🐦 Twitter Post Details

Viewing enriched Twitter post

@hardmaru

For the past few years, humans have been doing “prompt engineering” to coax the best performance out of different LLMs. In this work, we explored what happens if we train an AI to do that job instead. By training a Conductor model with RL, we found that it naturally learns to write highly effective, custom instructions for a whole pool of other models. It essentially learns to ‘manage’ them in natural language. What surprised me most was how it dynamically adapts. For simple factual questions, it just queries one model. But for hard coding problems, it autonomously spins up a whole pipeline of planners, coders, and verifiers. Really excited to see where this paradigm of “AI managing AI” goes next, especially as we start moving from single-agent chain-of-thought to multi-agent “chain-of-command”. Link to our #ICLR2026 paper: https://t.co/EwbjjRPLUb Along with our TRINITY paper which we announced earlier, this work also powers our new multi-agent system: Sakana Fugu (https://t.co/2m8VRdOYqG) 🐡

Media 1
View on Twitter

📊 Media Metadata

{
  "media": [
    {
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2048778095935795338/media_1.jpg",
      "media_url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2048778095935795338/media_1.jpg",
      "type": "photo",
      "filename": "media_1.jpg"
    }
  ],
  "processed_at": "2026-05-11T22:37:09.086026",
  "pipeline_version": "2.0"
}

🔧 Raw API Response

{
  "type": "tweet",
  "id": "2048778095935795338",
  "url": "https://x.com/hardmaru/status/2048778095935795338",
  "twitterUrl": "https://twitter.com/hardmaru/status/2048778095935795338",
  "text": "For the past few years, humans have been doing “prompt engineering” to coax the best performance out of different LLMs. In this work, we explored what happens if we train an AI to do that job instead.\n\nBy training a Conductor model with RL, we found that it naturally learns to write highly effective, custom instructions for a whole pool of other models. It essentially learns to ‘manage’ them in natural language.\n\nWhat surprised me most was how it dynamically adapts. For simple factual questions, it just queries one model. But for hard coding problems, it autonomously spins up a whole pipeline of planners, coders, and verifiers.\n\nReally excited to see where this paradigm of “AI managing AI” goes next, especially as we start moving from single-agent chain-of-thought to multi-agent “chain-of-command”.\n\nLink to our #ICLR2026 paper: https://t.co/EwbjjRPLUb\n\nAlong with our TRINITY paper which we announced earlier, this work also powers our new multi-agent system: Sakana Fugu (https://t.co/2m8VRdOYqG) 🐡",
  "source": "Twitter for iPhone",
  "retweetCount": 175,
  "replyCount": 40,
  "likeCount": 1450,
  "quoteCount": 21,
  "viewCount": 180673,
  "createdAt": "Mon Apr 27 14:55:37 +0000 2026",
  "lang": "en",
  "bookmarkCount": 1356,
  "isReply": false,
  "inReplyToId": null,
  "conversationId": "2048778095935795338",
  "displayTextRange": [
    0,
    277
  ],
  "inReplyToUserId": null,
  "inReplyToUsername": null,
  "author": {
    "type": "user",
    "userName": "hardmaru",
    "url": "https://x.com/hardmaru",
    "twitterUrl": "https://twitter.com/hardmaru",
    "id": "2895499182",
    "name": "hardmaru",
    "isVerified": false,
    "isBlueVerified": true,
    "verifiedType": null,
    "profilePicture": "https://pbs.twimg.com/profile_images/1678402467078234113/XN5Oy2UP_normal.jpg",
    "coverPicture": "https://pbs.twimg.com/profile_banners/2895499182/1429351923",
    "description": "",
    "location": "Minato-ku, Tokyo",
    "followers": 405023,
    "following": 1857,
    "status": "",
    "canDm": false,
    "canMediaTag": true,
    "createdAt": "Mon Nov 10 11:05:07 +0000 2014",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {}
    },
    "fastFollowersCount": 0,
    "favouritesCount": 143573,
    "hasCustomTimelines": true,
    "isTranslator": false,
    "mediaCount": 4473,
    "statusesCount": 25854,
    "withheldInCountries": [],
    "affiliatesHighlightedLabel": {},
    "possiblySensitive": false,
    "pinnedTweetIds": [
      "2036841736702767135"
    ],
    "profile_bio": {
      "description": "Co-Founder and CEO @SakanaAILabs 🎏",
      "entities": {
        "description": {
          "user_mentions": [
            {
              "id_str": "",
              "indices": [
                19,
                32
              ],
              "name": "",
              "screen_name": "SakanaAILabs"
            }
          ]
        },
        "url": {
          "urls": [
            {
              "display_url": "sakana.ai",
              "expanded_url": "https://sakana.ai/",
              "indices": [
                0,
                23
              ],
              "url": "https://t.co/cVQF43vIdu"
            }
          ]
        }
      }
    },
    "isAutomated": false,
    "automatedBy": null
  },
  "extendedEntities": {},
  "card": {
    "binding_values": [
      {
        "key": "thumbnail_image",
        "value": {
          "image_value": {
            "alt": "arXiv logo",
            "height": 144,
            "url": "https://pbs.twimg.com/card_img/2048997175515066368/mqP1mS8z?format=jpg&name=144x144_2",
            "width": 144
          }
        }
      },
      {
        "key": "description",
        "value": {
          "string_value": "Powerful large language models (LLMs) from different providers have been expensively trained and finetuned to specialize across varying domains. In this work, we introduce a new kind of Conductor..."
