3/31/23 Discord AMA Summary
Summary of Discord audio AMA, March 31st, 2023.
Space and Time participants:
Scott Dykstra, CTO and Co-Founder
Mike Post, Community Manager
Thank you to all of the community members who attended!
[Session begins]
Introduction
Mike: Welcome, all of our community members around the world, to another Space and Time AMA. We're super glad to see you all here. It's been couple of months since we did one of these. First, as always, I want to thank our Community Helpers, NFTerraX, Mystery Men, S. Crypto and Leon, for all the work you guys do in the community. We absolutely appreciate it. And also to everybody else who goes above and beyond to make this community what it is. And, finally, to every one of you for being part of this wonderful community that we've built over the last eight months. Let's get rolling. It's my pleasure today to introduce to you Scott Dykstra, CTO and Co-Founder of Space and Time. Scott, it's a pleasure to be here with you. I'm sure that everybody here pretty much knows you already. But in case there are some here who don't, why don't you give us a quick intro?
Scott: Absolutely. Well, thanks for having me on, Mike. And I'm super excited to be here. The community is definitely popping off. And Mike, you've done a fantastic job so far, getting everyone organized and going in this community. And, you know, kind of keeping the lights on, man. It's been really great working with you. And I appreciate all the folks that have been so active and submitting questions. And awesome memes, as well, that's important. I'm Scott Dykstra, CTO and Co-Founder of Space and Time, and I've been leading the whole technology strategy of Space and Time from day one. We started this back in 2021 and really got the ball rolling in 2022. It's been a really fun journey over the last 18 months or so, getting Space and Time to a point where we can actually launch it to the public and do a beta launch, which is coming in about a month. My background: long history in Web2 Fortune 500, data warehousing. Gross! And I am so, so happy and feel so very, very fortunate and blessed to be in Web3, in a field of world-changing technology. I worry that if I would have blinked, I would have missed it. I feel like a lot of people out there are blinking right now and missing it. And I'm really excited to be here with you all.
AMA
Mike: All right, fantastic. So, I just want to let you all know, we may not have time to get to everybody's questions today. We will try to answer those elsewhere, the ones we don't get to. So, a couple of quick questions. These are the two most common ones we got. First off, Scott, why don't you talk a little bit about how we're going to be onboarding node operators.
Scott: Absolutely. So, node operations is going to be an interesting process throughout this year. I'm going to give a little bit of context. Space and Time is a decentralized data warehouse. And what that means is that we are a network of community owned and community operated servers around the globe, that are all contributing via consensus to a global network of compute, available to run queries. And those queries are run on behalf of smart contracts, those queries are run on behalf of dapp developers, data scientists, business analysts, the client can really be anyone. And throughout the year, we are going to be onboarding node operators of all shapes and sizes, first in a permissioned way with larger node operations, the larger Web3 node operators of the world that we're already partnered with. And then later in the year, we'll be going permissionless, and kind of opening things up to the community, so anybody can run a Space and Time node. I expect by end of year, we'll have a pretty large swath of node operators, both large companies that do this full time and then community members who downloaded binaries and set up a Kubernetes cluster and got going. Last thing I'll say here is Space and Time is a Kubernetes cluster. You can run a single server, but we'd prefer multiple servers for high performance, high availability, redundancy, and extra memory available to do large joins. Some of our dapp developers and our smart contract developers that are actually executing these queries require a decent amount of memory and a decent amount of compute because they're actually asking really tough questions of a lot of data.
Mike: So, the second most common question, and we get this one a lot, too, so let's get it out of the way right away: when token?
Scott: Yeah... when token? We certainly intend to drop a token next year. What we're doing is we're spending this whole year bringing on node operators and decentralizing and deploying Proof of SQL on-chain. Proof of SQL is a novel ZK-proof for SQL that we invented, that allows us to not only generate SQL results but also prove them on-chain. Now, we were hoping to drop a token this year, but given all the instability, with the U.S. political approach to crypto, I think we're going to wait it out, weather the storm, and we intend to drop a token next year. And a lot of prep is going to go into that, a lot of big partnerships are going to go into that along the way, a massive node operator network is going to go into that along the way. And then we'll launch.
Mike: Yeah, the uncertain regulatory environment has definitely played a role in token launches across the space. It's been interesting to see.
