# Fact Briefing: 2025-03-28

## Overall Summary
Auto.fun launched as a token launchpad following the Eliza Studios website countdown, while ElizaOS users continued to report migration and plugin-configuration issues across v0.25.9/v1.0.0/v2. GitHub activity centered on community/UX features and test coverage, alongside new issues related to Twitter integration and a plugin dependency version resolution error.

## Key Facts

- The Eliza Studios website countdown completed and auto.fun launched, described as a launchpad.
- A team member (rhota) stated Spartan will be available for interaction on Discord first as v2 work continues, without waiting on X review processes.
- Binance Alpha 2.0 was reported as allowing users to purchase degenAI without requiring a Web3 wallet.
- Users reported running llama3.1:8b-instruct-q4_K_M on a GeForce 3060 (12GB VRAM) at approximately 15 tokens/sec with a 32K context window.
- Discord guidance stated that in v0.25.9, plugins should be added via the character configuration file using the fully qualified plugin name in the "plugins" array.
- GitHub issue #4101 reported an npm 'notarget' error for @elizaos/plugin-sql@^0.25.6 (no matching version found).
- GitHub issue #4102 reported that links and hashtags were not appearing properly in Twitter posts.
- The GitHub daily digest reported completion of a Discord community manager feature (PR #4099) and tests for each agent type in the the-org package (PR #4090).

## Open Questions

- Will the dev school be happening tomorrow?
- How can I update to the new version?
- Is version 0.25.9 OK? Agents don't show up on screen, always says disconnected?
- Can the Twitter plugin make my agent retweet, like or follow users?
- What happened to the liquidity?
- Countdown's over, now what?
- What's the best model provider method for Twitter that won't cost so much money compared to OpenAI?
- Is there any way to do tracing on Eliza's LLM interactions similar to LangSmith?

## Categories

### Twitter News Highlights
- Auto.fun promoted the platform with the tagline "press the fun button" and engaged with users about the service. (Sentiment: neutral)
- DankVR shared that BoomboxHeads is using an open-source tool to wrangle data from sources including Discord, GitHub, X, and markets, and posted a behind-the-scenes video of an automated 3D show workflow. (Sentiment: neutral)

### GitHub Updates

#### New Issues/PRs
- [Issue #4102: Twitter integration challenges: difficulties retrieving links and hashtags.](https://github.com/elizaos/eliza/issues/4102) by unknown - Status: open - Significance: User-reported functional problem affecting Twitter post formatting.
- [Issue #4101: Dependency issue: npm error for SQL plugin version.](https://github.com/elizaos/eliza/issues/4101) by unknown - Status: open - Significance: Blocks installation/usage when a required package version cannot be resolved.
- [Issue #4097: Inquiry about `ENABLE_TWITTER_POST_GENERATION` usage.](https://github.com/elizaos/eliza/issues/4097) by unknown - Status: open - Significance: Potentially indicates configuration drift or deprecated settings in v2.
- [Pull_request #4099: feat: Discord community manager](https://github.com/elizaos/eliza/pull/4099) by unknown - Status: merged - Significance: Adds automated community engagement (greetings/timeout actions) in Discord.
- [Pull_request #4090: test: the-org agent types](https://github.com/elizaos/eliza/pull/4090) by unknown - Status: merged - Significance: Improves reliability via expanded test coverage for agent types.
- [Pull_request #4085: fix: settings action hallucinations](https://github.com/elizaos/eliza/pull/4085) by unknown - Status: merged - Significance: Addresses incorrect/undesired agent behavior tied to settings actions.
- [Pull_request #4103: chore: reduce log level for null/undefined encryption/decryption errors](https://github.com/elizaos/eliza/pull/4103) by unknown - Status: merged - Significance: Reduces log noise to improve operational debugging signal.

#### Overall Focus
- GitHub work emphasized community engagement features (Discord community manager) and expanded automated testing for agent types, while triaging new Twitter integration and dependency-resolution issues.

