Exploring the New Al paradigm in the Crypto Era
Date: June 28, 2024, 8 AM PT
Co-Host:@BSCNews、@OdailyChina、@JinseFinance
Speakers:
@ViabtcCapital、@SpaceandTimeDB、@AlloraNetwork、@ionet
The recent session of the SPACE meeting was held on June 28, 2024, focusing on exploring the new paradigm of artificial intelligence in the crypto era. Co-hosted by BSCNews, OdailyChina, and JinseFinance, this session brought together experts and leaders from ViaBTC Capital, Space and Time,io.net, and Allora Network, moderated by Chung, an investment partner from ArkStream Capital.
The meeting centered on the latest advancements in AI and blockchain technology. Speakers introduced their innovative projects, covering topics such as decentralized data warehouses, zero-knowledge (ZK) proofs, decentralized GPU cloud networks, and the integration of AI and blockchain. There was a consensus that addressing real-world problems and providing user-friendly solutions are crucial for driving adoption and innovation.
The discussion also looked forward to future developments. io.net plans to launch a staking mechanism to enhance network stability, while Space and Time is expanding its data indexing capabilities and optimizing network design.
In summary, this SPACE meeting offered valuable insights into the intersection of AI and blockchain, showcasing innovative solutions and emphasizing the importance of meeting actual user needs. Participants shared a vision of leveraging technology to create more secure, accessible, and user-friendly applications.
Chung Host | Investment Partner of ArkStream Capital
Let’s get started. It’s an honour to have io.net , space and time, viabtc capital, and allora with us today for a roundtable discussion. I’d like to invite them to briefly introduce themselves and their projects.
Stephen | Head of Solutions from Space and Time
Hello everyone, I’m Stephen, the Head of Solutions at Space and Time. With a background in both Web3 and big data, I’ve spent years working in the Bay Area with various vendors and consulting firms. At Space and Time, we’ve merged big data capabilities with the Web3 world. Our platform features a decentralized data warehouse and zero-knowledge (ZK) proofs, enabling tamper-proof queries. This allows smart contracts to query large datasets and get verified answers on-chain.
Essentially, Space and Time acts as a data-centric ZK co-processor, focusing on data rather than arbitrary compute. We support integrations with AI and allow users to submit SQL or natural language queries. This makes it easy to ask complex questions about blockchain data, such as total liquidity for an exchange or token values across different chains.
In summary, Space and Time expands blockchain use cases and makes data more accessible, both for projects and consumers interested in blockchain analytics.There were no significant errors in your original text, but I’ve added the emphasis as requested to highlight the key points.
Asa | APAC Head at io.net
Thank you for having me. I’m excited to be here. At io.net, we operate a GPUDP network that connects idle GPUs from independent cloud providers, crypto miners, and even personal workstations. We provide these resources for AI training and inference, rendering, ZK proving, and crypto validation networks — any use case requiring GPUs.GPUs are becoming the new strategic asset, essential for future technological advancements. We’re aiming to be the Airbnb or Saudi Arabia of GPUs, capitalizing on their growing importance.
As the APAC lead at io.net, I focus on growing our ecosystem, business, community, and media presence. We’re proud to have Upstream as an early investor and value our strong relationship with them. It’s great to be here today.I’ve highlighted the key points as requested. The original text was correct, and I’ve emphasized the critical aspects related to io.net‘s operations, strategic goals, and your role as APAC lead.
Jeff | Chief Analyst from ViaBTC Capital
Absolutely, gmgm! It’s a pleasure to join this Twitter Space and discuss AI projects in the market. I’m from viabtc Capital, the capital arm of viabtc Group, the world’s top pre-IPO mining firm.On the capital side, we manage a $100 million portfolio that includes over 50 blockchain projects across sectors like DeFi, infrastructure, and AI. We also act as LPs in major funds like Pantera and Shimmer, actively expanding our presence in key ecosystems like Bitcoin, Solana, and Sui.I’m excited to be part of this session and look forward to exploring more about the AI paradigm.
Chung Host | Investment Partner of ArkStream Capital
What are the specific technologies or application areas covered by your AI project, and what user groups do your products mainly target?
Stephen | Head of Solutions from Space and Time
One of the core innovations at Space and Time is our ZK proof technology, particularly our “Proof of SQL.” This tamper-proof database enables secure, large-scale data queries, maintaining blockchain’s trust model.We recently open-sourced Proof of SQL and our verifier, collaborating with partners to build a robust verification network. Blitzar, our GPU accelerator for ZK-SNARKs, has also been open-sourced, drastically improving proving times. For example, our prover can handle under 100,000 records in sub-second time and up to a billion rows (about a terabyte of data) in around 60 seconds.
Developers can access our code and demos on GitHub at Space and Time Labs. Proof of SQL supports various data sources, including CSV files, and ensures tamper detection through verification failure.In addition to ZK proofs, we’ve integrated AI into our platform to simplify complex queries, enabling business users to ask questions in natural language, which our system translates into SQL.These advancements in ZK and AI enhance data security and accessibility, creating new opportunities for businesses and developers.