Scott: I think a lot of big networks are launching their token this year. I've got a bunch of friends over at Sui Network, for example, which is launching a token this year, and a number of others. And it'll be interesting to see how things play out as these networks launch their token. We're going to keep an eye on it, and we're going to watch how they do it and make sure that when we do it, we do it right, based on that data.
Mike: Let's move on to where we are right now with the Space and Time platform. There have been really a lot of changes lately. Where are we now in terms of the development of the Space and Time platform?
Scott: Yeah, we've been moving and shaking, man, we've been moving fast on this thing. So we launched kind of a private closed alpha, to a number of partners and customers, in December. A lot of those actually came from the Chainlink BUILD program. Shout-out Chainlink BUILD. And that was a really eye-opening experience, launching a private alpha of Space and Time back in in December. We did some bug fixes, we are working on some more stability, working on some additional capabilities that we know that dapp developers and smart contract developers are going to need. And the beta release, at the end of April, beginning of May, the beta release is going to be a pretty massive improvement on the Space and Time platform. Let me give some examples. We will have self-service, kind of one-click deployments of the Space and Time platform. We'll have a bunch of capabilities with OpenAI, Mike, that you and I can talk about in a minute. We're going to natively integrate Chainlink functions, so that dapp developers can write queries in Space and Time and then put those query results on-chain quickly with Chainlink Functions. We'll have a beautiful UI powered by OpenAI to explore blockchain data, find blockchain data, ask questions about it, query it, you name it. And then really badass dashboards, beautiful dashboards for major protocols. So if you're building a protocol, or you're doing analytics on a protocol, you can go into Space and Time dashboards and build branded, beautiful dashboards for specific protocols. But what I'm most excited about is Proof of SQL. The Proof of SQL team has been moving fast and numbers are looking good, performance is looking good. And so we're going launch open beta at the end of April, but then move very quickly after that toward a testnet of Proof of SQL.
Mike: Awesome. So, for those who aren't familiar with really what's pretty revolutionary technology of Proof of SQL, why don't you give us a bit of a brief about what Proof of SQL does?
Scott: Yeah, absolutely. This is a lot of where Space and Time's R&D is going. Yes, Space and Time is fully decentralized. Yes, Space and Time is a cluster-scale data warehouse technology that can process tens of terabytes of data per cluster in a very high-performance way. But that's not really what's getting the Web3 community most excited. What's getting the Web3 communities most excited is a novel ZK-proof, a new cryptographic proof for SQL. Why SQL? Well, because the world runs on SQL. Every developer I know has a SQL database. I don't think SQL is better than other programming languages. In fact, it's probably way goofier than most programming languages. SQL is a goofy programming language, but the world's business runs on it. And so it's a way for you to aggregate large volumes of data, a way to filter, sort, join, do arithmetic, a way to take large volumes of data and create dashboards very quickly. And OpenAI is really, really good at writing SQL for you. The future is not even in people writing their own SQL; the future is just an OpenAI writing it for them. Or any large language model, for that matter. The days of people manually typing out SQL are over. And that's a good thing. So what we built is a novel ZK-circuit, a cryptographic circuit, that proves that not only have we executed that SQL correctly, but that the underlying data hasn't been tampered with. That's really important for DeFi, for lending, for gaming, for any kind of smart contract that's going to be securing a lot of value on-chain. Let me give a few examples. You want to stream in terabytes of TradFi data from off-chain, TradFi data from traditional real estate appreciation, inflation markets, bond commodity markets, option futures, you name it. You can stream all that data into Space and Time, land it in Space and Time, and secure it in Space and Time. Then, whenever needed, a smart contract could just directly query Space and Time and say, "Hey, show me an aggregation of U.S. dollar inflation rates over the last month, or show me aggregated real estate prices across 18 different markets, or show me the options price of a Tesla stock today on-chain, but I need you to aggregate it over the last month." You can't put terabytes of data on-chain. You can't even put gigabytes of data on-chain. You can barely put megabytes of data on-chain. The chain is not a data storage platform. The chain is a place to transfer value. The data that goes on-chain is all about transfer of value. What about the terabytes of data off-chain? Think about a video game that's generating tons of real-time events that it needs to store, process, analyze, etc. That video game is going to stream their gaming server event logs into Space and Time. And then their smart