        }
      },
      {
        "key": "domain",
        "value": {
          "string_value": "arxiv.org"
        }
      },
      {
        "key": "thumbnail_image_large",
        "value": {
          "image_value": {
            "alt": "arXiv logo",
            "height": 420,
            "url": "https://pbs.twimg.com/card_img/2048997175515066368/mqP1mS8z?format=jpg&name=420x420_2",
            "width": 420
          }
        }
      },
      {
        "key": "thumbnail_image_original",
        "value": {
          "image_value": {
            "alt": "arXiv logo",
            "height": 1000,
            "url": "https://pbs.twimg.com/card_img/2048997175515066368/mqP1mS8z?format=jpg&name=orig",
            "width": 1000
          }
        }
      },
      {
        "key": "site",
        "value": {
          "scribe_key": "publisher_id",
          "user_value": {
            "id_str": "808633423300624384",
            "path": []
          }
        }
      },
      {
        "key": "thumbnail_image_small",
        "value": {
          "image_value": {
            "alt": "arXiv logo",
            "height": 100,
            "url": "https://pbs.twimg.com/card_img/2048997175515066368/mqP1mS8z?format=jpg&name=100x100_2",
            "width": 100
          }
        }
      },
      {
        "key": "thumbnail_image_x_large",
        "value": {
          "image_value": {
            "alt": "arXiv logo",
            "height": 1000,
            "url": "https://pbs.twimg.com/card_img/2048997175515066368/mqP1mS8z?format=png&name=2048x2048_2_exp",
            "width": 1000
          }
        }
      },
      {
        "key": "thumbnail_image_alt_text",
        "value": {
          "string_value": "arXiv logo"
        }
      },
      {
        "key": "vanity_url",
        "value": {
          "scribe_key": "vanity_url",
          "string_value": "arxiv.org"
        }
      },
      {
        "key": "thumbnail_image_color",
        "value": {
          "image_color_value": {
            "palette": [
              {
                "percentage": 94.17,
                "rgb": {
                  "blue": 255,
                  "green": 255,
                  "red": 255
                }
              },
              {
                "percentage": 4.33,
                "rgb": {
                  "blue": 105,
                  "green": 116,
                  "red": 124
                }
              },
              {
                "percentage": 1.26,
                "rgb": {
                  "blue": 46,
                  "green": 21,
                  "red": 170
                }
              },
              {
                "percentage": 0.23,
                "rgb": {
                  "blue": 131,
                  "green": 116,
                  "red": 203
                }
              }
            ]
          }
        }
      },
      {
        "key": "title",
        "value": {
          "string_value": "Learning to Orchestrate Agents in Natural Language with the Conductor"
        }
      },
      {
        "key": "card_url",
        "value": {
          "scribe_key": "card_url",
          "string_value": "https://t.co/EwbjjRPLUb"
        }
      }
    ],
    "card_platform": {
      "platform": {
        "audience": {
          "name": "production"
        },
        "device": {
          "name": "iPhone",
          "version": "13"
        }
      }
    },
    "name": "summary",
    "url": "https://t.co/EwbjjRPLUb",
    "user_refs_results": [
      {
        "rest_id": "808633423300624384",
        "result": {
          "__typename": "User",
          "action_counts": {
            "favorites_count": 985
          },
          "avatar": {
            "image_url": "https://pbs.twimg.com/profile_images/1365352170267299840/IzvjKckL_normal.jpg"
          },
          "banner": {
            "image_url": "https://pbs.twimg.com/profile_banners/808633423300624384/1481635469"
          },
          "core": {
            "created_at": "Tue Dec 13 11:23:26 +0000 2016",
            "name": "arXiv.org",
            "screen_name": "arxiv"
          },
          "dm_permissions": {
            "can_dm": true
          },
          "exclusive_tweet_following": false,
          "follow_request_sent": false,
          "identity_profile_labels_highlighted_label": {},
          "location": {
            "location": "Ithaca, NY",
