{
  "slug": "deleted-70-percent-ai-instructions-claude-got-smarter",
  "title": "I deleted 70% of my AI instructions and Claude got smarter",
  "date": "2026-08-03",
  "featuredClaim": "Deleting outdated AI instructions, not adding more, is what actually makes Claude perform better.",
  "description": "A guide on optimizing AI prompt instructions by removing outdated constraints that no longer serve a purpose. The author shares a methodology based on Anthropic's discovery that deleting 80% of its own instructions improved performance, along with practical templates for auditing and cleaning system prompts across settings, projects, and skills.",
  "keyPoints": [
    "Most AI instructions become outdated and create unnecessary processing overhead; deleting constraints the model no longer needs improves performance",
    "Use the key question: 'Would a brilliant new colleague in my field still need to be told this?' to determine what to keep versus delete",
    "Instructions live in three separate locations (Settings, Projects, Skills) and require systematic auditing across all three layers",
    "Only delete based on performance testing, not assumptions; wait for patterns of problems before adding instructions back"
  ],
  "topics": [
    {
      "id": "strategy",
      "slug": "ai-strategy",
      "label": "AI Strategy",
      "description": "Strategic planning and implementation approaches for AI adoption"
    },
    {
      "id": "implementation",
      "slug": "ai-implementation",
      "label": "Implementation",
      "description": "Hands-on implementation techniques and frameworks"
    },
    {
      "id": "tools",
      "slug": "ai-tools",
      "label": "AI Tools",
      "description": "Practical tools and platforms for AI implementation"
    }
  ],
  "claims": [
    "Anthropic deleted more than 80% of its coding tool's instructions without any drop in test scores.",
    "The author deleted 70 percent of personal Claude instructions and reported noticeably smarter, more useful output.",
    "Conflicting rules, like documenting appropriately versus never adding comments, force the model to arbitrarily choose.",
    "Claude instructions live across three separate locations: general settings, project files, and skill definitions.",
    "Skills only load into context when actively called, so idle skills cost nothing extra."
  ],
  "claimTitles": [
    "Anthropic's 80% Deletion Test",
    "Author's 70% Instruction Cut",
    "Conflicting Rules Confuse Models",
    "Three Instruction Storage Layers",
    "Skills Load On Demand"
  ],
  "originalUrl": "https://aiadopters.club/p/i-deleted-70-of-my-ai-instructions",
  "quote": "Would a brilliant new colleague, who already knows my field, still need to be told this?",
  "keyStatistics": [
    {
      "stat": "80%+",
      "context": "Share of instructions Anthropic removed from its Claude Code tool, with test scores remaining unchanged."
    },
    {
      "stat": "70%",
      "context": "Portion of personal AI instructions the author cut, which reportedly made Claude's responses sharper."
    },
    {
      "stat": "~66%",
      "context": "Estimated share of instructions cleared out when applying the 'brilliant new colleague' filter to a typical prompt set."
    },
    {
      "stat": "10 seconds",
      "context": "Suggested time to spend per instruction line when deciding whether it still needs to be kept."
    }
  ],
  "supportingContext": "The methodology centers on a single diagnostic question—whether a knowledgeable new hire would still need a given instruction—applied across three distinct layers where Claude stores context: general settings, project-level files, and skills. Practitioners are advised to delete instructions wholesale rather than trim them, then reintroduce only what proves necessary through a week of normal use, adding items back solely after a problem repeats rather than after a single incident. This mirrors Anthropic's own internal practice of stripping instructions to zero and rebuilding line by line to test actual utility. The approach distinguishes between corrective instructions—meant to patch mistakes models no longer make—and preference instructions like tone, audience, or style, which remain essential and should not be cut. The caveat that newer model versions show longer system prompts suggests the deletion trend may not be as uniformly aggressive as headline figures imply.",
  "canonicalUrl": "https://kbanc.com/claims-library/deleted-70-percent-ai-instructions-claude-got-smarter",
  "markdownUrl": "https://kbanc.com/md/claims-library/deleted-70-percent-ai-instructions-claude-got-smarter.md",
  "jsonUrl": "https://kbanc.com/api/claims/deleted-70-percent-ai-instructions-claude-got-smarter.json",
  "source": {
    "publisher": "AI Adopters Club",
    "title": "I deleted 70% of my AI instructions and Claude got smarter",
    "url": "https://aiadopters.club/p/i-deleted-70-of-my-ai-instructions"
  }
}