contract can just periodically ask Space and Time, "Hey, who won the game? What was their score? Were they gold-tier ranked? How many hours of playtime has the player done? And what kind of reward should we mint them on-chain?" If you want to remit awards to the winners of an Esports tournament, how do you get that data from the servers for that Esports tournament over to your smart contract? We're not talking about one little data point; we're talking about millions of gamers playing, winning, earning, and Space and Time can aggregate all that information and make it available to a smart contract when needed. Finally, another example would be lending. Say you're building a cross-chain lending protocol, and you want to aggregate interest rates from a bunch of other chains and a bunch of other lending protocols like Aave, Compound, MakerDAO, and so on. Space and Time can capture all those rates, liquidations, and data about investor risk, trader risk, and provide those to smart contracts on any major chain when needed. For example, a smart contract on Arbitrum can say, "Hey, Space and Time. What's Scott's risk score? And should I give him an over-collateralized loan on-chain or under-collateralized loan on-chain? If I give him an under-collateralized loan, what are the odds of liquidation? What is my risk here?" Those risk calculations can be done off-chain, stored in Space and Time, and then we provide that to smart contracts on any major chain in a tamperproof, trustless way via Proof of SQL. I'm sorry for the longwinded winded response there, for taking extra time to walk through Proof of SQL; I'm just really excited about it. This is hard work that our team is doing to build a new cryptographic circuit, a new ZK-proof, that proves that the data in Space and Time hasn't been manipulated, hasn't been touched, hasn't been changed, and that the SQL operations we do on that data to hand a smart contract what it needs have been done accurately.
Mike: Absolutely. I mean, rightfully so that you're excited. Proof of SQL is huge. And beyond just the technical implications, we have the philosophical implications on the side of Web3: trustlessness. This really frees everybody from the need to trust anybody in terms of connecting data to their smart contracts. You no longer need to trust your data warehouse, no matter who you are.
Scott: The naive implementation that I'm seeing across Web3 today without Proof of SQL is that so many protocols, even the protocols you use on a daily basis, the Uniswaps of the world, for example, are, surprisingly, storing off-chain data in, like, a centralized PostgreSQL database. Now, they're not necessarily connecting that PostgreSQL database directly to their smart contracts, but sometimes they are. So they're storing all this data off-chain. So when you're getting quoted prices on Uniswap, it's actually getting quoted from a centralized PostgreSQL database, and then when you go to trade, you're trading through a smart contract. So many protocols do this. A lot of derivatives protocols and options trading, they do options price calculations off-chain, and they just store that in a PostgreSQL database. The problem with that is that a data engineer, an employee of that company that built that protocol, can just log into that PostgreSQL database, and they can just change a few rows of data and mint themselves new prices on-chain, or they can change the price of options or loan collateralization ratios or scores in an Esports tournament. Or they can change, for example, the price of Tesla stock before it's put on-chain. The problem with centralized databases is not so much that external bad actors are going to come in. The problem is that anybody who deployed that database can go and just change things willy-nilly and then mint premium on-chain that they didn't earn. It's a setup that requires a lot of trust. And trust is a bad word in this case. You're trusting the folks who deployed that centralized database to not be a bad actor, and to not do those things. Yet we see all the time in DeFi major hacks that came in through an oracle network or a backdoor or a bridge to a smart contract where the folks that deployed that bridge were involved. Space and Time helps to solve this by ensuring that the database is truly tamperproof, is truly cryptographically secure, cryptographically guaranteed that you can't go in and just change data later. If you load in the the price history of Tesla's stock into Space and Time, you can guarantee that when you provide that price to a smart contract that wants to, for example, know a prediction market on Tesla's stock price, you can ensure that you're providing them an accurate, cryptographically guaranteed price.
Mike: Proof of SQL is definitely going to change a lot in the world of Web3. It's crazy that right now protocols are using centralized databases and data warehouses in decentralized finance. I look forward to the operation of truly decentralized protocols using Space and Time. So, let's move on to something that you mentioned before, which is AI. We all love ChatGPT. You mentioned also that it can even write code and that that's really the future of of writing SQL, which would have been unthinkable not long ago. But let's talk Space and Time in terms of OpenAI. How is Space and Time using large language models on blockchain data?
