{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreibnb54gja5ciz7nrh2ddys36qwny4v3pcqu57wubl6ch3zbrnbibe",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mo4v6pbirjj2"
  },
  "path": "/t/june-2026-theme-through-time-chatgpt-api-image-generative-art-gallery-prompt-tips-and-help/1378298?page=98#post_1977",
  "publishedAt": "2026-06-12T22:18:48.000Z",
  "site": "https://community.openai.com",
  "tags": [
    "(click for more details)"
  ],
  "textContent": "Oh this is so cool!\nI did something similar a while ago. But more with archetypes.\n\nFound these as of now, SFW versions. Still fenced to be safe.\n\nExamples (click for more details)\n\nAlso did something I call Three Levels of Semantic Entropy, this was done with GPT Image 1.5.\n\nThe idea here was, a Mime and a Berserker stand in front of each other and have to intimidated each other without attacking each other. Along those lines was the prompt. And then for each level I kept thee same base prompt but started adding agnostic semantic adjectives that would enhance the scene but not the interaction. Level 1 proved great. Level 2 GPT started setting them on all variations a first attack, subtle one. But on Level 3 it was an All out war and the thing was: I never once asked for them to attack each other. Mime has to Mime the best, Berserker has to yell the best.\n\nLevel 1\n\n\n\nLevel 2\n\n\n\nLevel 3\n",
  "title": "June 2026 (Theme: Through Time) — ChatGPT / API Image Generative Art Gallery, Prompt Tips, and Help"
}