{
  "type": "tweet",
  "id": "2093022661677081004",
  "url": "https://x.com/Scobleizer/status/2093022661677081004",
  "twitterUrl": "https://twitter.com/Scobleizer/status/2093022661677081004",
  "text": "RT @kwindla: Introducing PhoneLLM, an open model for voice agents.\n\nGPT 5.6 Terra performance on typical voice agent tasks at 1/3 the laten…",
  "source": "Twitter for iPhone",
  "retweetCount": 51,
  "replyCount": 27,
  "likeCount": 512,
  "quoteCount": 9,
  "viewCount": 28408,
  "createdAt": "Thu Aug 27 17:07:43 +0000 2026",
  "lang": "en",
  "bookmarkCount": 583,
  "isReply": false,
  "inReplyToId": null,
  "conversationId": "2093022661677081004",
  "displayTextRange": [
    0,
    140
  ],
  "inReplyToUserId": null,
  "inReplyToUsername": null,
  "author": {
    "type": "user",
    "userName": "Scobleizer",
    "url": "https://x.com/Scobleizer",
    "twitterUrl": "https://twitter.com/Scobleizer",
    "id": "13348",
    "name": "Robert Scoble",
    "isVerified": false,
    "isBlueVerified": true,
    "verifiedType": null,
    "profilePicture": "https://pbs.twimg.com/profile_images/1915614118876504066/zVnfpAMf_normal.jpg",
    "coverPicture": "https://pbs.twimg.com/profile_banners/13348/1777608401",
    "description": "",
    "location": "San Jose, California, USA",
    "followers": 601540,
    "following": 56672,
    "status": "",
    "canDm": true,
    "canMediaTag": true,
    "createdAt": "Mon Nov 20 23:43:44 +0000 2006",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {}
    },
    "fastFollowersCount": 0,
    "favouritesCount": 637102,
    "hasCustomTimelines": true,
    "isTranslator": false,
    "mediaCount": 6734,
    "statusesCount": 257457,
    "withheldInCountries": [],
    "affiliatesHighlightedLabel": {
      "label": {
        "badge": {
          "url": "https://pbs.twimg.com/profile_images/1981734716748234752/FT4VUW6r_bigger.jpg"
        },
        "description": "Boardy",
        "url": {
          "url": "https://twitter.com/boardyai",
          "urlType": "DeepLink"
        },
        "userLabelDisplayType": "Badge",
        "userLabelType": "BusinessLabel"
      }
    },
    "possiblySensitive": false,
    "pinnedTweetIds": [
      "2092287044882252222"
    ],
    "profile_bio": {
      "description": "San Francisco/Silicon Valley AI | Robots, holodecks, BCIs, analysis of new things | Ex-Microsoft, Rackspace, Fast Company | Wrote eight books about the future.",
      "entities": {
        "description": {},
        "url": {
          "urls": [
            {
              "display_url": "alignednews.com/ai",
              "expanded_url": "https://alignednews.com/ai",
              "indices": [
                0,
                23
              ],
              "url": "https://t.co/8L5xphk0qQ"
            }
          ]
        }
      }
    },
    "isAutomated": false,
    "automatedBy": null
  },
  "extendedEntities": {},
  "card": null,
  "place": {},
  "entities": {
    "user_mentions": [
      {
        "id_str": "16375739",
        "indices": [
          3,
          11
        ],
        "name": "kwindla",
        "screen_name": "kwindla"
      }
    ]
  },
  "quoted_tweet": null,
  "retweeted_tweet": {
    "type": "tweet",
    "id": "2093014818647339026",
    "url": "https://x.com/kwindla/status/2093014818647339026",
    "twitterUrl": "https://twitter.com/kwindla/status/2093014818647339026",
    "text": "Introducing PhoneLLM, an open model for voice agents.\n\nGPT 5.6 Terra performance on typical voice agent tasks at 1/3 the latency and 1/18 the cost.\n\nFor voice agents, we need models that are both very low latency and very good at tool calling and instruction following.\n\nThere's a trade-off here, and we often have to compromise on either latency or capability when building voice agents. With PhoneLLM (and the training and data stack that made this model possible) we're fixing this problem.\n\nFor the last couple of years, most of the effort in frontier model development has gone towards leveraging test-time compute. Which is awesome! Models of all shapes and sizes are available that perform really, really well ... if you have \"thinking\" turned on for your model.\n\nBut if you need your agent to respond at voice conversation speed, you can't use thinking models.\n\nPhoneLLM is a full-weights fine-tune of NVIDIA Nemotron Nano 30B. We trained on a wide range of real-world telephone and customer support use cases. The training focused on taking the excellent Nano 30B base capabilities and teaching the model to do typical voice agent tasks with thinking disabled.\n\nThe results are really good: accurate tool calling and concise, on-topic responses in long conversations.\n\nAnd fast: TTFAT measured server-side is <100ms if you run PhoneLLM on a lightly loaded B200. :-)\n\nBut seriously, when we characterize model latency, we do it with full, end-to-end, batched request simulations using real Pipecat voice agent pipelines.\n\nYou can serve more than 80 concurrent agents on a single B200 with P95 end-to-end TTFAT <600ms. Including network overhead. That's an LLM cost-per-minute around $0.0025. (1/4 of a cent.) At a latency lower than any third-party API offers today.\n\nMore details about this model, including weights on @huggingface, how to spin it up with one click on Modal, and a starter project repo you can clone, are in the thread ...",
    "source": "Twitter for iPhone",
    "retweetCount": 51,
    "replyCount": 27,
    "likeCount": 512,
    "quoteCount": 9,
    "viewCount": 28408,
    "createdAt": "Thu Aug 27 16:36:33 +0000 2026",
    "lang": "en",
    "bookmarkCount": 583,
    "isReply": false,
    "inReplyToId": null,
