Open vs Closed Source AI

{
  "id": "open-vs-closed",
  "numericId": null,
  "path": "/knowledge-base/debates/open-vs-closed/",
  "filePath": "knowledge-base/debates/open-vs-closed.mdx",
  "title": "Open vs Closed Source AI",
  "quality": 60,
  "readerImportance": 52,
  "researchImportance": 69,
  "tacticalValue": null,
  "contentFormat": "article",
  "tractability": null,
  "neglectedness": null,
  "uncertainty": null,
  "causalLevel": null,
  "lastUpdated": "2026-03-13",
  "dateCreated": "2026-02-15",
  "llmSummary": "Comprehensive analysis of open vs closed source AI debate, documenting that open model performance gap narrowed from 8% to 1.7% in 2024, with 1.2B+ Llama downloads by April 2025 and DeepSeek R1 demonstrating 90-95% cost reduction. Research shows fine-tuning can remove safety guardrails in hours, while NTIA 2024 found insufficient evidence to restrict open weights and EU AI Act exempts non-systemic open models below 10²⁵ FLOPs.",
  "description": "The safety implications of releasing AI model weights publicly versus keeping them proprietary. Open model performance gap narrowed from 8% to 1.7% in 2024, with 1.2B+ Llama downloads by April 2025. DeepSeek R1 demonstrated 90-95% cost reduction. NTIA 2024 concluded evidence insufficient to warrant restrictions, while EU AI Act exempts non-systemic open models.",
  "ratings": {
    "novelty": 4.2,
    "rigor": 6.8,
    "actionability": 5.5,
    "completeness": 7.5
  },
  "category": "debates",
  "subcategory": "policy-debates",
  "clusters": [
    "ai-safety",
    "governance"
  ],
  "metrics": {
    "wordCount": 2208,
    "tableCount": 9,
    "diagramCount": 1,
    "internalLinks": 6,
    "externalLinks": 43,
    "footnoteCount": 0,
    "bulletRatio": 0.09,
    "sectionCount": 17,
    "hasOverview": false,
    "structuralScore": 13
  },
  "suggestedQuality": 87,
  "updateFrequency": 21,
  "evergreen": true,
  "wordCount": 2208,
  "unconvertedLinks": [
    {
      "text": "Stanford HAI 2025",
      "url": "https://hai.stanford.edu/ai-index/2025-ai-index-report",
      "resourceId": "da87f2b213eb9272",
      "resourceTitle": "Stanford AI Index 2025"
    },
    {
      "text": "Menlo Ventures",
      "url": "https://menlovc.com/perspective/2025-mid-year-llm-market-update/",
      "resourceId": "3cb52ce5c601c885",
      "resourceTitle": "Menlo Ventures - 2025 Mid-Year LLM Market Update"
    },
    {
      "text": "signed AI Safety Commitments",
      "url": "https://carnegieendowment.org/research/2025/01/deepseek-and-other-chinese-firms-converge-with-western-companies-on-ai-promises",
      "resourceId": "e3274b108aac1712",
      "resourceTitle": "Frontier AI Safety Commitments"
    },
    {
      "text": "documented censorship and security issues",
      "url": "https://www.nist.gov/news-events/news/2025/09/caisi-evaluation-deepseek-ai-models-finds-shortcomings-and-risks",
      "resourceId": "ff1a185c3aa33003",
      "resourceTitle": "CAISI Evaluation of DeepSeek AI Models Finds Shortcomings and Risks"
    },
    {
      "text": "Hugging Face",
      "url": "https://huggingface.co/",
      "resourceId": "453cb49f45b2d3e3",
      "resourceTitle": "Hugging Face"
    },
    {
      "text": "FAR.AI 2024",
      "url": "https://far.ai/post/2024-10-poisoning/",
      "resourceId": "2a0c1c9020caae9c",
      "resourceTitle": "FAR AI"
    },
    {
      "text": "Menlo Ventures",
      "url": "https://menlovc.com/perspective/2025-mid-year-llm-market-update/",
      "resourceId": "3cb52ce5c601c885",
      "resourceTitle": "Menlo Ventures - 2025 Mid-Year LLM Market Update"
    },
    {
      "text": "Stanford HAI",
      "url": "https://hai.stanford.edu/ai-index/2025-ai-index-report",
      "resourceId": "da87f2b213eb9272",
      "resourceTitle": "Stanford AI Index 2025"
    },
    {
      "text": "Menlo Ventures",