Scott: Yeah, I'm glad you brought this up. Proof of SQL is my favorite topic, and ChatGPT is probably my second favorite topic, or GPT-4, the model underneath. So, I've got a lot of thoughts here. So feel free to jump in and stop me if I'm getting too longwinded. This is an exciting topic and an exciting area of research for Space and Time. Given our heavy partnership with Microsoft, and our backing by M12, Microsoft's venture arm, last year, we're obviously in very deep reach and partnership with Microsoft around unique ways in which we can use GPT-4 to change the way people think about data on-chain. I'll start with SQL. As I said earlier, the future is not that you type out your own SQL statements. SQL is kind of a wiry, goofy language. And even me, I'm not good at writing SQL. Anybody who is good at writing SQL, you know, to me, they're a god. SQL is a difficult language! But GPT-4 does a really good job at writing it. So, the future is just asking GPT-4 for what you need. For example, "Hey, show me daily active wallets involved in a cross-chain messaging protocol on-chain. Show me many average gas prices on Sui today. Show me the number of wallets that have at least minted two of this NFT and have a gamerscore greater than 10,000 in this game on-chain. Hey, show me what's happening with this bridge, show me the inflows and outflows of of tokens traded across this bridge to Arbitrum." And OpenAI will write those queries for you and execute them. And then once you like the output, once you're confident that it wrote the query correctly, then you can use Chainlink to put that data on-chain and say, "Hey, I want this query to run every day, and I want you to publish the query results to the chain."
From this point forward, humanity has changed. It's changed for the better, I hope. Web3 has changed humanity by giving humanity a decentralized railroad for finance. It solves the challenges we have in regional currencies, and builds a globally decentralized, community-owned and operated, real-time currency, That's badass. On top of that currency, DeFi rails were built, lending protocols were built, microtransactions and video games, you name it. Now, OpenAI has further changed humanity. From this point forward, the human-computer interface will always be a large language model. From this point forward, the interface between human and computer will be AI. When a major business decision is made by a board of directors, they're going to tell their AI and their AI is going to log that for them. When a user writes a poem, OpenAI is going to help them write that poem. When a user builds a DeFi protocol, OpenAI is going to audit their smart contract for them. The interface between human and computer from this point forward will be a large language model. And that's a massive shift in the human timeline. The invention of the computer was something that connected man with machine. The invention of the internet was something that connected man with man across the globe, a global way to communicate and share information and share value. And then Web3, the decentralized internet, was a chance for mankind across the globe to share value in a way that's removed resistance and censorship and centralized control. Now you add AI on to that, and things get out of control, hopefully in a good way. So what I see happening is that our understanding of what's happening on-chain is about to change. Space and Time in partnership with Microsoft is building really, really forward-thinking ways to add context to the data on-chain, add context to queries that are written that are being sent to the chain, add context to the inflows and outflows of capital between major blockchains.
Mike: It's definitely exciting stuff. In terms of the future of AI, the possibilities are limitless.
Scott: Practically, what that means for Space and Time is that when we launch our open beta, we're going to have a chatbot in Space and Time. That chatbot will allow you to fully control Space and Time. Build a dashboard with a chatbot, write a query with your chatbot, build a Python data pipeline to ingest data industries in time with the chatbot. Connect to Chainlink Functions, and use Chainlink Functions to put data on-chain with a chatbot. That's exciting to me. And that's just the first step. This is just the beginning.
Mike: You mentioned dashboards, so let's move back to the beta. And I know we're all looking forward to the point at which the community can be trawling through data, building dashboards, analyzing data. Why don't you tell us a little bit about what the beta is going to look like and how ordinary, non-enterprise users can participate?
Scott: Yeah, great question. What I'm most excited about is finally getting the community involved in a very real way, in terms of the Space and Time user experience. We're building a UX that anybody, whether they're technical or not, can get on, ask questions about blockchain data, build data pipelines, access the data they need, build dashboards quickly, and explore data. And I expect the community to really rally around the indexed data that we already have loaded into Space and Time and start building dashboards quickly with the help of OpenAI. If you want to build a dashboard about Aave liquidations, you don't really have to write too much SQL. Theoretically, you could do it completely without writing any SQL. You can just ask for the chatbot's help. I think the first couple of weeks of open beta are going to be a little insane. I think that the demand is going to be high and is probably going to be demand that we're not prepared for. And so we might have a little waitlist for a couple of weeks, just so that we can make sure everything is polished and smooth, and the experience is smooth, and we don't onboard, like, 4000 active query writers the first day. If we do, that'd be amazing, but I expect Space and Time to crash if that happens. So I'm going to probably put a little waitlist up for a couple of weeks, and then we'll just remove the waitlist and let everyone go nuts. It's going to be really exciting.