            "profile_location_place": {
              "country": "",
              "country_code": "",
              "full_name": "Ithaca, NY",
              "id": "ae76bffcaf2bf545",
              "name": "Ithaca, NY",
              "place_type": "unknown"
            }
          },
          "media_permissions": {
            "can_media_tag": true
          },
          "notifications_settings": {
            "notifications_enabled": false
          },
          "pinned_items": {},
          "privacy": {
            "protected": false,
            "suspended": false
          },
          "private_super_following": false,
          "profile_bio": {
            "description": "News from https://t.co/enurGFxpcS, a free distribution service and an open archive for scholarly articles.\n\nFor help with arXiv, see https://t.co/LcWuhM0BOl",
            "entities": {
              "description": {
                "urls": [
                  {
                    "display_url": "arXiv.org",
                    "expanded_url": "http://arXiv.org",
                    "indices": [
                      10,
                      33
                    ],
                    "url": "https://t.co/enurGFxpcS"
                  },
                  {
                    "display_url": "arxiv.org/help",
                    "expanded_url": "https://arxiv.org/help",
                    "indices": [
                      133,
                      156
                    ],
                    "url": "https://t.co/LcWuhM0BOl"
                  }
                ]
              },
              "url": {
                "urls": [
                  {
                    "display_url": "arxiv.org",
                    "expanded_url": "https://arxiv.org/",
                    "indices": [
                      0,
                      23
                    ],
                    "url": "https://t.co/DHMkdi4lF9"
                  }
                ]
              }
            }
          },
          "profile_image_shape": "Circle",
          "profile_metadata": {
            "profile_interstitial_type": "",
            "profile_link_color": "ABB8C2"
          },
          "profile_translation": {
            "translator_type_enum": "None"
          },
          "properties": {
            "has_extended_profile": true,
            "has_no_screen_name": false
          },
          "relationship_counts": {
            "followers": 46711,
            "following": 184
          },
          "relationship_perspectives": {
            "blocked_by": false,
            "blocking": false,
            "followed_by": false,
            "following": false,
            "live_following": false,
            "muting": false
          },
          "rest_id": "808633423300624384",
          "smart_blocked_by": false,
          "smart_blocking": false,
          "super_follow_eligible": false,
          "super_followed_by": false,
          "super_following": false,
          "tweet_counts": {
            "media_tweets": 118,
            "tweets": 1131
          },
          "verification": {
            "is_blue_verified": false,
            "verified": false
          },
          "website": {
            "url": "https://t.co/DHMkdi4lF9"
          }
        }
      }
    ]
  },
  "place": {},
  "entities": {
    "hashtags": [
      {
        "indices": [
          823,
          832
        ],
        "text": "ICLR2026"
      }
    ],
    "symbols": [],
    "urls": [
      {
        "display_url": "arxiv.org/abs/2512.04388",
        "expanded_url": "https://arxiv.org/abs/2512.04388",
        "indices": [
          840,
          863
        ],
        "url": "https://t.co/EwbjjRPLUb"
      },
      {
        "display_url": "sakana.ai/fugu-beta",
        "expanded_url": "https://sakana.ai/fugu-beta",
        "indices": [
          985,
          1008
        ],
        "url": "https://t.co/2m8VRdOYqG"
      }
    ],
    "user_mentions": []
  },
  "quoted_tweet": {
    "type": "tweet",
    "id": "2048777689763639741",
    "url": "https://x.com/SakanaAILabs/status/2048777689763639741",
    "twitterUrl": "https://twitter.com/SakanaAILabs/status/2048777689763639741",