    "conversationId": "2093014818647339026",
    "displayTextRange": [
      0,
      280
    ],
    "inReplyToUserId": null,
    "inReplyToUsername": null,
    "author": {
      "type": "user",
      "userName": "kwindla",
      "url": "https://x.com/kwindla",
      "twitterUrl": "https://twitter.com/kwindla",
      "id": "16375739",
      "name": "kwindla",
      "isVerified": false,
      "isBlueVerified": true,
      "verifiedType": null,
      "profilePicture": "https://pbs.twimg.com/profile_images/1790772534914551808/YpwkVUIl_normal.jpg",
      "coverPicture": "https://pbs.twimg.com/profile_banners/16375739/1502226088",
      "description": "",
      "location": "San Francisco, CA",
      "followers": 15061,
      "following": 3923,
      "status": "",
      "canDm": true,
      "canMediaTag": true,
      "createdAt": "Sat Sep 20 07:14:14 +0000 2008",
      "entities": {
        "description": {
          "urls": []
        },
        "url": {}
      },
      "fastFollowersCount": 0,
      "favouritesCount": 8351,
      "hasCustomTimelines": true,
      "isTranslator": false,
      "mediaCount": 1389,
      "statusesCount": 6544,
      "withheldInCountries": [],
      "affiliatesHighlightedLabel": {
        "label": {
          "badge": {
            "url": "https://pbs.twimg.com/profile_images/1855992730360713216/_x37w4M7_bigger.jpg"
          },
          "description": "Daily",
          "url": {
            "url": "https://twitter.com/trydaily",
            "urlType": "DeepLink"
          },
          "userLabelDisplayType": "Badge",
          "userLabelType": "BusinessLabel"
        }
      },
      "possiblySensitive": false,
      "pinnedTweetIds": [],
      "profile_bio": {
        "description": "Infrastructure and developer tools for real-time voice, video, and AI. @trydaily // ᓚᘏᗢ // @pipecat_ai",
        "entities": {
          "description": {},
          "url": {
            "urls": [
              {
                "display_url": "machine-theory.com",
                "expanded_url": "https://machine-theory.com/",
                "indices": [
                  0,
                  23
                ],
                "url": "https://t.co/plyseTkcW0"
              }
            ]
          }
        }
      },
      "isAutomated": false,
      "automatedBy": null
    },
    "extendedEntities": {
      "media": [
        {
          "additional_media_info": {
            "monetizable": false
          },
          "display_url": "pic.x.com/CGMNR6Ltej",
          "expanded_url": "https://x.com/kwindla/status/2093014818647339026/video/1",
          "ext_media_availability": {
            "status": "Available"
          },
          "id_str": "2093013293560709120",
          "indices": [
            281,
            304
          ],
          "media_key": "13_2093013293560709120",
          "media_results": {
            "result": {
              "media_key": "13_2093013293560709120"
            }
          },
          "media_url_https": "https://pbs.twimg.com/amplify_video_thumb/2093013293560709120/img/y0b0JFTxKUQ4XDS8.jpg",
          "original_info": {
            "focus_rects": [],
            "height": 1080,
            "width": 1920
          },
          "sizes": {
            "large": {
              "h": 1080,
              "resize": "fit",
              "w": 1920
            },
            "medium": {
              "h": 675,
              "resize": "fit",
              "w": 1200
            },
            "small": {
              "h": 383,
              "resize": "fit",
              "w": 680
            },
            "thumb": {
              "h": 150,
              "resize": "crop",
              "w": 150
            }
          },
          "type": "video",
          "url": "https://t.co/CGMNR6Ltej",
          "video_info": {
            "aspect_ratio": [
              16,
              9
            ],
            "duration_millis": 208200,
            "variants": [
              {
                "content_type": "application/x-mpegURL",
                "url": "https://video.twimg.com/amplify_video/2093013293560709120/pl/812cD5MwWOHcHcNp.m3u8?tag=29&v=334"
              },
              {
                "bitrate": 256000,
                "content_type": "video/mp4",
                "url": "https://video.twimg.com/amplify_video/2093013293560709120/vid/avc1/480x270/n_x0Y_YslgrFXqTy.mp4?tag=29"
              },
              {
                "bitrate": 832000,
                "content_type": "video/mp4",
                "url": "https://video.twimg.com/amplify_video/2093013293560709120/vid/avc1/640x360/hevousAkHgFkn1YV.mp4?tag=29"
              },
              {
                "bitrate": 2176000,
                "content_type": "video/mp4",
                "url": "https://video.twimg.com/amplify_video/2093013293560709120/vid/avc1/1280x720/d0RJRniRpAnJPS4I.mp4?tag=29"
              },
              {
                "bitrate": 10368000,
                "content_type": "video/mp4",
                "url": "https://video.twimg.com/amplify_video/2093013293560709120/vid/avc1/1920x1080/XQiBNIbvUKdPv5GK.mp4?tag=29"
              }
            ]
          }
        }
      ]
    },
    "card": null,
    "place": {},
    "entities": {
      "hashtags": [],
      "smarttags": [],
      "symbols": [],
      "timestamps": [],
      "urls": [],
      "user_mentions": [
        {
          "id_str": "778764142412984320",
          "indices": [
            1828,
            1840
          ],
          "name": "Hugging Face",
          "screen_name": "huggingface"
        }
      ]
    },
    "quoted_tweet": null,
    "retweeted_tweet": null,
    "article": null
  },
  "article": null
}