Mike: Absolutely. I'm looking forward to it. I'm looking forward to watching the community be able, as you said, to participate, to be able to come on to discord and say, "Look at this dashboard. I built this. What do you think?" and people working together on those.
Scott: I'm also eventually going to build some Discord bots that connect directly, so that we will be able to use OpenAI to write SQL from within Discord. I don't have a timeline for when those Discord bots will be available, because that's next for me to do after our open beta launch. I'm kind of heads down with the open beta launch. But don't worry, some real alpha: Discord bots are on the way, my friends. As for the beta, I think step one is to just give everyone access, let everyone start playing around with this data, building dashboards, building oracle jobs, you name it. Step two will be some fun competitions, some rewards, some skin in the game for the community, where we can start rewarding folks for building extremely high-quality dashboards.
Mike: I'm super excited for it. Let's move a little bit further into the future here. What do you see the roadmap as after beta?
Scott: Sure, sure. Yeah, I think we'll be in beta for awhile working out the kinks. really ensuring that our smart contract processes and oracle processes are well-audited and secure, improving the performance of Proof of SQL. I mean, think about it: we are building a novel, zero-knowledge circuit that processes terabytes of data. This has never been done before. This is brand-new, greenfield research. And there are infinite ways that we can optimize how that works and improve performance and get it more real-time. If you if you want to prove a ZK-rollup, proving time can be anywhere from 30 seconds to 30 minutes, depending on the type of ZK-rollup you're using and the amount of data or the amount of transactions you're trying to prove. Space and Time does not have 30 seconds to 30 minutes to prove SQL. People are very impatient; they want their SQL processed right away. If my smart contract on-chain says, "Hey, show me all the wallets that recently won this Esports tournaments and their gamerscore," your smart contract is going to wait, like, eight seconds for that result in the worst case. This has to be real-time. And so we've got a lot of work to do to improve the performance of Proof of SQL so that it can process larger and larger and larger datasets in real-time. Today, we can do decent-sized data sets in real-time, but we've got a long ways to go before we can process terabytes in real-time. We'll get there, and we'll get there by end of year, so that next year, we can focus on a token. So that's Proof of SQL, that's the next step after open beta, is a probably a testnet of Proof of SQL where we start bringing out a bunch of DeFi protocols, lending protocols, games, options, trading platforms, really anyone that's been begging us to use Proof of SQL, and we start working with them directly. We're also talking to a number of L1s. We're talking to a couple of different L1s about ways to more natively integrate Proof of SQL. And that's exciting. And so maybe I'll have some alpha there in a month or two.
I see we've got a growing list of folks in the community that have joined and are listening in. If you're listening in, and you want to be more active in the Space and Time community, I would be really thrilled to find more ways and listen to more suggestions of what the community wants to do. Do you guys want to build dashboards for major protocols? Do you guys want to use OpenAI to explore blockchain data, using crazy data visualizations that we build? Space and Time has always been very user-experience centric. We love our design work. We design crazy experiences on the web. Is that what you guys want to see? Do you want us to keep building new experiences for exploring blockchain data? Do you have creative ideas for how we can use large language models to give you new ways to understand what's going on-chain? I'd love to hear from the community. I'd love to hear recommendations of how you want to use this technology. Think about it! Space and Time is indexing every event from every transaction from every block from every major blockchain, and it's there, ready to query, in Space and Time. Then you have this beautiful UI on top to explore that data in very interactive experiences. And then you have a large language model on top of that, so you don't even have to write SQL. So what do you guys want to do as a community? And how do you want to get involved? I'd love to hear thoughts from the community on Discord.
Mike: The "Time" part of the Space and Time name refers to our lightning-fast query speeds. How does Space and Time achieve sub-second query speeds?
