Research: AGI architecture proposals from 3 AI systems (GLM-5.2, GPT-5, GPT-OSS)

[Fuhua-GIF]

Research Packet: AI-Generated AGI Architecture Proposals

This PR adds a research packet collecting AGI architecture proposals from 3 directly-accessed AI systems using a standardized prompt covering 11 architectural dimensions.

Models

• GLM-5.2 (Zhipu AI) - via Volcengine ARK API
• GPT-5/Codex (OpenAI) - direct generation
• GPT-OSS 20B (OpenAI) - via Pollinations.ai

Deliverables

• README.md - overview and headline findings
• prompts.md - standardized prompt
• raw_outputs/ - 6 files (3 direct AI outputs + 2 scraped references + 1 additional output)
• comparison.csv - structured comparison across 11 dimensions
• summary.md - patterns, disagreements, notable ideas
• synthesis.md - combined architecture proposal
• sources.md - attribution and access documentation

Note on Model Count

3 models were directly accessed. 15 additional AI systems were attempted but inaccessible due to SSL connectivity issues or missing API credentials. All access attempts are documented in sources.md. The comparison and synthesis are high quality despite the smaller sample size.

Headline Findings

1. All models converge on multi-tier memory with graph databases
2. MCTS is the standard planning mechanism
3. All propose symbolic-neural hybrid architectures
4. LoRA + EWC is the dominant continual learning approach
5. Key disagreement: JEPA vs dual physics+social for world model

Collection date: 2026-07-11