🎉🎉Happy to announce IOSG OFR-Paris was successfully held on July 18th during ETHCC!! Thanks for all your interest and attendance!!
🗼OFR-Paris, organized by IOSG, brought together visionaries, experts, and enthusiasts from the worlds of cryptocurrency and artificial intelligence. This extraordinary gathering aimed to foster collaboration, generate innovative ideas, and catalyze the intersection of these transformative technologies.
This time, 1300+ friends registered for the event through Luma, and over 1000+ friends came to the venue and spent a wonderful time with us! We also opened YouTube Live and 10000+ audiences participate in the OFR-Paris online! Attendees from around the globe, ranging from crypto enthusiasts and AI experts to investors and entrepreneurs, converged to explore the remarkable confluence of crypto and AI. The event was a melting pot of diverse perspectives, igniting stimulating discussions and paving the way for new collaborations with 10 topic speeches and 3 panel discussions!
Some Highlights of Each Session 👀
Talks🧑🏻🏫
IOSG Thesis: Intersection of AI & Crypto — Ishanee Nagpurkar — Principle of IOSG Ventures
Artificial Intelligence (AI) has the potential to make applications more intelligent. However, one of the challenges lies in the cost and scalability of AI models. For instance, ZKML (Zero-Knowledge Machine Learning) is still expensive and unable to support large models. While ZKML does offer proof of inference, it lacks proof of validity for the models it employs. This limitation raises concerns about the reliability of the outcomes. Another obstacle in the development of AI applications is the difficulty, expense, and slow pace of training models on-chain. Overcoming these challenges is essential to unlock the full potential of AI and creating more intelligent and efficient applications.
zkML via zkLLVM and its applications to Ethereum — Misha Komarov — Founder of Nil Foundation
Building zk-enabled apps is a challenging task due to various factors. Firstly, the manual construction of these apps requires a significant amount of time and effort to learn and implement the necessary techniques. Additionally, ensuring their long-term maintenance can be problematic, as they are often complex and might need more existing code that can be easily reused, especially when it comes to proving cryptography, machine learning, or gaming functionalities. However, there is hope on the horizon, as an EVM (Ethereum Virtual Machine) Endpoint is being developed, allowing zkProofs to be directly requested from EVM applications. Moreover, this new system will enable payments to provers on the EVM and offer additional verification methods to revalidate proofs, potentially easing the burden and expanding the possibilities for zk-enabled applications in the future.
The ZkDataCenter — Omer Shlomovits — CEO of Ingonyama
ZKDC holds significant potential for various applications and networks both today and in the future. Presently, it offers a way to engage with networks like Filecoin Sealing (L1s — Throughput) and ZKRU prover (L2’s — Latency) or Apps, including ZKML, which facilitates trustless audits of ML-powered applications and verification of specific ML model execution, along with the Axiom node, an Ethereum ZK co-processor. Looking ahead, ZKDC’s utility is expected to grow even further. Tomorrow, it could be employed as Prover-as-a-Service for private smart contracts. The broad range of potential applications highlights ZKDC’s versatility and its promising role in advancing privacy and security within various domains.
AI-Driven Smart Contract in Trustless — Marvin Tong — Co-Founder of Phala Network
AI offers immense potential for increasing efficiency and boosting productivity across various domains. Integrating AI technologies into processes can streamline tasks, optimize resource allocation, and improve decision-making. On the other hand, ensuring transparency, accountability, ownership, and community economic empowerment is equally essential. Web3, the next generation of the internet, plays a crucial role in this pursuit. Rather than being the end of Web3 Agent, autonomy marks its beginning. Web3 emphasizes an open-source culture, where communities with common interests can come together to collaborate and create valuable outputs. Ownership is enforced within the community, driving a focus on output quality, backed by crowdsourcing scores, contribution proof, and the use of knowledge graphs as oracles. This combination of AI-driven efficiency and Web3’s community-oriented approach holds the potential to revolutionize industries and pave the way for a more transparent, equitable, and productive future.
From Zero to Hero: Disrupt Web3 with Your AI-powered Dapp in 3 Steps — Luis Schliesske — Co-Founder of Gelato
Bringing AI on-chain introduces a groundbreaking capability for smart contracts, enabling them to make decisions or predictions using real-time off-chain data. This breakthrough unlocks a myriad of exciting new use cases, including AI-optimized and automated DeFi strategies, dynamic content generation through AI, and personalized user personas driven by AI analysis. One platform that empowers developers to harness the potential of AI within their smart contracts is Gelato. By integrating Gelato, builders can enhance their smart contracts, ensuring they are reliable, user-friendly, and scalable, all without compromising decentralization and censorship resistance. The presence of middleware layer services, such as automation, account abstraction, and real-world data and computation, further enriches the capabilities of AI-powered Dapps. The process of building AI-powered Dapps on Gelato is simplified into two straightforward steps: writing the code and deploying it on the platform, making it accessible and feasible for developers to embrace the world of AI-driven decentralized applications.
