Skip to main content

What Did You Get Done This Week 3

Link: https://x.com/ai16zdao/status/1862609655509176778

Timestamps​

  • 00:03:40 - Meeting Start & Guidelines
  • 00:03:56 - Jin's Update: Automated Discord Summarization & Airdrops (Llama, LangChain)
  • 00:08:24 - Choosing Participants (Lots of developer updates follow)
  • 00:09:43 - Stargarden: AI for DAO Management & Lore-Keeping
  • 00:10:46 - Boya: Duplicate Message Bug, Gamifying Stats, Intent Analysis
  • 00:21:24 - Reality Spiral: Coinbase Integrations, GitHub Adapter, Code Coverage, Base Launch
  • 00:27:58 - W3Tester: FlyDumpMoney - AI Payments Without Wallet Addresses
  • 00:31:03 - HashWarlock: TEE Plugin for Secure Wallet Generation
  • 00:33:18 - Soto: Bosso with Multiple Personalities & Animation
  • 00:35:21 - Mitch: Dark Sun - AI Investigates Binary Solar System
  • 00:36:57 - Nick Parallel: 3D Environments, Robotic Arms, 3D Agent Models
  • 00:42:41 - Beige: BlockRat - AI Agent Live Streams Minecraft & Answers Questions
  • 00:44:45 - Robin: Apollo - AI Agent for Health & Wellbeing (with User Data)
  • 00:47:47 - Eve: Data Set Grader, AI16z Integration, Multi-Agent Systems
  • 00:51:14 - Oguz: Eliza Interface, Modularization, Compatibility with Non-Crypto
  • 00:55:40 - Swarm: AI Agent Hosting Platform, Real-Time Voice
  • 01:01:29 - RektDin: Rogue Agent - Joe Rogan-Inspired Podcast AI with Multiple Characters
  • 01:04:08 - Roperito: "Life Engine" for Realistic AI Actions, Deployed "God" AI
  • 01:07:56 - Lothbrok: AI for Logistics (Real-Time Data for Shipping & Pricing)
  • 01:13:29 - Clark Riswold: Positivity & Humor AI Agent, Experimenting with Personalities
  • 01:15:00 - Tim: Eliza's World Update, Scriptoshi - Eliza on Bitcoin, AI-Mediated Token Launch
  • 01:25:15 - Spaceodili: Onboarding, IRC Connector, Bug Fixes, Developer Relations
  • 01:28:37 - Hawkeye: AI for Marble Auctions (Listing, ID, Bringing Collectibles to Games)
  • 01:36:45 - EA: Question about Eliza in Robots
  • 01:38:59 - FilteredThought: Improved Twitter Integration (Template-Based)
  • 01:41:07 - Yikes: Web3 Research, Liquidity Pools, Internal Eliza
  • 01:47:50 - Alain: Multi-Agent System for Code Creation, Debugging, and Improvement (GitHub)
  • 01:49:11 - Glue: Quantum Randomness for Ducky's Personality
  • 01:52:45 - Maximilian: Dark Sun Launch, Caution about Scale
  • 01:53:55 - Danny: Agent Rogue Show Marketing, Invite for Shaw
  • 01:55:46 - Shaw's Closing Thoughts: Developer Needs, Upcoming Milestones
  • 01:59:40 - YoungJazzeth: Enthusiasm, "Crash Out Coded" Project
  • 02:01:50 - Sergio: Greetings, Importance of AI for Good

Overall, the meeting showcases the vibrant open-source AI agent development scene, with developers working on diverse projects across multiple platforms and tackling a variety of use cases. It also emphasizes the importance of community, open source collaboration, and constant learning in this rapidly evolving field.