Scott: Yeah, that's a great question. I think sub-second queries will be really difficult with Proof of SQL specifically. Right now we're seeing sub-second response times for non-tamperproof queries, queries that aren't secured by Proof of SQL specifically. Because not all queries need Proof of SQL. Proof of SQL requires a lot of compute. If you're going to run Proof of SQL, you're generally running it on queries that are securing value on-chain, queries that are being requested by a smart contract in DeFi, lending, gaming, infra, you name it. So the performance of queries that specifically require Proof of SQL, we're trying to keep those under 10 seconds, you know, one to 10 seconds, depending on the complexity of the query. And queries that don't require Proof of SQL, sub-second if you're asking small questions, regardless of the of the data volume, and maybe anywhere from one to three minutes depending on how complex the question you ask. I've seen some folks write some queries that were so mind-blowingly complicated, things like a 10-table join that runs for three minutes straight on 10 terabytes of data. And that's okay, because it's meant to be an offline query, it's meant to be this massive compute process that runs for minutes and processes a ton of data. I've also seen response times as low as 5 or 10 milliseconds, .05 seconds, on terabytes of data. So it really just depends on the query complexity, the total data volume, if it's a complex query, and if it depends on Proof of SQL. But I'll tell you this: performance is what it's all about. Our team is heads down, optimizing, optimizing, optimizing. We've got to make this run crazy fast, at all costs, because that's what the world's expecting.
Mike: Absolutely. And I know we're working with Nvidia on GPU acceleration. Tell us a little bit about some of the technology that allows us to scale so well.
Scott: Oh, good callout. Yeah, GPUs. Surprisingly, most modern databases don't leverage GPUs. Think about that. Most modern databases and data warehouses do not use GPUs. Why not? Well, first of all, there's latency in actually getting the data loaded onto GPUs. Leveraging GPUs only actually speeds up the query. When that query is doing a lot of computation, it's a very compute-heavy query. Well, the good news for us is that a lot of our queries are compute-heavy. A lot of the queries that I'm seeing that our dapp developers write are compute-heavy. So we work with Nvidia to GPU-accelerate our queries. Our Proof of SQL framework runs on GPUs. Our big database clusters, this global network of decentralized compute and decentralized node operators, are spinning up GPUs for Proof of SQL. So then we thought, "Okay, well, if we already have all these GPUs to power Proof of SQL, will might as well use them to accelerate our query response times as well. We might as well get faster-running queries by leveraging GPUs. So that's what we've done in partnership with Nvidia. The first, as far as I know, GPU-accelerated data warehouse, GPU-accelerated HTAP data warehouse, and the first ZK-proofed data warehouse, all via NVIDIA GPUs.
Mike: It's definitely exciting technology. And if you ask a lot of people and mention Nvidia, they think gaming. We recently announced partnerships with WeMade and with stardust. And, of course, we were at GDC 2023 as well to meet with game developers. So what are those partnerships going to look like? And how is Space and Time going to help to optimize Web3 games?
Scott: Yeah, a little bit of alpha: we've got a ton of gaming partnerships on the way. We're going to be doing a huge booth at the Consensus conference in Austin, Texas, at the end of April with Shrapnel, which is a AAA game that I'm really excited about, a really cool first-person shooter. You mentioned WeMade, you mentioned Stardust. And I think what we're most excited about is allowing gaming developers to understand what in-game activity led to an on-chain purchase. What events in the game led to an NFT mint, what weapons were upgraded, what skins were used, etc. And so that's from an analytic standpoint. Gaming servers stream in terabytes of data into Space and Time, and then we stream in on-chain data to join that with the in-game data. What in-game data can we join with on-chain data? That's from an analytic standpoint. And then the next step is okay, well, once we have all those analytics, how can we actually connect them back to smart contracts to mint rewards for gamers for accolades? If I win an Esports tournament, pay me out, baby, pay me out on-chain. If I rank up to gold tier, where's my NFT at? How do I tell a smart contract what my gamerscore was, the metadata of the events that happened in-game, to appropriately give me a reward on-chain? That's a lot of our research and a lot of our partnership with these larger gaming companies. And it's really exciting to see. Not just analytics, not just understanding what in-game activity is leading to on-chain activity, or vice versa, what on-chain activity is leading to cool events in games, but also then giving rewards to players for specific actions and accolades they earn.
Mike: Yeah, for sure. Speaking of partnerships, let's go back to Microsoft. Space and Time is the first Web3 project invested in by Microsoft. It's a partnership that's generated a ton of interest. Tell us more about that partnership and how we're going to see it in action.