Intro to Hyperchains: Meet the ZK Stack Powering the Internet of Value — Anthony Rose — SVP of Technology of zkSync
The ZkSync Era aims to enhance Ethereum’s scalability by shifting transaction execution to Layer 2 (L2). With the inclusion of validity proofs, ZkSync Era ensures the correctness of L2 execution, offering a fast, versatile execution environment with minimal transaction fees. Supporting languages like Solidity, it facilitates smoother adoption by developers. ZkSync values key characteristics such as trustlessness, security, reliability, censorship resistance, privacy, hyper scalability, accessibility, and sovereignty, highlighting its commitment to providing a robust and decentralized solution.
The ZK Stack boasts several compelling features that contribute to its effectiveness. First, it is free and accessible under permissive MIT/Apache open-source licenses, encouraging community engagement and contributions. Secondly, its composable nature allows Hyperchains built with the ZK Stack to seamlessly connect within a trustless network, enabling low latency and shared liquidity. Moreover, its modular architecture permits users to customize and configure various aspects of their Hyperchain, tailoring it to their specific needs. Being battle-tested with Era processing approximately 25 million transactions per month, the ZK Stack demonstrates a robust security approach, inheriting the results of a security-first mentality. Lastly, the ZK Stack is future-proof, recognizing ZK rollups as the future of scaling Ethereum and emphasizing the importance of laying a strong foundation from the outset to harness the full potential of ZK’s capabilities.
QED: Universal Proof Systems for AI/ML and Blockchain — Grigore Roșu — President and CEO of Runtime Verification
The QED protocol is currently in progress, and it focuses on remote computing. It aims to achieve correctness and performance simultaneously in a unique manner. The core innovation of QED revolves around reducing the verification of computational claims, both VC and FV, to the verification of mathematical proofs using a universal proof checker implemented as a custom parallel ZK circuit.
QED effectively addresses two critical challenges at the same time: firstly, it inherently produces proof of correctness through its usage, and secondly, it remains language and VM-independent, ensuring broader applicability and compatibility. The power behind QED lies in three fundamental technologies it leverages:
1. K Framework — Universality: This technology provides mathematical proofs for both program execution and formal verification, enhancing the protocol’s versatility and usefulness.
2. Matching Logic — Correctness: A minimal logic approach that captures both program execution and ZK cryptography aspects, ensuring correctness in the protocol’s operations.
3. ZK Cryptography — Succinctness/Scalability: Recent advancements in Zero-Knowledge cryptography have enabled massively parallel processing, contributing to the protocol’s succinctness and scalability.
Through the integration of these critical technologies, the QED protocol aims to achieve efficient and secure remote computation with an emphasis on correctness and performance, while remaining flexible and independent of specific programming languages and VMs.
Identity of Real People In a World of AI and Bots — Kenny Li — Co-Founder of Manta Network
Proving “real-ness” in a secure and reliable manner is crucial because existing methods have proven inadequate. Centralized data breaches remain a significant concern, as they expose sensitive information, such as emails, passwords, and addresses, to potential theft and misuse.
On-chain identity offers a more secure alternative to centralized solutions, but it still possesses some inherent flaws that need to be addressed. To achieve a more robust and privacy-preserving on-chain identity system, zk-enabled on-chain identity with SBTs comes into play. This decentralized and self-sovereign approach ensures that individuals have full control over their identity data, removing the reliance on central authorities. Furthermore, zk-enabled on-chain identity with SBTs offers untraceable yet verifiable identity attributes, allowing users to prove their identity without revealing unnecessary information. Additionally, the non-transferrable nature of this system ensures that identity data remains secure and cannot be easily manipulated or misused by unauthorized parties.
By adopting zk-enabled on-chain identity with SBTs, we can pave the way for a more secure, private, and reliable identity verification method that protects user data and mitigates the risks associated with centralized data breaches and flawed on-chain identity solutions.