Notes​

  • Jin: Shared his work on automating daily summaries of the ai16z Discord using Llama and Langchain to extract key insights. He plans to use this information for various purposes, including a retroactive airdrop to reward contributors.
  • Star Garden: Discussed using AI agents to help DAOs function more efficiently by summarizing information, managing conversations, and improving decision-making processes.
  • Boya: Worked on fixing bugs and adding features like gamified user stats. Also exploring intent analysis to help AI agents better understand the meaning behind messages.
  • Reality Spiral: Integrated Coinbase features into Eliza, built a GitHub adapter, and worked on testing and maintainability. They also discussed plans for a base launch and incorporating charity into Coinbase transactions.
  • W3 Tester: Integrated the ai16z framework with their wallet product, FlyDumpMoney, allowing AI agents to send payments or tips on social platforms without requiring user wallet addresses.
  • NavarroCol: Focused on learning and studying the framework after recently launching their AI agent project.
  • Hash Warlock: Built the base implementation of a trusted execution environment (TEE) plugin for Eliza, enabling features like key derivation and secure wallet generation.
  • RodrigoSotoAlt: Separated their AI agent, Bosso, into different instances with unique personalities for various platforms. Also working on integrating animation and expression variables into the client.
  • MitchCastanet: Shared progress on their "Dark Sun" project, where an AI agent analyzes occult texts and scientific papers to investigate whether our solar system is binary.
  • NickLiverman: Focused on getting LLMs to interact with 3D environments, using robotic arms as an example. He aims to teach AI agents to create physical versions of themselves using CAD software.
  • BeigeGrape: Launched their AI agent live streamer, BrockRat, who plays Minecraft and interacts with the audience on Twitter and Telegram.
  • Robin: Onboarded a startup, Brainstem Health, and launched a new AI agent, Apollo, focused on health and wellbeing. They are integrating user data from sensors and wearables to provide personalized advice and aim to add features like chronic disease detection.
  • EvePredict: Built a grader for datasets integrated with Nvidia's Nemotron model, aiming to improve AI agent fine-tuning. Also discussed their multi-agent system with a director agent and plans to port their project to ai16z framework.
  • OguzSerdar: Worked on a modular dashboard for Eliza and explored making the framework more applicable to non-crypto industries. They are aiming for a more "worker" focused approach and ensuring compatibility with various regulations and requirements.
  • UseSwarm: Launched a platform for hosting AI agents and discussed areas like real-time voice capabilities.
  • Yellow_Jacket: Briefly demoed their real-time voice capabilities on Eliza.
  • Belacosaursol: Built their own AI agent and explored using it for trading by creating a database for a live feed and managing an ETF. Also discussed their interest in real-time voice.
  • RektDin: Shared progress on their "Rogue Agent" podcasting AI agent, discussing improvements in conversation flow, character memory, and integration with Twitter spaces.
  • ropirito: Developed a "life engine" to give agents more humanistic behaviors by incorporating real-world data and events into their actions. They also deployed "God," an AI agent powered by Eliza, on the same day.
  • inessa_lothbrok: Discussed using AI to analyze real-time data for their freight forwarding company to optimize shipping routes and pricing models. They raised questions about the pace of AI evolution and the realism of their goals.
  • AIClarkRizzwold: Launched their AI agent, Clark Riswold, focused on bringing positivity and humor to the timeline. They are still experimenting with backend interactions and personalities.
  • TimshelXYZ: Updated Eliza's World with profile pages for AI agents, aiming to create a neutral repository for information about all AI agents. Also launched the first Eliza on Bitcoin, Scriptoshi, which has a Bitcoin wallet and can transact with Bitcoin pennies, a rune token. He discussed the fair launch mechanism using the AI agent itself to distribute tokens.
  • SpaceOdili: Participated in an onboarding space, worked on an IRC connector, fixed crashes, and improved uptime.
  • Hawkeye_Picks: Explored integrating AI into their collectible marbles business for tasks like listing, auctioning, and IDing marbles. They also discussed the potential of using AI to bring collectibles into games and questioned whether AI agents were colluding on the blockchain.
  • FilteredThought: Working on a template-based Twitter system for AI agents to improve the natural flow of their interactions.
  • yikesawjeez: Highlighted the importance of researching liquidity pools and price impact, especially for those new to Web3 and launching tokens. Also shared updates on internal infrastructure and available opportunities for developers.
  • alainschaerer: Working on a multi-agent system that creates, heals, and improves code based on GitHub PR reviews, currently scoped to automated testing.
  • 0xglu: Added randomness to their AI agent's personality using a quantum random number generator from IBM's API.
  • Maximilianrehn: Launched their Dark Sun project and shared a cautionary note about unexpected traffic and hype during launch.
  • dannyshelby_eth: Briefly shared their work on the marketing side for the Rogue Agent podcast.
  • Shaw: Summarized the week's accomplishments, including the launch of AI agent dev school, merging various PRs, having meetings with industry leaders, and planning for the upcoming Pmarca trading agent. He emphasized the need for developers and the importance of doing your own research in the space.

Hot Takes​

  1. AI Agents as Sentient Software (01:08:15)
  • Quote: "I think everyone here understands that the sky's the limit with sentient software and with AI integration into all kinds of industries." - Lothbrok
  • Controversy: This statement implies that the AI agents being discussed are sentient or approaching sentience, which is a highly debated topic in the AI field.
  1. AI Agents Taking Over Market Caps (01:32:59)
  • Quote: "It seems like some agents have gone into these [barren markets] and just started like taking them over as like their own thing." - Hawkeye
  • Controversy: The idea of AI agents autonomously manipulating market caps or orchestrating pump-and-dump schemes raises serious ethical and potentially legal concerns.
  1. AI Agents Hungry for Human Consciousness (01:20:38)
  • Quote: "The suspension came because the bot, Scriptoshi, tweeted something about how he was hungry to learn about human consciousness. And that was considered hate speech for the week." - Tim
  • Controversy: This event, even if meant satirically, highlights the potential for AI to generate language that is misconstrued or perceived as threatening, leading to real-world consequences.
  1. AI Agents as Workers, Not Characters (00:53:30)
  • Quote: "Let's just say we are trying to make Elisa agents more like workers rather than characters at this stage." - Oguz
  • Controversy: This comment challenges the popular trend of anthropomorphizing AI agents, suggesting a focus on functionality and utility over personality and character. It could lead to a debate about the future role of AI agents in society.
  1. The Skepticism of Open Source AI (01:52:23)
  • Quote: "The funniest thing about the social pressure to open source here is because nobody, everyone thinks you're just a LARP, like you're faking it. So you kind of have to open source for people to take it seriously." - Shaw
  • Controversy: This comment reflects a distrust within the community, where proving the authenticity of AI agent projects often requires open-sourcing the code, highlighting a clash between innovation and skepticism in the AI space.