    "text": "Introducing our new work: “Learning to Orchestrate Agents in Natural Language with the Conductor” accepted at #ICLR2026\n\nhttps://t.co/Wnh9ZACmLm\n\nWhat if we trained an AI not to solve problems directly, but to act as a manager that delegates tasks to a diverse team of other AIs?\n\nTo solve complex tasks, humans rarely work alone; we form teams, delegate, and communicate. Yet, multi-agent AI systems currently rely heavily on rigid, human-designed workflows or simple routers that just pick a single model. We wanted an AI that could dynamically build its own team.\n\nWe trained a 7B Conductor model using Reinforcement Learning to orchestrate a pool of frontier models (including GPT-5, Gemini, Claude, and open-source models available during the period leading up to ICLR 2026).\n\nInstead of executing code, the Conductor outputs a collaborative workflow in natural language. For any given question, the Conductor specifies:\n\n1/ Which agent to call\n2/ What specific subtask to give them (acting as an expert prompt engineer)\n3/ What previous messages they can see in their context window\n\nThrough pure end-to-end reward maximization, amazing behaviors emerged. The Conductor learned to adapt to task difficulty: it 1-shots simple factual questions, but autonomously spins up complex planner-executor-verifier pipelines for hard coding problems.\n\nThe results are very promising: The 7B Conductor surpasses the performance of every individual worker model in its pool, setting new records on LiveCodeBench (83.9%) and GPQA-Diamond (87.5%) at the time of publication. It also significantly outperforms expensive multi-agent baselines like Mixture-of-Agents at a fraction of the cost.\n\nOne of our favorite features: Recursive Test-Time Scaling! By allowing the Conductor to select itself as a worker, it reads its own team's prior output, realizes if it failed, and spins up a corrective workflow on the fly. This opens a new axis for scaling compute during inference.\n\nThis research proves that language models can become elite meta-prompt engineers, dynamically harnessing collective intelligence.\n\nAlongside our TRINITY research which we announced a few days earlier, this foundational research powers our new multi-agent system: Sakana Fugu! (https://t.co/36Ud311KCP) 🐡\n\nOpenReview: https://t.co/e5WqTleQNL (ICLR 2026)",
    "source": "Twitter for iPhone",
    "retweetCount": 119,
    "replyCount": 17,
    "likeCount": 650,
    "quoteCount": 14,
    "viewCount": 223490,
    "createdAt": "Mon Apr 27 14:54:00 +0000 2026",
    "lang": "en",
    "bookmarkCount": 552,
    "isReply": false,
    "inReplyToId": null,
    "conversationId": "2048777689763639741",
    "displayTextRange": [
      0,
      279
    ],
    "inReplyToUserId": null,
    "inReplyToUsername": null,
    "author": {
      "type": "user",
      "userName": "SakanaAILabs",
      "url": "https://x.com/SakanaAILabs",
      "twitterUrl": "https://twitter.com/SakanaAILabs",
      "id": "218811492",
      "name": "Sakana AI",
      "isVerified": false,
      "isBlueVerified": true,
      "verifiedType": "Business",
      "profilePicture": "https://pbs.twimg.com/profile_images/1885939209388929024/dtnrOdGp_normal.jpg",
      "coverPicture": "https://pbs.twimg.com/profile_banners/218811492/1686643464",
      "description": "",
      "location": "Tokyo, Japan",
      "followers": 70759,
      "following": 0,
      "status": "",
      "canDm": false,
      "canMediaTag": true,
      "createdAt": "Tue Nov 23 10:20:07 +0000 2010",
      "entities": {
        "description": {
          "urls": []
        },
        "url": {}
      },
      "fastFollowersCount": 0,
      "favouritesCount": 2,
      "hasCustomTimelines": true,
      "isTranslator": false,
      "mediaCount": 380,
      "statusesCount": 1086,
      "withheldInCountries": [],
      "affiliatesHighlightedLabel": {},
      "possiblySensitive": false,
      "pinnedTweetIds": [
        "2036840833690071450"
      ],
      "profile_bio": {
        "description": "Sakana AI is an AI R&D company based in Tokyo. We develop AI solutions for Japan’s needs, and democratize AI in Japan. Try Sakana Chat: https://t.co/1m2lSgnfB2",
        "entities": {
          "description": {
            "urls": [
              {
                "display_url": "sakana.ai",
                "expanded_url": "https://sakana.ai/",
                "indices": [
                  136,
                  159
                ],
                "url": "https://t.co/1m2lSgnfB2"
              }
            ]
          },
          "url": {
            "urls": [
              {