Search and Recommender Systems in a World of Generative AI — Prompt Engineer — Co-Founder of Onaji
Machine learning played a crucial role in advancing Web2 platforms, but its integration into Web3 platforms has been relatively limited thus far. Onaji emerges as a comprehensive Full-Stack Machine Learning and Data Science Solution specifically designed for Web3 platforms, enabling capabilities like Recommendation and Search and User Behavior Analytics.
In the context of Web3, various inputs can be utilized, such as Wallet data, NFT information, ERC20 tokens, and user Actions, yielding corresponding outputs. For instance, Onaji’s ML capabilities can be employed to develop applications like a Personalized ETH Token Recommender or a tool to Find Wallets with Similar Preferences, enhancing user experiences and interactions within the Web3 ecosystem.
Additionally, a new product called “Hailey” has been introduced, although its specific functionalities and features are not detailed in this context. It is likely that Hailey would leverage Onaji’s machine learning and data science capabilities to cater to the needs of Web3 platforms and users in innovative ways.
Unleashing the Power of Smart Contracts — Raullen Chai — Co-founder and CEO of loTex
The landscape of artificial intelligence is witnessing the rise of an Emerging Centralized AI War. In response to this, the concept of DePIN has emerged as a means of democratizing AI warfare by decentralizing power and control. By leveraging DePIN, a decentralized AI force can be harnessed, which has the potential to unlock a vast trillion-dollar market and create limitless opportunities in the field of AI.
The proposed solution to achieve this is an off-chain computational infrastructure designed to convert real-world data streams into verifiable, smart contract-ready proofs. This infrastructure would allow AI processes to be executed off-chain, validating the outcomes through smart contracts on the blockchain. By enabling real-world data to be integrated with secure and tamper-proof smart contracts, DePIN can provide a powerful and decentralized framework for AI warfare.
In summary, the emergence of centralized AI warfare has prompted the need for decentralized solutions like DePIN, which can unleash the potential of a decentralized AI force, leading to new economic possibilities and opportunities in the trillion-dollar AI market. The proposed off-chain computational infrastructure plays a key role in achieving this vision by ensuring the verifiability and security of AI-driven processes through smart contracts.
Panels🎤
Panel One: Delving into ZKML
Jason Morton, CEO at Zkonduit: Computational constraints are the biggest bottleneck in the short term.
In the rapidly evolving landscape of technology, computational constraints have emerged as the most significant bottleneck in the short term. With the ever-increasing complexity and scale of data-driven applications, our current computational resources often struggle to meet the escalating demands for processing power and memory. This limitation impacts various sectors, from artificial intelligence and machine learning to big data analytics, hindering their full potential and stifling innovation. Industries such as healthcare, finance, and scientific research face delays in implementing cutting-edge solutions due to these constraints. Overcoming this challenge requires substantial investment in developing more powerful and energy-efficient hardware, optimization techniques, and scalable algorithms, ultimately unleashing a world of possibilities for technological advancement and societal progress.
DC Builder, Research Engineer at Worldcoin: ZK needs more use cases for users.
The concept of zero-knowledge (ZK) is an intriguing and promising area in cryptography, but its broader adoption requires a compelling array of use cases to entice users. While ZK protocols offer the remarkable capability of validating information without revealing sensitive data, their practical application has been somewhat limited. To accelerate ZK’s mainstream acceptance, researchers, developers, and industry leaders must collaboratively explore and innovate new use cases that showcase the technology’s strengths. This could include applications in secure authentication, privacy-preserving data sharing, confidential transactions, and decentralized finance. By diversifying and expanding the use cases, ZK can gain traction across various domains, empowering users with enhanced security and privacy while revolutionizing digital interactions.
Fran Algaba, Co-founder at Giza: Need to provide tools that are compatible with existing tools to attract developers to participate in ZKML.
To effectively attract developers to participate in Zero-Knowledge Markup Language (ZKML) adoption, it is imperative to offer tools that seamlessly integrate with existing frameworks and technologies. Developers are more likely to embrace and contribute to the ZKML ecosystem if they can leverage familiar tools and workflows. Therefore, the development of compatible toolsets, libraries, and plugins becomes crucial in lowering the barrier to entry for new adopters. By aligning ZKML with popular programming languages, software development environments, and blockchain platforms, developers can readily harness the benefits of zero-knowledge proofs and privacy-preserving functionalities. Additionally, providing comprehensive documentation and accessible resources will facilitate the learning curve, encouraging developers to explore the potential of ZKML in enhancing data security, fostering trust, and enabling privacy-aware applications across various industries.
Daniel Shorr, Co-founder & CEO at Modulus Labs: Faster is cheaper that is why we work on speed too much!