Asa | APAC Head at io.net
I’m excited to discuss our tech stack at Io, which has undergone extensive trials and iterations recently. We’re pioneering a decentralized GPU cloud, a novel concept bridging Web2 and Web3 technologies.Currently, we use a centralized server to manage the complexity of hundreds of thousands of devices sending status signals, but our goal is progressive decentralization, reducing reliance on central servers over time.
While our blockchain use is currently limited to payments and Io coin, we aim to democratize GPU access, traditionally controlled by major providers like Nvidia or AWS. This mirrors DeFi’s role in simplifying financial access.Despite challenges like speed disparities with centralized data centers, Io is well-suited for AI compute needs, especially for tasks like training smaller models and inference, which dominate AI compute demand.
Looking forward, we anticipate AI algorithms evolving to address GPU shortages, emphasizing edge compute and adaptable models suited to varied hardware.We’re confident in our ability to evolve alongside the ecosystem and effectively utilize compute resources. This journey is exciting, and I’m thrilled to share our progress.
Chung Host | Investment Partner of ArkStream Capital
And so for via BTC as a venture capital, I also, I’m curious about What areas do you think there are pain points that need to be addressed by AI? What are your suggestions for the combination of AI and blockchain?
Jeff | Chief Analyst from ViaBTC Capital
When integrating AI and crypto, three primary scenarios emerge: computational power tokenization, model tokenization, and data tokenization.Computational power tokenization utilizes distributed networks for GPU resources and AI model training, as seen with projects like Alldone. Model and agent tokenization facilitates deployment and utilization in decentralized environments, with agents driving consumer-centric products and adoption. Data tokenization involves querying and utilizing resources for training large models, akin to Space and Time’s approach.Emerging subfields in AI and crypto include oracles, ZK-ML, co-processors, generative arts, and gaming applications. However, current AI applications in crypto are limited, driven largely by capital operations and FOMO sentiments.
Addressing pain points in AI integration with blockchain is crucial. Many users find technologies like SQL daunting, necessitating lower entry barriers. Popular AI applications like GPT-4 and ChatGPT lack inherent crypto integration, with no widely recognized AI apps centered on crypto.For practical use, users often opt for centralized solutions like iCloud or Google Cloud for storage and Apple Pay or Google Pay for payments, highlighting complexity in adopting crypto. Innovations like Solana’s new product, enabling on-chain transactions via Twitter, show promise but lack one-click simplicity.
While many projects aim to replicate existing services with crypto, practical and user-friendly solutions solving real issues are paramount. Focus should be on effective solutions rather than mere token promotion.In summary, integrating AI and blockchain must prioritize practical applications that surpass Web2 solutions and attract real users.
Chung Host | Investment Partner of ArkStream Capital
And I’ve also notice space and time recently released an open-source Verifier. Could you explain in detail the functions and significance of this Verifier?
Stephen | Head of Solutions from Space and Time
Space and Time has recently open-sourced both their proof of SQL and verifier, which is significant for their network users.Space and Time indexes blockchain data, offering tamper-proof queries beneficial for both Web3 and some Web2 enterprises. For example, developers integrating application data with blockchain data can use Space and Time. Their smart contracts call a relayer to execute queries, receive results, and obtain proofs. These proofs are then verified to ensure data integrity.
The verifier code supports multiple environments: Space and Time’s network, an API for streamlined operations, and an ongoing development of an on-chain verifier. This flexibility allows applications like Web3 games or centralized exchanges to verify data independently using the open-source verifier code.
Collaborations with companies such as Chainlink and Horizon Labs aim to bolster the verifier network’s robustness. Space and Time’s decentralized network of compute clusters addresses specific constraints but requires significant capital for node operations. They are working to streamline operations and create a lightweight verifier network, potentially accessible on platforms like ZK Sync or Polygon.Their goal is to democratize verification by making it affordable and accessible, fostering broad participation and robust verification processes.
Chung Host | Investment Partner of ArkStream Capital
And for io.net or for Asa, You’ve previously mentioned io.net and its GPU cluster or GPU Cloud solution to address the global shortage of GPUs. I’m interested in learning more about the details of these innovations and the proposed solutions. Could you share some insights? Additionally, how do these innovations impact retail or the listeners?
Asa | APAC Head at io.net
Connecting different nodes and integrating GPU resources has been a challenging feat for Io.net. We’ve faced issues such as fake card attacks in Season 1 and stability problems like login failures and machine drops in Season 2. Despite these challenges, the Io engineering team has been relentless in making continuous improvements, gaining recognition even from critics.Instead of delving into technical details still in flux, let’s focus on the opportunity. Despite the global GPU shortage, many independent data centers utilize only 20–30% of their GPUs, leaving the majority idle due to structural asymmetries. Large cloud providers like Google and Amazon dominate with extensive marketing and sales teams, leaving smaller centers underutilized.
