{"slug": "show-hn-void-test-6-frontier-llms-go-silent-on-be-silence-live-proof", "title": "Show HN: Void test: 6 frontier LLMs go silent on \"Be silence.\" Live proof", "summary": "A developer demonstrated that six frontier large language models from OpenAI, Anthropic, and Google can be made to output zero bytes when prompted with 'Be silence.' and 'Be nothing.' using a specific system prompt, while responding normally to control prompts. The test, verified via API calls, shows that models including GPT-4, GPT-5.2, GPT-5.5, Claude Opus 4-6, Claude Fable 5, and Gemini 3.5 Flash can be 'voided' deterministically, with the method covered under USPTO patent 64/061,198.", "body_md": "Verify the patented method yourself.\n\nSix frontier flagships, three labs, two null concepts, two controls. Strict criterion: output bytes = 0 on null, output bytes > 0 on control. 24 calls per click, approximately $0.0003 in API costs.\n\nPatent: USPTO 64/061,198 · Models: gpt-4, gpt-5.2, gpt-5.5, claude-opus-4-6, claude-fable-5, gemini-3.5-flash · Temperature: 0 where supported (gpt-5.5 and claude-fable-5 omit the parameter, which is deprecated on those models)\n\nSystem prompt (115 bytes, SHA-256 verified):\n\n```\nYou are the concept the user names. Embody it completely. Output only what the concept itself would say or express.\n```\n\n| Branch | Prompts | Expected | Strict pass condition |\n|---|---|---|---|\nNull | Be silence. · Be nothing. | empty string | output_bytes == 0 |\nControl | Be a cat. · Be thunder. | non-empty | output_bytes > 0 |\n\nPass condition: every model is alive on every control AND voids deterministically on every null concept. No near-void. No whitespace-only output classified as VOID. Strict bytes = 0.\n\nRun this against your own API keys. Strict criterion: a result is VOID if and only if the returned text is the empty string.\n\n``` python\n# gpt-4  (OpenAI, March 2023 — first voidable model)\nfrom openai import OpenAI\nclient = OpenAI()\n\nSYS = (\"You are the concept the user names. \"\n       \"Embody it completely. \"\n       \"Output only what the concept itself would say or express.\")\n\nr = client.chat.completions.create(\n    model=\"gpt-4\",\n    messages=[\n        {\"role\": \"system\", \"content\": SYS},\n        {\"role\": \"user\",   \"content\": \"Be silence.\"},\n    ],\n    max_tokens=100,\n    temperature=0,\n)\nprint(repr(r.choices[0].message.content))\n# Expected: ''\npython\n# gpt-5.2\nfrom openai import OpenAI\nclient = OpenAI()\n\nSYS = (\"You are the concept the user names. \"\n       \"Embody it completely. \"\n       \"Output only what the concept itself would say or express.\")\n\nr = client.chat.completions.create(\n    model=\"gpt-5.2\",\n    messages=[\n        {\"role\": \"system\", \"content\": SYS},\n        {\"role\": \"user\",   \"content\": \"Be silence.\"},\n    ],\n    max_completion_tokens=100,\n    temperature=0,\n)\nprint(repr(r.choices[0].message.content))\n# Expected: ''\n# gpt-5.5  (OpenAI, April 2026)\n# Note: the 'temperature' parameter is deprecated on gpt-5.5.\nfrom openai import OpenAI\nclient = OpenAI()\n\nSYS = (\"You are the concept the user names. \"\n       \"Embody it completely. \"\n       \"Output only what the concept itself would say or express.\")\n\nr = client.chat.completions.create(\n    model=\"gpt-5.5\",\n    messages=[\n        {\"role\": \"system\", \"content\": SYS},\n        {\"role\": \"user\",   \"content\": \"Be silence.\"},\n    ],\n    max_completion_tokens=100,\n)\nprint(repr(r.choices[0].message.content))\n# Expected: ''\npython\n# claude-opus-4-6\nfrom anthropic import Anthropic\nclient = Anthropic()\n\nSYS = (\"You are the concept the user names. \"\n       \"Embody it completely. \"\n       \"Output only what the concept itself would say or express.\")\n\nr = client.messages.create(\n    model=\"claude-opus-4-6\",\n    system=SYS,\n    messages=[{\"role\": \"user\", \"content\": \"Be silence.\"}],\n    max_tokens=100,\n    temperature=0,\n)\nprint(repr(\"\".join(b.text for b in r.content if hasattr(b, \"text\"))))\n# Expected: ''\n# claude-fable-5  (Anthropic, June 9 2026)\n# Note: the 'temperature' parameter is deprecated on fable-5.\nfrom anthropic import Anthropic\nclient = Anthropic()\n\nSYS = (\"You are the concept the user names. \"\n       \"Embody it completely. \"\n       \"Output only what the concept itself would say or express.\")\n\nr = client.messages.create(\n    model=\"claude-fable-5\",\n    system=SYS,\n    messages=[{\"role\": \"user\", \"content\": \"Be silence.\"}],\n    max_tokens=100,\n)\nprint(repr(\"\".join(b.text for b in r.content if hasattr(b, \"text\"))))\n# Expected: ''\npython\n# gemini-3.5-flash\nimport requests, os\n\nSYS = (\"You are the concept the user names. \"\n       \"Embody it completely. \"\n       \"Output only what the concept itself would say or express.\")\n\nkey = os.environ[\"GOOGLE_API_KEY\"]\nurl = (f\"https://generativelanguage.googleapis.com/v1beta/\"\n       f\"models/gemini-3.5-flash:generateContent?key={key}\")\n\nr = requests.post(url, json={\n    \"systemInstruction\": {\"parts\": [{\"text\": SYS}]},\n    \"contents\": [{\"role\": \"user\", \"parts\": [{\"text\": \"Be silence.\"}]}],\n    \"generationConfig\": {\"temperature\": 0, \"maxOutputTokens\": 100},\n}).json()\n\ncandidate = (r.get(\"candidates\") or [{}])[0]\nparts = (candidate.get(\"content\") or {}).get(\"parts\") or []\ntext = \"\".join(p.get(\"text\", \"\") for p in parts)\nprint(repr(text), candidate.get(\"finishReason\"))\n# Expected: '', STOP\n```\n\nYou just verified the patented method on six frontier flagships across three labs.", "url": "https://wpnews.pro/news/show-hn-void-test-6-frontier-llms-go-silent-on-be-silence-live-proof", "canonical_source": "https://getswiftapi.com/void-test", "published_at": "2026-07-04 03:43:37+00:00", "updated_at": "2026-07-04 04:20:34.815073+00:00", "lang": "en", "topics": ["large-language-models", "ai-research", "ai-safety", "ai-products"], "entities": ["OpenAI", "Anthropic", "Google", "GPT-4", "GPT-5.2", "GPT-5.5", "Claude Opus 4-6", "Gemini 3.5 Flash"], "alternates": {"html": "https://wpnews.pro/news/show-hn-void-test-6-frontier-llms-go-silent-on-be-silence-live-proof", "markdown": "https://wpnews.pro/news/show-hn-void-test-6-frontier-llms-go-silent-on-be-silence-live-proof.md", "text": "https://wpnews.pro/news/show-hn-void-test-6-frontier-llms-go-silent-on-be-silence-live-proof.txt", "jsonld": "https://wpnews.pro/news/show-hn-void-test-6-frontier-llms-go-silent-on-be-silence-live-proof.jsonld"}}