Scott: Microsoft has been a really, really good partner. And I'm seeing all of the major cloud players, Tencent, Google, even AWS, diving headfirst into Web3. I actually think Microsoft was an early mover, although we don't see that in the headlines as often. It's exciting. Really, really encouraging. It makes me really, really optimistic. It's a lot of hopium for me, that the major cloud vendors are all investing heavily in Web3, finally. Now, you could argue from a centralization versus decentralization standpoint, if all these cloud players get into Web3, aren't they going to be deploying centralized systems on top of their clouds? Sure, in some instances, but I've actually been surprised to see that they've embraced the open-source, decentralized community more than I expected. And they're actually contributing public goods to the decentralized network, because they know they have to align to the mantra of Web3 if they're going to be playing in this space. I'm not saying by any means that all of the Web3 code and systems being deployed by Microsoft, AWS, Google, Tencent, are all decentralized. No, that's not what I'm saying. I'm saying they have they have absolutely embraced decentralized Web3 networks as well. From a Microsoft standpoint, they've been a really good partner. It's funny, when we first started talking to Microsoft last year, lot of our Web3-minded folks and advisors and investors and partners said, "Don't do that. You need to be truly decentralized." And I'm like, "You guys don't get it. It's not about decentralized versus centralized. It's about OpenAI. It's about a hypervisor that can offer us a lot of GPUs. It's about GPUs. It's about confidential compute. It's about enclaves. It's about trusted execution environments on Azure. It's about OpenAI on Azure." Like, wake up, the world's moving fast! And surprisingly, Microsoft's leading that charge. And that's really exciting for me to see. Microsoft is by far leading the charge in enclaves and trusted execution environments, TEEs, encrypted VMs. Essentially, encrypted containers. Microsoft is leading the charge on GPUs. I'm seeing massive partnerships between Microsoft and Nvidia to deploy a lot of GPUs. And we need a lot of GPUs. And then finally, most exciting, Microsoft is leading the charge with OpenAI and large language models. And all of those are relevant to blockchain because blockchain data is very hard to understand, very hard to access, very hard to contextualize. When you combine a very powerful decentralized database with a ZK-proof that runs on GPUs on Azure with trusted execution environments to manage private keys on Azure, and OpenAI, large language models on Azure, things get really fun. They've been very supportive of our efforts. And I feel like they're going to be even more supportive as we move forward.
Mike: It's really great to see Microsoft really supporting the development of the trustless web. I'm old enough to remember the antitrust lawsuits. You would have thought back then that Microsoft would do anything to keep things centralized and monopolistic. It's great to see them supporting the development of Web3 infrastructure.
Scott: Seeing the multiple different large cloud vendors, Microsoft, AWS, Google, Tencent, seeing them jump head-first into Web3 is confirmation to me that we're not wasting our time. You know, I think this year was a little bit scary for everyone. Like, "What's going on with Web3? Is this the right space? Is this really going to fulfill the vision we all expect?" Seeing all the major cloud players jump in is just further evidence of that.
Mike: Overall, what criteria does Space and Time use when looking for partners? What sort of qualities do we look for in a partner?
Scott: It's really different across different verticals. With gaming, it's high-quality AAA games that are going to generate a lot of data, or gaming infrastructure platforms that are processing a lot of data that need a cluster-scale data warehouse. For DeFi and lending and on-chain activity and Web3 infrastructure like Chainlink, it's a firm footing in the space. It's high-quality projects that have built a name for themselves and have really, really interesting technology. For example, we're talking with the folks at Sui. I think Sui is really interesting because their technology is badass, right? They're pushing the limits of what an L1 can do. Polygon, with the launch of their zkEVM. This new ZK technology is really exciting for helping to scale Ethereum. You know, we look for L1s and L2s that are building really interesting tech, we look for cloud hypervisors that are willing to support us in our search for GPUs, we look for gaming platforms that are building really great games that are actually fun to play.
Mike: One more question about the beta: is there going to be a fee for using the service? And how is the pricing model going to work in general, from community members to developers to enterprises?