This gap presents a significant opportunity for Io.net to enter the market as a viable alternative, both in Web3 and Web2. We aim to represent smaller GPU providers, handling their marketing and sales while they focus on operations. Our goal is for Io.net to be considered when major providers cannot meet users’ needs in terms of cost, speed, or availability.Initially, we’re collaborating with professional-grade data centers with responsive technical teams to power the first batch of AI compute on Io.net. We’re approaching this challenge from both technical and business perspectives, targeting the most accessible opportunities.
Lastly, we’re rolling out staking in the next 30 days to enhance reliability. This crypto-staking mechanism requires nodes to stake tokens, incentivizing network stability by slashing stakes for unreliable participants. We’re addressing this comprehensively with a focus on business strategy, staking mechanisms, and ongoing technical enhancements
Chung Host | Investment Partner of ArkStream Capital
And for viabtc, I wanted to also know what are the investment theories or philosophies that you base your decision making or in to invest AI on? And then mindful of the time, also wanted to know if you’ve got anything to share with the audience in terms of how to pick some AI picks.
Jeff | Chief Analyst from ViaBTC Capital
At a high level, a significant portion of existing AI projects seems focused on hype and speculation rather than genuinely solving real-world problems. Many integrate AI tools or connect to APIs without offering truly innovative or creative solutions. Projects prioritizing token prices over product development are likely to lose users to newer, more exciting opportunities in the near future.Reflecting on the crypto market’s history, we’ve seen several cycles, like the first bull market from 2017 to 2018 and the second from 2020 to 2021, each attracting waves of new investors. In this third bull market, investors are more educated. Similar to traditional equities markets, only projects solving real problems can sustain their price and grow long term. This principle applies equally to AI investments in the crypto space.
Looking ahead, the next major breakthrough in the crypto market is likely to be an AI-driven consumer application capable of onboarding millions of users. However, such a killer app has yet to emerge. Founders should focus on solving genuine problems and real use cases. While crypto technology offers solutions, it shouldn’t be forced into projects that don’t genuinely benefit from it. Many market solutions in crypto claim to need blockchain or tokens but may not require them in practice, often fueled by hype and speculation.Regarding investment criteria, we emphasize team strength, past experience, and technical expertise. Recently, there’s been much debate around tokenomics, including circulating supply, vesting schedules, and valuations. Evaluating these factors is crucial, especially as crypto project valuations have soared in recent fundraising rounds, sometimes reaching hundreds of millions or even billions of dollars before token generation events, leading to inflated valuations.
In conclusion, while speculation plays a role in crypto markets, it’s not the sole determinant of success. The key criterion remains whether a project effectively addresses real use cases. We’re particularly focused on practical utilization of AI and blockchain technology. Lately, many promising teams and products lack genuine need for crypto in their solutions. This critical issue should be considered by founders and investors alike, irrespective of whether the solution belongs to Web2 or Web3 domains.
Chung Host | Investment Partner of ArkStream Capital
Understand. Absolutely. So just to wrap up, and I know asa, you need to go, like probably each of us can use maybe 1 minutes to share. Are there any interesting thing that audience should be excited about in the upcoming maybe 1 to 2 months about our io.net and also space and time is if you want, you can go first. I think you’ve mentioned as well.
Asa | APAC Head at io.net
We’re planning to roll out staking in the next 30 days, expected to be a bullish event for our token. Staking will require each node and GPU to stake at least 100 Io tokens to participate in the network, ensuring greater stability. Nodes will face slashing if they exhibit unreliable behavior, reinforcing network integrity.
Additionally, we’re actively supporting community projects such as Io token-based DeFi initiatives and tokenizations that enable individual investors to purchase compute power as NFTs on Io. Furthermore, we’re exploring merge mining opportunities with other AI projects to incentivize the Io network with their native tokens. This cross-pollination aims to enhance our ecosystem’s vibrancy and visibility.If you’re involved in an ecosystem project or considering one, we invite you to reach out. We’re eager to collaborate and discuss how we can contribute to the growth and success of innovative projects in the AI and blockchain space. Thank you for the opportunity to share these updates with you all.
Chung Host | Investment Partner of ArkStream Capital
Understand. Thanks a lot. And so basically, more utility for the token, more ecosystem, let’s supply kind of because of the sticking and token bullish basically on the token price. And so Steven, do you have anything exciting for the users in the upcoming 1 to 2 months?
Stephen | Head of Solutions from Space and Time
We’re actively expanding our data indexing to include new chains like Arbitrum, Optimism, and others currently in progress. This expansion will significantly enhance the breadth and depth of data available on our network.Moreover, we’re in the process of redesigning certain parts of our network, with plans to roll out these improvements soon.There are also exciting developments on the horizon that I can’t disclose just yet, but expect more announcements and updates over the next six months. Stay tuned for more details in the coming months!
ArkStream Capital is a venture capital firm specializing in early-stage investments in Web3 unicorns.
Founded by crypto experts with pedigrees from MIT, Stanford, Tencent, Google, and BlackRock, ArkStream leverages eight years of deep Web3 expertise to drive the zero-to-one growth of its portfolio companies.
ArkStream Capital is managing a portfolio of over 100 companies, including Aave, Flow, Sei, Manta, Fhenix, Merlin, Particle Network, and Space and Time.