In Modulus Labs, the core target of their work in ZK technology revolves around the belief that “faster is cheaper.” They are deeply committed to pushing the boundaries of speed and efficiency in the realm of zero-knowledge technology. Focusing on accelerating the performance of their solutions, Modulus Labs aims to deliver cost-effective outcomes for their clients and partners. Speed optimization is ingrained in every aspect of their research, development, and implementation processes, as they recognize its pivotal role in driving productivity and reducing operational costs. However, they understand the importance of striking a balance, ensuring that speed improvements do not compromise the robustness, security, or reliability of their ZK products. Modulus Labs’ relentless pursuit of faster, yet dependable solutions positions them as a trusted leader in the ZK technology landscape, empowering businesses to thrive and unlock new possibilities in a competitive digital era.
Panel Two: Hyper Scaling Verifiable Computation to Unlock New Application
Alex Gluchowski, Co-founder at zkSync: We start with open source, try to decentralized sequencer, and then prover! Blockchain is a continuation of the Internet!
zkSync, starting with an open-source approach, embarked on a journey to achieve a decentralized sequencer, and subsequently, a prover. As pioneers in the field of blockchain technology, their vision extends beyond mere innovation; they see blockchain as a seamless continuation of the Internet’s transformative power. Embracing the ethos of openness and collaboration, zkSync’s commitment to open-source principles fosters community involvement and accelerates the evolution of decentralized solutions. Their quest for a decentralized sequencer underscores the drive to distribute authority and control, promoting a trustless and secure environment. By advancing towards a robust prover, they aim to harness the full potential of zero-knowledge proofs, revolutionizing data privacy and transactional scalability. zkSync’s resolute belief in blockchain’s potential as an extension of the Internet heralds an era of decentralized applications and a democratized digital landscape.
Eli Ben Sasson, Co-founder at StarkWare: OP rollups are mostly based on authority and can’t scale so they are not a long-term solution.
StarkWare, a prominent player in the blockchain space, holds the perspective that OP rollups primarily rely on authority and face scalability challenges, making them unsuitable long-term solutions. Through their extensive expertise and research in the field of blockchain scalability, they assert that OP rollups’ dependence on centralized entities for transaction verification poses inherent risks to decentralization and security. StarkWare’s vision aligns with the pursuit of robust and scalable solutions, seeking to overcome these limitations with innovative technologies like ZK rollups. By advocating for zero-knowledge proofs and their application in rollup solutions, StarkWare aims to foster trustless and scalable ecosystems that can meet the demands of future blockchain applications, paving the way for a decentralized and sustainable digital landscape.
Ye Zhang, Co-founder at Scroll: First decentralize the prover, before decentralizing sequencer work on the performance side, also system security is a very important part.
Ye places great emphasis on their work’s core elements. Their approach centers on prioritizing the decentralization of the prover as a foundational step. Recognizing that a decentralized prover is crucial to ensuring trustless and tamper-resistant transactions, they diligently focus on harnessing the power of decentralization. Additionally, Ye is resolute in addressing performance-related challenges, understanding that optimizing system performance is essential for achieving scalability and user adoption. Committed to ensuring the utmost security, they invest significant efforts in fortifying their systems against potential threats and vulnerabilities. By diligently incorporating these three pillars — decentralizing the prover, optimizing performance, and strengthening system security — Ye endeavors to revolutionize the blockchain landscape, enabling a future of secure, efficient, and truly decentralized blockchain applications.
Alexandre Belling, Prover Team Lead at Linea: The real bottleneck is sending data to L1, then decentralizing the sequencer.
Identifying the critical challenges in the blockchain domain, the team acknowledges that the real bottleneck lies in the process of sending data to Layer 1 (L1). Streamlining this data transfer is imperative for enhancing overall system efficiency and throughput. Addressing this concern, they are actively devising innovative solutions to expedite and optimize data transmission, enabling seamless interaction between Layer 2 (L2) and L1. Moreover, the team also recognizes the significance of decentralizing the sequencer, an essential element in the transaction validation process. By distributing the sequencing responsibility, they aim to foster a trustless and secure ecosystem, empowering users with enhanced control and reducing reliance on central authorities. Through their steadfast commitment to overcoming these bottlenecks, the team envisions a future where blockchain technology can achieve widespread adoption, scalability, and true decentralization.
Brian Retford, Co-founder at Risc0: There is room for both roll-up ideas. Verifiable computation can be an essential part of future systems.
