have um a lot to share on this very very soon but this is the the next phase of space mesh once we've kind of stabilized things with those earlier improvements I listed and finally going back to the very first thing I said right the whole point of this why we're here in the first place is to build these Global decentralized permissionless credibly neutral networks where Ordinary People anywhere in the world regardless of you know the extent to which you're enfranchised by existing Financial networks can can join create value generate value distribute value globally right it's you can think of it as decentralizing um social media networks that we have today right removing those those uh rent seeking centralized operators um I've always seen BL blockchain as a platform for this and I think this is my set of criteria uh something this is a I just kind of came up with this I don't know if it's perfect let me know what you think I'll write and speak more about this but number one the thing we're measuring the thing we're proving should be widely endowed right that means it should be a thing that is accessible to humans everywhere right number two it has to be objectively measurable and measurable in a decentralized way why because the protocol is the only thing that can do that in incredibly neutral fashion as opposed to having human Arbiters which is again the whole point of decentralization ideally it's not too gameable by machines this is really a touchy subject today with the rapid advances in Ai and ideally it's the extent to which it's subject to those economies of scale that I mentioned is is not too severe so actually hold on before I before I jump to the to that uh how to proof of work proof of stake proof of humanity things like that do here not super well right so proof of work proof of stake are not super widely endowed there's not that many people who have 32 eth or Bitcoin as6 um those things are pretty gameable uh proof of humanity there's a number of these projects I won't name any specifically the biggest issue is them is they're not necessarily measurable um in an objective and decentralized way if you're scanning eyeballs or something like there's no way for a protocol to to do that in a decentralized way yet maybe someday we'll have it with zero knowledge proofs um I think we can do better so as I've meditated on this for many many years the thing I've come up with is this idea and I'm just painting a very broad picture here I I think of it as proof of creativity I'm not the first person to come up with this idea there's new foundation and other projects have been working on this and have gotten further than I have but I do think this is the goal right I think that the thing that checks all these boxes as much as possible is creativity right it's very widely endowed there are creative people talented people all over the world you don't necessarily need an education a credit card a bank account to to be creative um it is not immediately obvious how it's subjective measurable but if you think about the way social media networks have tweets and likes and warps and all this kind of fun stuff um I think that is a fair starting point for objective measurement this next one is very controversial I'm going to go out on a limb and say that I don't think that AI tools uh are able to compete with humans at this moment nor will they be in the near future in real creativity uh controversial topic I'd love to have that debate uh I could be wrong but I have a good feeling about about that um real art writing things like this and uh to what extent is creativity subject to economies of scale I mean this is debatable but I think it's less subject to it than than some of the existing protocols like proof of work and stuff so um that's all I've got thank you so much for your attention I I have a vision that you know we'll build space mesh into a proof of creativity Network so stay tuned we have our work cut out for us but that's where we're going um some resources so uh I mentioned a substack I write about all these topics including space mesh in long form so check that out um on Twitter I'm very much a boomer but I'm now on warcast as well shout out to warcast there's a uh there's a Meetup later today I think right um and yeah SPAC smashh website and join our Discord chat. spaco I hang out there I'm going to leave you with a quote from one of my personal Heroes jiren lineer anyone who's not familiar with jiren he has a number of excellent books on this topic he is the greatest thinker in the world who has thought deeply about the dangers of centralized social media centralized networks centralized technology um this is our goal is to solve this problem thank you