Base(d) AI House: A Celebration of AI Innovation on Base
On November 11th, Magnet Labs proudly hosted its fourth major event, Base(d) AI House, during the highly anticipated Devcon Blockchain Week in Bangkok. This milestone event was made possible with the support of Base, Masa, Virtuals Protocol, and Pond, along with contributions from Coinbase CDP team, Payman and Failsafe.
Designed as a platform for discovery, collaboration, and innovation, Base(d) AI House provided a vibrant environment where attendees could:
-
Explore cutting-edge AI-powered projects built on Base.
-
Learn from top-tier builders in the ecosystem.
-
Gain strategic insights from thought leaders during discussions.
-
Engage directly with the dynamic AI and Base communities, fostering connections and collaborations.
The entire event was hosted by Lude Trappe, co-founder of Magnet Labs, kicking off with the Keynote session.
Keynote
The keynote sessions offered an immersive glimpse into the forefront of AI innovation on Base, including live demos from Magnet Labs, and trailblazing presentations from leading projects in the ecosystem.
Magnet Labs: Live Demos of Ardio and Meme Blindbox
Luke Fan, CEO and Co-founder, captivated the audience with live demonstrations of two groundbreaking innovations:
Ardio
Ardio, the first Action Agent developed by Magnet Labs, is redefining blockchain usability by enabling natural language commands to execute on-chain tasks. Users can transfer tokens, interact with protocols, and navigate blockchain systems simply by chatting with Ardio, eliminating the complexities of managing multiple decentralized applications (DApps).
This user-centric approach not only streamlines blockchain interaction but also significantly lowers the barriers to entry for newcomers in the crypto space.
Meme Blind Box (MBB)
Meme Blind Box empowers users to create memes effortlessly by interacting directly with the MBB Twitter account. The process is collaborative: users can define key token parameters such as name, icon, and description, while the AI determines attributes like total token supply, allocations for contributors and liquidity pools.
The result is a meme that evolves dynamically based on community feedback, culminating in a unique and creative expression. The product’s expansion to Farcaster and Discord is already underway, promising even broader reach and engagement.
Virtuals Protocol
Jansen Teng, CEO and Co-founder, explored whether on-chain AI agents are merely a passing trend or the future of decentralized applications.
Virtuals Protocol is poised to impact various sectors by enabling:
-
Gaming: Creating dynamic characters that adapt based on player interactions.
-
Metaverse Experiences: Offering personalized virtual companions that evolve through user engagement.
-
Digital Art: Facilitating unique interactions with AI-driven art pieces that respond to viewer input.
Pond
Dylan Zhang, CEO and Co-founder, presented Pond’s innovative crypto-native AI models, demonstrating how they leverage on-chain data to train AI agents tailored for blockchain applications.
Key Features of Pond's AI Models:
-
On-Chain Analysis: Processes and simplifies complex blockchain data streams into actionable insights, enabling clear understanding of on-chain interactions.
-
Predictive Intelligence: Leverages advanced algorithms to forecast market movements and identify potential security threats in blockchain environments.
-
DeFi Applications: Enables dynamic fee optimization for liquidity providers, enhances security through threat prediction, and delivers personalized NFT recommendations.
Masa
Brendan Playford, Co-founder, introduced a conversational AI agent powered by real-time Twitter data. This agent utilizes the Masa protocol to access and process live data from Twitter, supporting various applications including sentiment analysis and trend monitoring.
Using Masa’s data scraping tools, he showcased how users could create a Twitter meme agent that deploys actions based on live Twitter insights.
Watch its demo: https://youtu.be/Sb4mKDgbaKQ?si=-QvSMao20CNU74Dp
Failsafe
Ari Medvinsky, CTO, demonstrated Audit AI, an advanced generative AI-powered tool for auditing smart contracts in real-time. This tool represents a leap forward in enhancing the security of Solidity-based blockchain projects.
Key Features of Audit AI:
-
Real-time Detection: Facilitates continuous monitoring and instant vulnerability identification in smart contracts, replacing traditional time-consuming audit processes.
-
Cost-effective Security: Provides unlimited automated auditing capabilities at a fraction of manual audit costs, making comprehensive security accessible to all projects.
-
Quick Fix Guidance: Delivers instant, actionable remediation suggestions when vulnerabilities are detected, enabling developers to implement security fixes efficiently.
Coinbase CDP
Lincoln Murr, Product & Onchain AI, concluded the keynote with a presentation on building AI agents using AgentKit, an essential toolkit for developers creating AI-driven blockchain solutions.
Key Features of AgentKit:
-
Autonomous AI Agents: Facilitates the creation of AI agents that can operate independently on the blockchain.
-
Integration with Web2 APIs: Combines traditional web API functionalities with blockchain data, enhancing the versatility of the agents.
-
Support for Multiple LLMs: Works with various large language model providers, including OpenAI and Claude, for natural language processing capabilities.