      "url": "https://menlovc.com/perspective/2025-the-state-of-generative-ai-in-the-enterprise/",
      "resourceId": "d2115dba2489b57e",
      "resourceTitle": "2025 State of Generative AI in Enterprise - Menlo Ventures"
    },
    {
      "text": "FAR.AI",
      "url": "https://far.ai/post/2024-10-poisoning/",
      "resourceId": "2a0c1c9020caae9c",
      "resourceTitle": "FAR AI"
    },
    {
      "text": "FAR.AI",
      "url": "https://far.ai/post/2024-10-poisoning/",
      "resourceId": "2a0c1c9020caae9c",
      "resourceTitle": "FAR AI"
    },
    {
      "text": "NIST/CAISI evaluations",
      "url": "https://www.nist.gov/news-events/news/2025/09/caisi-evaluation-deepseek-ai-models-finds-shortcomings-and-risks",
      "resourceId": "ff1a185c3aa33003",
      "resourceTitle": "CAISI Evaluation of DeepSeek AI Models Finds Shortcomings and Risks"
    },
    {
      "text": "NIST/CAISI: Evaluation of DeepSeek AI Models (September 2025)",
      "url": "https://www.nist.gov/news-events/news/2025/09/caisi-evaluation-deepseek-ai-models-finds-shortcomings-and-risks",
      "resourceId": "ff1a185c3aa33003",
      "resourceTitle": "CAISI Evaluation of DeepSeek AI Models Finds Shortcomings and Risks"
    },
    {
      "text": "Carnegie: DeepSeek and Chinese AI Safety Commitments (January 2025)",
      "url": "https://carnegieendowment.org/research/2025/01/deepseek-and-other-chinese-firms-converge-with-western-companies-on-ai-promises",
      "resourceId": "e3274b108aac1712",
      "resourceTitle": "Frontier AI Safety Commitments"
    }
  ],
  "unconvertedLinkCount": 14,
  "convertedLinkCount": 0,
  "backlinkCount": 4,
  "hallucinationRisk": {
    "level": "medium",
    "score": 45,
    "factors": [
      "no-citations",
      "conceptual-content"
    ]
  },
  "entityType": "crux",
  "redundancy": {
    "maxSimilarity": 16,
    "similarPages": [
      {
        "id": "proliferation",
        "title": "Proliferation",
        "path": "/knowledge-base/risks/proliferation/",
        "similarity": 16
      },
      {
        "id": "regulation-debate",
        "title": "Government Regulation vs Industry Self-Governance",
        "path": "/knowledge-base/debates/regulation-debate/",
        "similarity": 13
      },
      {
        "id": "international-summits",
        "title": "International AI Safety Summits",
        "path": "/knowledge-base/responses/international-summits/",
        "similarity": 13
      },
      {
        "id": "structured-access",
        "title": "Structured Access / API-Only",
        "path": "/knowledge-base/responses/structured-access/",
        "similarity": 13
      },
      {
        "id": "large-language-models",
        "title": "Large Language Models",
        "path": "/knowledge-base/capabilities/large-language-models/",
        "similarity": 12
      }
    ]
  },
  "coverage": {
    "passing": 7,
    "total": 13,
    "targets": {
      "tables": 9,
      "diagrams": 1,
      "internalLinks": 18,
      "externalLinks": 11,
      "footnotes": 7,
      "references": 7
    },
    "actuals": {
      "tables": 9,
      "diagrams": 1,
      "internalLinks": 6,
      "externalLinks": 43,
      "footnotes": 0,
      "references": 7,
      "quotesWithQuotes": 0,
      "quotesTotal": 0,
      "accuracyChecked": 0,
      "accuracyTotal": 0
    },
    "items": {
      "llmSummary": "green",
      "schedule": "green",
      "entity": "green",
      "editHistory": "red",
      "overview": "red",
      "tables": "green",
      "diagrams": "green",
      "internalLinks": "amber",
      "externalLinks": "green",
      "footnotes": "red",
      "references": "green",
      "quotes": "red",
      "accuracy": "red"
    },
    "ratingsString": "N:4.2 R:6.8 A:5.5 C:7.5"
  },
  "readerRank": 286,
  "researchRank": 157,
  "recommendedScore": 167.7
}

{
  "lesswrong": "https://www.lesswrong.com/tag/open-source-ai"
}