### Discord Updates
- **#💻-coders:** Users discussed version migration (v0.25.9 → v1.0.0 → v2), plugin configuration (notably Twitter/Discord/custom plugins), local LLM setups (Ollama), and API rate limiting (Anthropic 50k input tokens/min). Guidance included shifting to v2 CLI-based character management and updated core imports (composePrompt; useModel(ModelType.OBJECT_SMALL)). (Key Participants: jin, kaisermerkle, winded4752, new.moon)
- **#spartan_holders:** Community discussed Spartan’s blocked X account and the option to rebrand rather than wait for recovery. A team member confirmed work on Spartan v2 with Discord-first availability, and Binance Alpha 2.0 purchasing of degenAI without a Web3 wallet was mentioned. (Key Participants: rhota)
- **#🥇-partners:** Conversation focused on the Eliza Studios countdown and the subsequent appearance/launch of auto.fun as a launchpad, plus a request for feedback on a "taming_info" blog post. (Key Participants: jin, Patt)
- **#associates:** Members compared GPT-4o image generation to earlier ComfyUI/ControlNet/LoRA workflows; GPT-4o was described as enabling consistent multi-character scenes up to around five characters via natural language prompts. (Key Participants: jin, Patt)
- **#discussion:** Users asked about ai16z staking (answered as 2-sided LP with SOL on daos.fun), ticker-change progress (metadata upgrade), and raised unanswered questions about version updates and agent disconnection behavior in v0.25.9. (Key Participants: Patt)

### User Feedback
- Multiple users requested clearer migration documentation between v0.25.9 and v1.0.0/v2, including character creation and file placement guidance for v2. (Sentiment: neutral)
- Users reported Twitter plugin/client issues across versions, including duplicate status errors and incorrect/unrelated tweet content, and requested better Twitter plugin capability documentation. (Sentiment: mixed)
- A request was made for tracing of LLM interactions similar to LangSmith. (Sentiment: neutral)
- Users requested support for more cost-effective model providers for Twitter workflows compared to OpenAI. (Sentiment: neutral)

### Strategic Insights

#### Discord-first communications due to X account constraints
Project communications and user interaction for Spartan are being repositioned toward Discord-first availability while X/Twitter access and review processes remain a constraint.

*Implications/Questions:*
  - What product and moderation features are required to support higher-volume Discord interaction if it becomes the primary interface?

#### Migration friction and documentation discoverability
Repeated questions about v0.25.9/v1.0.0/v2 differences, character configuration, and plugin setup suggest that migration and onboarding documentation is a central bottleneck for developer adoption.

*Implications/Questions:*
  - Should the project publish a single versioned migration playbook and pin it in Discord/GitHub to reduce repeated support load?

#### Growing emphasis on local inference for agent workloads
Community discussion includes concrete local LLM throughput and context-window reports (e.g., 8B instruct models on consumer GPUs), indicating ongoing interest in local model deployments for RAG/agents.

*Implications/Questions:*
  - Are there reference configurations (hardware + model + settings) that should be standardized and documented for repeatability?

### Market Analysis
- ai16z staking was described as currently available only via a 2-sided LP with SOL on daos.fun. (Relevance: Defines current yield/staking path and may affect holder expectations and onboarding.)
- A crypto market update reported ai16z trading between $0.1856 and $0.1989 (~7% increase between reported prices) and WBTC between $86,771.64 and $87,067.89. (Relevance: Provides near-term price context for project token discussions.)
- Binance Alpha 2.0 was reported as enabling purchases of degenAI without a Web3 wallet. (Relevance: Reduces onboarding friction for retail acquisition through a centralized platform flow.)
- Community chat included real-time tracking of Solana tokens on pump.fun, including large reported market cap moves (e.g., Abstract Pepe) and use of bot commands for monitoring. (Relevance: Reflects attention on Solana meme/AI token dynamics that may overlap with launchpad narratives.)