-
Extensible Design: Offers a template structure for developers to add custom functionalities easily.
Fireside Chat with Jesse Pollak and Luke Fan
Key Insights from Jesse, the #001 builder of Base:
-
Convergence of AI and On-Chain Technologies
Jesse highlighted the simultaneous growth of AI (like LLMs and agents) and on-chain systems (decentralized economies). He emphasized their complementary nature, where AI drives participation and efficiency, and on-chain systems ensure authenticity and transparency.
-
Evolution of the Global Economy
Jesse compared the current shift to on-chain systems to the internet’s transformation of communication. However, he noted that while the internet centralized financial and governance systems, on-chain aims to create borderless, inclusive economies.
-
Regional Pain Points
Jesse illustrated the need for on-chain solutions by referencing regional challenges:
Kenya: Limited creator rewards and trust in platforms.
Canada: Disconnected financial systems from global economies.
India: Advanced payment systems but restricted access to global capital.
-
Lowering Entry Barriers
He stressed the importance of making blockchain accessible without requiring extensive technical knowledge. Simpler engagement drives exponential adoption.
-
AI Agents as Catalysts
Jesse described personal use cases, such as managing Base NFTs, to showcase how AI simplifies interactions with on-chain assets.
-
Call to Action for Builders
Jesse urged builders to tackle real-world problems with creativity, using AI to unlock the full potential of a decentralized economy.
Panel
Panel l: Agents Unleashed: A New AI Economy on Base
Moderator: Mattie Biniek, Magnet Labs
Panelists:
-
Luke Fan (Magnet Labs)
-
Lincoln Murr (Coinbase CDP)
-
Jansen Teng (Virtuals Protocol)
Discussion Highlights:
-
The Agentic Web
Panelists explained the concept of the "Agentic Web," as outlined in Coinbase’s report, focusing on AI agents leveraging crypto infrastructure to drive economic growth. They highlighted Base’s thriving ecosystem for agent development.
-
Magnet’s Action Agents
Luke broke down what Magnet’s Action Agents are all about, and for those interested, what are the ways users can start engaging with these agents on Base.
-
Coinbase CDP Tools
Lincoln outlined how tools like the Smart Wallet, MPC Wallet, and AgentKit lower the barrier for developers to create on-chain solutions, fostering innovation in the Agentic Web.
-
Memecoin vs. Initial Agent Offering
Jansen described Virtuals Protocol’s unique approach of combining memecoin launchpads with IAOs, emphasizing the potential for co-owned AI and innovative token models.
Panel ll: Infrastructure for AI Agents: Exploring Support Across Different Layers
Moderator: Mattie Biniek, Magnet Labs
Panelists:
-
Calanthia (Masa)
-
Tyllen Bicakcic (Payman)
-
Dylan Zhang (Pond)
Discussion Highlights:
-
Infrastructure Layers for AI Agents
Drawing from Coinbase’s Crypto x AI Landscape report, panelists discussed key layers critical for AI agent infrastructure: Payments (Payman), Data (Masa), and Middleware (Pond).
-
Real-World Payments
Tyllen shared how Payman builds infrastructure enabling AI economies to operate in real-world scenarios, leveraging Base’s tools for scalability and reliability.
-
Real-Time Data Integration
Calanthia explained Masa’s focus on combining diverse data sources like Twitter to power conversational agents. He stressed the need for precise PMF (Product-Market Fit) to identify impactful real-time data use cases.
-
On-Chain LLMs
Dylan debated the feasibility of hosting LLMs on-chain, emphasizing use cases where decentralized AI models bring unique advantages, such as data ownership and trustless interactions.
Closing
The Base(d) AI House showcased the immense potential of AI and blockchain convergence, sparking inspiration and collaboration among builders, developers, and entrepreneurs. From visionary keynote to the insights shared by industry leaders during panel discussions, the event underscored the transformative opportunities at the intersection of these technologies.
As AI-powered agents and robust infrastructure layers continue to evolve, the path toward a decentralized, inclusive, and transparent global economy becomes clearer. Builders are encouraged to leverage their expertise, tackle real-world challenges, and contribute to this groundbreaking shift.
Together, we are not just imagining the future of the on-chain economy—we are building it!
Resources
To learn more about us, you can check our Website, follow us on X, join our Discord, Telegram, and read more on Blog.
Try our first Action Agent Ardio Alpha.
🧲 About Magnet Labs
We’re building the infrastructure for an AI-powered blockchain experience. Key Innovations:
-
Programmable Action Agent: Provides programmable, upgradeable, and composable AI user interfaces for a variety of crypto scenarios.
-
ActionVM: Provides an off-chain execution environment, supporting dynamic loading of action sets and compatibility with various LLMs.
-
Creator Hub: Offers a collaborative playground where users can build and refine agents or actions using no-code/low-code tools. In return they can share revenue based on their contributions.