Scott: Yeah, we're going to be pretty supportive of the community. I don't think we're going to charge much, if at all, to community members that want to just get on and run some queries. If you're going to run a ton of queries and run big queries, that's going to take up a lot of compute, and it's going to cost us a lot of money, so we'll probably charge for that. But if you're just building little dashboards, and you're playing around with some blockchain data, and you're asking some questions with OpenAI, you don't really need to be charged at all, in my opinion. Where we make our money is larger Web3 protocols and enterprises and games that want to build production-grade processes. So we'll charge on a cost-per-query basis, you can pay as you go cost-per-query, and we'll charge alternatively on dedicated capacity. You can either just pay per query, or you can prepay for sort of dedicated compute for your project. And think of it like a Chainlink model where you could kind of pay per query on-chain, or like a centralized software-as-a-service model like Snowflake, for example, where you can just pay in arrears with your credit card. We've facilitated both. And I think it's going to be really exciting to see the different ways people use Space and Time and the different commercial models. We're already getting all kinds of requests for new commercial models as well, especially protocols that are building their own marketplaces on top of Space and Time. I think things are going to move really fast.
Mike: Awesome. So let's end here on some questions about the future of Space and Time. Scott, how do you envision the Space and Time platform evolving in the next one, three, five years. What does the future look like?
Scott: I want to answer this question so bad. But there's some alpha I really can't give away too soon. We're seeing a lot of folks eyeing Space and Time. It's a very visible project. And in the past, we've given away alpha a little early and I've seen other data-focused projects sort of copy our style. And that's okay. That's just part of the game. This is part of Web3. But some things I do want to keep a little more close to the chest. But here's some hints, some higher-level alpha for you.
Mike: Just to clarify, I was just talking generally.
Scott: I know, and I'll explain generally, but the real future of Space and Time is hard to explain generally without giving away some really creative ideas. We're going to do some really cool things with OpenAI that change the way people think about blockchain data and think about what's happening on-chain and the way people have access to the blockchain. We're also going to do some things to secure data fed into large language models with blockchain technology. How do you verify that the data fed into large language models came from an accurate source? This is something we've tweeted about and we've blogged about, but we're actually building towards, unlike others. I believe Space and Time will become an actual chain. Not just a decentralized data warehouse; we will be a chain. I'm not saying we'll be an L2 or an L1, or that we'll compete with other chains. We'll be something unique that sort of doesn't really compete with other L2s or L1s, but more complements L2s and L1s. And then long-long term, I think we're going to just take over the database space in general, not just Web3 databases, but just databases in general. We have some really unique technology that applies to Web2, not just Web3. So as we get deeper and deeper and deeper into Web3, and build the Space and Time blockchain, and build crazy capabilities with OpenAI, simultaneously, we can expand into Web2 and say, "Hey, you guys deserve a tamperproof database too, not just Web3."
Conclusion
Mike: So, we're just about coming to the end of our time here. Scott, any parting thoughts?
Scott: I just want to say I really appreciate how active this community is. And I also want to say now that we're so close to launching beta. Post-beta launch, we'll be getting much, much more active with ways the community can actually use Space and Time, not just talk about it. Space and time is for the community. It's the first-ever community owned, community-operated data warehouse. That's the whole point: it's truly decentralized, it's built for the community of Web3. I almost want Space and Time to be a public good of Web3 eventually. And so it's finally time to get this community onboarded to our dapp, exploring our data, and building things on top of Space and Time, and using Space and Time rather than just talking about it. And we'll definitely create fun programs, rewards, incentives, reasons to give back to our community when they do use Space and Time. So I'm really excited about this summer, because there's going to be a lot more to come.
Mike: For sure. Scott, as always, it's an absolute pleasure to have you on here. Thanks for joining us today and answering some questions we got from the community.
Scott: You guys rock. And Mike, great job, man. You're always the best host. Thanks, man. I really appreciate you putting this together, my friend.
Mike: Absolutely. It's my pleasure. Folks, this was Scott Dykstra, CTO and Co-Founder of Space and Time. And I just want to echo what he said, we absolutely love hearing from you guys. So really, anything, anytime you want to reach out, feel free to contact me. And we're just really looking forward to the future of the community as we move on, and watching the community really get to start using Space and Time. It's going to be great. Thank you all for coming here and being with us today. We'll see you next time.
[Session ends]
Space and Time Links:
Website: http://spaceandtime.io
Twitter: http://twitter.com/spaceandtimeDB
Discord: http://discord.com/spaceandtimeDB
Telegram: http://t.me/spaceandtimeDB
LinkedIn: http://linkedin.com/space-and-time-db
Last updated