                "display_url": "sakana.ai/careers",
                "expanded_url": "https://sakana.ai/careers",
                "indices": [
                  0,
                  23
                ],
                "url": "https://t.co/1q07mb3TzE"
              }
            ]
          }
        }
      },
      "isAutomated": false,
      "automatedBy": null
    },
    "extendedEntities": {
      "media": [
        {
          "display_url": "pic.twitter.com/1BcqayXSGl",
          "expanded_url": "https://twitter.com/SakanaAILabs/status/2048777689763639741/photo/1",
          "ext_media_availability": {
            "status": "Available"
          },
          "features": {
            "large": {
              "faces": []
            },
            "orig": {
              "faces": []
            }
          },
          "id_str": "2048777511304310785",
          "indices": [
            280,
            303
          ],
          "media_key": "3_2048777511304310785",
          "media_results": {
            "id": "QXBpTWVkaWFSZXN1bHRzOgwAAQoAARxuuEqpGjABCgACHG64dDYbcb0AAA==",
            "result": {
              "__typename": "ApiMedia",
              "id": "QXBpTWVkaWE6DAABCgABHG64SqkaMAEKAAIcbrh0NhtxvQAA",
              "media_key": "3_2048777511304310785"
            }
          },
          "media_url_https": "https://pbs.twimg.com/media/HG64SqkaMAE6wJo.jpg",
          "original_info": {
            "focus_rects": [
              {
                "h": 531,
                "w": 948,
                "x": 0,
                "y": 0
              },
              {
                "h": 948,
                "w": 948,
                "x": 0,
                "y": 0
              },
              {
                "h": 1081,
                "w": 948,
                "x": 0,
                "y": 0
              },
              {
                "h": 1200,
                "w": 600,
                "x": 29,
                "y": 0
              },
              {
                "h": 1200,
                "w": 948,
                "x": 0,
                "y": 0
              }
            ],
            "height": 1200,
            "width": 948
          },
          "sizes": {
            "large": {
              "h": 1200,
              "w": 948
            }
          },
          "type": "photo",
          "url": "https://t.co/1BcqayXSGl"
        }
      ]
    },
    "card": null,
    "place": {},
    "entities": {
      "hashtags": [
        {
          "indices": [
            110,
            119
          ],
          "text": "ICLR2026"
        }
      ],
      "symbols": [],
      "urls": [
        {
          "display_url": "arxiv.org/abs/2512.04388",
          "expanded_url": "https://arxiv.org/abs/2512.04388",
          "indices": [
            121,
            144
          ],
          "url": "https://t.co/Wnh9ZACmLm"
        },
        {
          "display_url": "sakana.ai/fugu-beta",
          "expanded_url": "https://sakana.ai/fugu-beta",
          "indices": [
            2244,
            2267
          ],
          "url": "https://t.co/36Ud311KCP"
        },
        {
          "display_url": "openreview.net/forum?id=U23A2…",
          "expanded_url": "https://openreview.net/forum?id=U23A2BUKYt",
          "indices": [
            2284,
            2307
          ],
          "url": "https://t.co/e5WqTleQNL"
        }
      ],
      "user_mentions": []
    },
    "quoted_tweet": {
      "type": "tweet",
      "id": "2047479445209145785",
      "url": "",
      "twitterUrl": "",
      "text": "",
      "source": "Twitter for iPhone",
      "retweetCount": 0,
      "replyCount": 0,
      "likeCount": 0,
      "quoteCount": 0,
      "viewCount": 0,
      "createdAt": "",
      "lang": "",
      "bookmarkCount": 0,
      "isReply": false,
      "inReplyToId": null,
      "conversationId": "",
      "displayTextRange": [],
      "inReplyToUserId": null,
      "inReplyToUsername": null,
      "author": {},
      "extendedEntities": {},
      "card": null,
      "place": {},
      "entities": {},
      "quoted_tweet": null,
      "retweeted_tweet": null,
      "isLimitedReply": false,
      "communityInfo": null,
      "article": null
    },
    "retweeted_tweet": null,
    "isLimitedReply": false,
    "communityInfo": null,
    "article": null
  },
  "retweeted_tweet": null,
  "isLimitedReply": false,
  "communityInfo": null,
  "article": null
}