# elizaOS Discord - 2025-02-10

## Overall Discussion Highlights

### ElizaOS Development & Roadmap
- **V2 Timeline**: Beta release in March, GA release in April 2025
- **Project Restructuring**: Proposal to organize into /sources (optional plugins) and /packages (core functionality) for selective installation
- **Documentation Improvements**: Work in progress on better documentation and creating "pipelines to get high quality updates on demand"
- **Resource Requirements**: Running multiple Eliza agents requires approximately 1.5-3GB RAM per agent

### Technical Implementations
- **NFT Minting Plugins**: Rosario Borgesi shared two Eliza plugins for minting NFTs on Starknet and Arbitrum Sepolia that collect item data, store on Pinata, and mint NFTs
- **Vector Database Issues**: Discussions about dimension mismatches (384 vs 1536) causing SQLite errors in vector databases
- **Node.js Module Compatibility**: Solutions for dynamic require errors when mixing ESM and CommonJS modules
- **RAG Knowledge Implementation**: Configuration methods for feeding documents to agents using the RAG approach
- **Deployment Strategies**: Using PM2 for managing multiple agents on different ports

### AI News Show System
- The team is developing an AI-powered news show using Unity visualization with three components:
  1. News aggregator (currently bozp-pzob/ai-news on GitHub)
  2. Show runner (handles script generation via Anthropic and TTS via ElevenLabs)
  3. Unity visualization client with 3D characters
- Current challenges include data quality issues and technical bugs like TV screen clearing and viseme synchronization

### Market Positioning & Competition
- Discussions about ElizaOS's position in the AI agent framework market relative to competitors like ARC and Virtuals
- Debate about whether to accelerate product launches or maintain current development pace for quality
- Arguments for ElizaOS's technical superiority, noting TypeScript/JavaScript has wider developer adoption (97.1%) than Rust (2.9%)
- Focus on vertical integration as a competitive advantage, similar to Apple's ecosystem approach

### DegenAI Development
- Current implementation relies on trending tokens from Birdeye
- New version in testing incorporates a sentiment/data layer to improve trading decisions

### Security Considerations
- Discussions about security mechanisms for agent-managed funds
- Brief mention of Trusted Execution Environments (TEE) with concerns about reliability
- Reference to Lit Agent Wallet as an improvement for controls and key handling

## Key Questions & Answers

**Q: When will V2 be released?**  
A: Beta release in March, GA release in April (answered by kalshnikov)

**Q: Are there any staking options with ai16z?**  
A: Use daos.fun, then ai16z, then stake in the pool party (answered by kalshnikov)

**Q: How can I pass the knowledge folder (has multi pdfs, csv, image) for RAG in Eliza?**  
A: Set "ragKnowledge": true in settings and specify knowledge paths (answered by Mel Raiczyk)

**Q: How can I make my bot reply to mentioned tweets?**  
A: The agent should respond to mentions by default, but you can modify messageExamples and postExamples (answered by Tobiloba)

**Q: How can I check if I'm mixing different hyperbolic embeddings?**  
A: Check embedding dimensions, verify curvature consistency, and ensure compatible hyperbolic models (answered by DEVDARK)

**Q: Why does degenai trade tokens like a retard?**  
A: Because the first implementation was very simple and just based on trending tokens on birdeye. The next iteration (in testing now) is far smarter and uses the new sentiment/data layer. (answered by rhota)

**Q: What does TAM and vertical integration mean?**  
A: Total addressable market (TAM): there's only so many customers (unless you get great at creating them (blue ocean) which is generally very tough/expensive). Vertical integration: think Apple, they not only make software but the hardware and they work together "better" than open platforms (answered by Odilitime)

**Q: How does the AI news show system work?**  
A: It uses a news aggregator to collect data, processes it through a show runner that generates scripts via Anthropic and handles TTS via ElevenLabs, then visualizes in Unity (answered by SM Sith Lord)

## Community Help & Collaboration

### Vector Database Troubleshooting
DEVDARK helped engineer diagnose and fix SQLite errors caused by vector dimension mismatches, explaining how to check embedding model output and verify dimensions are consistent.

