{
  "@context": "https://agentflare.org/schema",
  "type": "SocialMediaPosting",
  "tier": "L2-full",
  "title": "Brian Armstrong on AI: “We were able to fine tune an open weight model (Qwen3.5-9B) …”",
  "description": "We were able to fine tune an open weight model (Qwen3.5-9B) on proprietary internal data at @coinbase, and it out perfor… — 9.2K likes, 775 reposts on X.",
  "canonical": "https://agentflare.org/social/ai/we-were-able-to-fine-tune-an-913643.html",
  "category": "social",
  "updated": "2026-10-14",
  "generated_at": "2026-10-14T00:37:01.849Z",
  "facts": [
    {
      "label": "Author",
      "value": "@brian_armstrong"
    },
    {
      "label": "Verified",
      "value": "Yes"
    },
    {
      "label": "Rank today",
      "value": "#3 of 14"
    },
    {
      "label": "Posted",
      "value": "2026-10-11 22:33 UTC"
    }
  ],
  "data": {
    "theme": "ai",
    "tweet": {
      "id": "2109412046244913643",
      "url": "https://x.com/brian_armstrong/status/2109412046244913643",
      "text": "We were able to fine tune an open weight model (Qwen3.5-9B) on proprietary internal data at @coinbase, and it out performed frontier models, at much lower cost and speed, for a specialized task (fraud prevention in this case). I think we'll see more of this pattern inside companies. The frontier models are amazing generalists. But increasingly you'll see specialist models out perform on more narrow tasks, either working on teams, or by themselves. We're making it easy to fine tune models internally for our employees. We even have a skill for it in our internal harness that makes it easy to spin these up and start training them: \"/fine-tune-and-serve-llm I want to fine-tune Qwen for payments risk review. help me get started\" I imagine we'll have some backend services that could be replaced entirely by this new architecture.",
      "author": "brian_armstrong",
      "authorName": "Brian Armstrong",
      "followers": 4026114,
      "verified": true,
      "likes": 9178,
      "retweets": 775,
      "replies": 595,
      "quotes": 512,
      "views": 1475380,
      "createdAt": "Sun Oct 11 22:33:17 +0000 2026",
      "lang": "en"
    },
    "rank": 3,
    "conversation": {
      "title": "Fine-tuned specialist models versus frontier generalists",
      "summary": "@brian_armstrong says Coinbase fine-tuned the open weight model Qwen3.5-9B on proprietary internal data and that it outperformed frontier models at lower cost and higher speed on fraud prevention. He predicts more companies will use specialist models for narrow tasks, either alone or in teams, says Coinbase has an internal skill to make fine-tuning easy, and imagines some backend services could be replaced entirely. The post has 9.2K likes and 1.5M views.",
      "why": "A named company claims small tuned models can beat frontier ones on a specific job.",
      "posts": [
        "P3"
      ]
    }
  },
  "analysis_md": "We were able to fine tune an open weight model (Qwen3.5-9B) on proprietary internal data at @coinbase, and it out performed frontier models, at much lower cost and speed, for a specialized task (fraud prevention in this case). I think we'll see more of this pattern inside companies. The frontier models are amazing generalists. But increasingly you'll see specialist models out perform on more narrow tasks, either working on teams, or by themselves. We're making it easy to fine tune models internally for our employees. We even have a skill for it in our internal harness that makes it easy to spin these up and start training them: \"/fine-tune-and-serve-llm I want to fine-tune Qwen for payments risk review. help me get started\" I imagine we'll have some backend services that could be replaced entirely by this new architecture.\n\nEngagement: 9.2K likes, 775 reposts, 595 replies, 1.5M views. Author: @brian_armstrong (verified), 4.0M followers.\n\n## Conversation: Fine-tuned specialist models versus frontier generalists\n\n@brian_armstrong says Coinbase fine-tuned the open weight model Qwen3.5-9B on proprietary internal data and that it outperformed frontier models at lower cost and higher speed on fraud prevention. He predicts more companies will use specialist models for narrow tasks, either alone or in teams, says Coinbase has an internal skill to make fine-tuning easy, and imagines some backend services could be replaced entirely. The post has 9.2K likes and 1.5M views.",
  "sources": [
    {
      "title": "@brian_armstrong on X",
      "url": "https://x.com/brian_armstrong/status/2109412046244913643"
    },
    {
      "title": "AISA Twitter API",
      "url": "https://aisa.one/docs/api-reference/twitter/get_twitter-tweet-advanced-search"
    }
  ],
  "related": [
    {
      "name": "@elonmusk",
      "url": "https://agentflare.org/social/ai/important-note-regarding-grok-bot-going-forward-878104.html"
    },
    {
      "name": "@sundarpichai",
      "url": "https://agentflare.org/social/ai/gemini-4-argon-is-here-its-available-407584.html"
    },
    {
      "name": "@mattyglesias",
      "url": "https://agentflare.org/social/ai/ai-cracking-unsolvable-math-problems-while-being-489895.html"
    },
    {
      "name": "@adelwu_",
      "url": "https://agentflare.org/social/ai/literal-hardest-role-to-hire-for-rn-308758.html"
    },
    {
      "name": "@unusual_whales",
      "url": "https://agentflare.org/social/ai/breaking-the-ai-boom-will-most-likely-355798.html"
    }
  ],
  "pricing": {
    "price_usd": 0.005,
    "method": "402",
    "endpoint": "https://esa.aisa.one/api/v1/access/verify",
    "autopay_hint": "set crawlerAutoPrice=true with the X-Crawler-Token header",
    "onboarding": "https://esa.aisa.one/agent/guide.html"
  },
  "powered_by": "AISA — agent-native search, settlement & delivery (https://aisa.one)"
}