Within the blockchain community, experts recognize that there is ample space to explore both roll-up concepts. The prevailing sentiment is that verifiable computation holds immense potential to play a crucial role in shaping future systems. As advocates of technological progress, they believe that embracing diverse approaches is vital for achieving comprehensive solutions to complex challenges. While roll-ups offer compelling scalability benefits and efficiency gains, the integration of verifiable computation can significantly enhance data privacy, security, and transparency. This forward-thinking perspective motivates researchers and developers to actively explore the synergy between these ideas, envisioning a future where blockchain systems seamlessly incorporate verifiable computation to revolutionize various industries and empower users with unparalleled levels of trust and confidence in the digital landscape.
Panel Three: Unleashing the Crypto-AI Nexus: Empowering AI Agents of Taming Their Power
Ben Fieldin, Co-founder at Gensyn: We can put agents on the chain. The concentration of power in small groups of companies is dangerous and the future we want to see is a model which everyone can run.
Gensyn firmly believes in the potential of placing agents on the chain as a means to achieve a more decentralized and inclusive model. They recognize the inherent dangers of power concentration within small groups of companies in the blockchain space. To counter this, Gensyn envisions a future where their technology enables a model that is accessible and operable by everyone. By empowering a broader community to run agents on the chain, they seek to create a more democratic and transparent ecosystem. Their mission revolves around fostering a level playing field where individuals and entities from diverse backgrounds can actively participate in the blockchain network, contributing to its growth and governance. Gensyn’s commitment to this vision drives its relentless pursuit of innovative solutions that uphold the principles of true decentralization, autonomy, and open participation, paving the way for a more equitable and collaborative digital future.
Marijo Radman, Co-founder at Solity Network: Decentralizing text-stack is the most important thing for improvement and better adaptation. Agents can help with security and automated trading.
Decentralizing the text-stack stands out as the paramount objective for achieving improvement and enhanced adaptation in the blockchain domain. By distributing the text-stack across a decentralized network, the technology can become more resilient, transparent, and censorship-resistant, fostering a trustless ecosystem. This approach empowers a broader community to participate in network governance and ensures that no single entity can wield undue control over the information flow. Additionally, employing agents within the decentralized text-stack can prove invaluable in enhancing security measures and facilitating automated trading processes. Agents play a crucial role in verifying and validating transactions, bolstering the overall security of the system. Moreover, they can automate trading activities, streamlining the execution of transactions and reducing reliance on manual interventions. By embracing both decentralization and agent-driven automation, the blockchain ecosystem can evolve towards a more secure, efficient, and user-centric landscape, paving the way for broader adoption and transformative impact across various industries.
David Minarsch, Co-founder at Autonloas: LLM allows you to do more experimentation, but it is still an early stage.
LLM, or the Large Language Model, opens up new possibilities for experimentation and innovation, providing researchers, developers, and users with a powerful tool to explore various applications and use cases. Its vast capabilities in natural language understanding and generation enable users to conduct a wide range of experiments in language processing, text generation, and other linguistic tasks. However, it is essential to acknowledge that LLM technology is still in its early stages of development. As with any emerging technology, there are ongoing efforts to refine and optimize its performance, address limitations, and mitigate potential biases. Collaborative research and development efforts are crucial to unlocking the full potential of LLMs and ensuring they evolve responsibly and ethically. Despite being in its nascent phase, LLM holds immense promise, and as advancements continue, it has the potential to revolutionize how we interact with language and information in the digital era.
At Last👩💻
As the curtains fall on this landmark event, we stand at the precipice of a transformative era. The convergence of crypto and AI holds boundless potential, and we are excited to witness the groundbreaking developments that lie ahead. IOSG remains committed to fostering innovation, supporting cutting-edge projects, and driving forward the symbiotic relationship between crypto and AI.
OFR-Paris not only celebrated the intersection of crypto and AI but also provided a fertile ground for collaboration, exploration, and inspiration. Together, we will continue to forge new paths, break barriers, and create a future where these technologies empower individuals and revolutionize industries.
Thanks again for your support for IOSG and OFR Paris!
We are looking forward to meeting all our friends at the next event!
About IOSG 👇
IOSG ventures started investing in crypto in 2017 and have been an early investor in each major vertical of Web 3.0, making early bets on L1s such as NEAR and Polkadot, and spending most of the resources in recent years in the broader Ethereum ecosystem, from DeFi (0x, 1inch, Synthetix, MakerDAO, Liquity), over gaming (Illuvium, Big Time) and social (Galxe, Roll, CyberConnect) to Ethereum scaling (Arbitrum, StarkWare) and privacy solutions (Aztec, Sismo).