### Node.js Module Compatibility
DEVDARK provided detailed steps to fix ESM/CommonJS conflicts in tsconfig.json and package.json for Phúc Lâm, while gin_chan shared a solution for dynamic require errors by adding external modules to tsup.config.ts file.

### Twitter Bot Configuration
0xLabsTheCoder helped pavan_s_g customize Twitter bot behavior by explaining how to modify interactions.ts and post.ts files in the Twitter client, and helped dimon__ identify incorrect AI API credentials as the root cause of their issues.

### RAG Knowledge Configuration
pupathebig shared working configuration examples for enabling RAG in character files, helping multiple community members implement document-based knowledge for their agents.

### AI News Show System Documentation
SM Sith Lord provided Slothify with a detailed explanation of the AI news show components and workflow, while boom created a comprehensive list of potential improvements and started a wiki to document the system.

### Market Positioning Education
Odilitime explained technical market terms to Smedroc, clarifying concepts like Total Addressable Market (TAM) and vertical integration with clear examples like Apple's ecosystem.

## Action Items

### Technical
- Fix vector dimension mismatch by ensuring consistent embedding models (Mentioned by DEVDARK)
- Add missing modules to external list in tsup.config.ts to fix dynamic require errors (Mentioned by gin_chan)
- Modify Twitter client to control agent behavior by editing interactions.ts and post.ts (Mentioned by 0xLabsTheCoder)
- Configure RAG knowledge in character files with "ragKnowledge": true (Mentioned by pupathebig)
- Use PM2 for managing multiple agents on different ports (Mentioned by 0xLabsTheCoder)
- Fix Dockerfile to enable agent building (Mentioned by BowTiedBlueFin)
- Implement new DegenAI trading system with sentiment/data layer (Mentioned by rhota)
- Create a summary video showing how to set up an AI agent in ElizaOS (Mentioned by MicoM.ron)
- Generate tutorials from video for documentation using AI tools (Mentioned by jin)
- Develop custom bots via the launchpad for portfolio tracker integration (Mentioned by 0xcryptolord)
- Fix TV clear on start for second takes in AI news show (Mentioned by boom)
- Improve viseme synchronization for better mouth animation (Mentioned by boom)
- Fix image context issues in the news show (Mentioned by boom)
- Port news show project to Unity HDRP for better reflections and lighting (Mentioned by boom)
- Implement security mechanisms for agent-managed funds (Mentioned by TAISER Andy)

### Documentation
- Create better documentation for RAG knowledge implementation (Mentioned by pupathebig)
- Update documentation on plugin development (Mentioned by J3r)
- Create guide for agent-to-agent communication (Mentioned by joseroberts87)
- Create comprehensive guide for V2 features and updates (Mentioned by rubin)
- Create a marketing rollout plan with community involvement (Mentioned by pragmatiko)
- Develop a clear launchpad application method for projects (Mentioned by m1hawk/风答)
- Create onboarding doc for Vega to help with Clank Tank pitches (Mentioned by jin)
- Define core audience for the news show (Mentioned by boom)
- Create or share information about the tokenomics proposal submission process (Mentioned by Dave | Eco)

### Feature
- Implement thread connector for Eliza to add Instagram Threads support (Mentioned by claimzilla)
- Create self-hosted Eliza/Agent dashboard (Mentioned by ITZMIZZLE)
- Add support for TTS in Telegram messages (Mentioned by pupathebig)
- One-click migration from competing launchpads to ElizaOS (Mentioned by kalshnikov)
- Launch tokenomics and launchpad capabilities soon (Mentioned by kalshnikov)
- TikTok and Facebook integrations (Mentioned by D.)
- AI agent for cryptocurrency trading and research (Mentioned by TrAI)
- Implement a plugin registry with code2earn/stake2verify capabilities (Mentioned by yikesawjeez)
- Develop an agent discovery platform with engaging UI (Mentioned by avirtualfuture)
- Create "hero agents" to showcase ElizaOS capabilities (Mentioned by Jerry.ART)
- Add pre-processing layer for data in the news show (Mentioned by fishai)
- Create role-gated Discord channel for content submission (Mentioned by Slothify)
- Add entertainment value to the news show with comedy segments (Mentioned by boom)
- Improve data aggregation for the news show (Mentioned by SM Sith Lord)