🐦 Twitter Post Details

Viewing enriched Twitter post

@LiorOnAI

Google just gave language models real long-term memory. A new architecture learns during inference and keeps context across millions of tokens. It holds ~70 percent accuracy at 10 million tokens. š—§š—µš—¶š˜€ š—®š—æš—°š—µš—¶š˜š—²š—°š˜š˜‚š—æš—² š—¹š—²š—®š—æš—»š˜€ š˜„š—µš—¶š—¹š—² š—¶š˜ š—æš˜‚š—»š˜€ Titans adds a neural long-term memory that updates during generation. Not weights. Not retraining. Live learning. • A small neural network stores long-range context • It updates only when something unexpected appears • Routine tokens get ignored to stay fast This lets the model remember facts from far earlier text without scanning everything again. š—œš˜ š—øš—²š—²š—½š˜€ š˜€š—½š—²š—²š—± š˜„š—µš—¶š—¹š—² š˜€š—°š—®š—¹š—¶š—»š—“ š—°š—¼š—»š˜š—²š˜…š˜ Attention stays local. Memory handles the past. • Linear inference cost • No quadratic attention blowups • Stable accuracy past two million tokens š—œš˜ š—®š—¹š—¹š—¼š˜„š˜€ š˜†š—¼š˜‚ š˜š—¼ š—Æš˜‚š—¶š—¹š—± š—»š—²š˜„ š—øš—¶š—»š—±š˜€ š—¼š—³ š—®š—½š—½š˜€ You can process full books, logs, or genomes in one pass. You can keep state across long sessions. You can stop chunking context just to survive limits.

Media 1

šŸ“Š Media Metadata

{
  "media": [
    {
      "type": "photo",
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2012555843087073541/media_0.jpg?",
      "filename": "media_0.jpg"
    }
  ],
  "processed_at": "2026-01-18T17:32:30.600091",
  "pipeline_version": "2.0"
}

šŸ”§ Raw API Response

{
  "type": "tweet",
  "id": "2012555843087073541",
  "url": "https://x.com/LiorOnAI/status/2012555843087073541",
  "twitterUrl": "https://twitter.com/LiorOnAI/status/2012555843087073541",
  "text": "Google just gave language models real long-term memory.\n\nA new architecture learns during inference and keeps context across millions of tokens.\n\nIt holds ~70 percent accuracy at 10 million tokens.\n\nš—§š—µš—¶š˜€ š—®š—æš—°š—µš—¶š˜š—²š—°š˜š˜‚š—æš—² š—¹š—²š—®š—æš—»š˜€ š˜„š—µš—¶š—¹š—² š—¶š˜ š—æš˜‚š—»š˜€\nTitans adds a neural long-term memory that updates during generation.\nNot weights. Not retraining. Live learning.\n\n• A small neural network stores long-range context\n• It updates only when something unexpected appears\n• Routine tokens get ignored to stay fast\n\nThis lets the model remember facts from far earlier text without scanning everything again.\n\nš—œš˜ š—øš—²š—²š—½š˜€ š˜€š—½š—²š—²š—± š˜„š—µš—¶š—¹š—² š˜€š—°š—®š—¹š—¶š—»š—“ š—°š—¼š—»š˜š—²š˜…š˜\nAttention stays local. Memory handles the past.\n\n• Linear inference cost\n• No quadratic attention blowups\n• Stable accuracy past two million tokens\n\nš—œš˜ š—®š—¹š—¹š—¼š˜„š˜€ š˜†š—¼š˜‚ š˜š—¼ š—Æš˜‚š—¶š—¹š—± š—»š—²š˜„ š—øš—¶š—»š—±š˜€ š—¼š—³ š—®š—½š—½š˜€\nYou can process full books, logs, or genomes in one pass.\nYou can keep state across long sessions.\nYou can stop chunking context just to survive limits.",
  "source": "Twitter for iPhone",
  "retweetCount": 65,
  "replyCount": 31,
  "likeCount": 426,
  "quoteCount": 6,
  "viewCount": 24401,
  "createdAt": "Sat Jan 17 16:01:19 +0000 2026",
  "lang": "en",
  "bookmarkCount": 265,
  "isReply": false,
  "inReplyToId": null,
  "conversationId": "2012555843087073541",
  "displayTextRange": [
    0,
    269
  ],
  "inReplyToUserId": null,
  "inReplyToUsername": null,
  "author": {
    "type": "user",
    "userName": "LiorOnAI",
    "url": "https://x.com/LiorOnAI",
    "twitterUrl": "https://twitter.com/LiorOnAI",
    "id": "931470139",
    "name": "Lior Alexander",
    "isVerified": false,
    "isBlueVerified": true,
    "verifiedType": null,
    "profilePicture": "https://pbs.twimg.com/profile_images/1980366446975766528/LPbXxZYl_normal.jpg",
    "coverPicture": "https://pbs.twimg.com/profile_banners/931470139/1761077189",
    "description": "",
    "location": "",
    "followers": 109076,
    "following": 1982,
    "status": "",
    "canDm": true,
    "canMediaTag": false,
    "createdAt": "Wed Nov 07 07:19:36 +0000 2012",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {}
    },
    "fastFollowersCount": 0,
    "favouritesCount": 6604,
    "hasCustomTimelines": true,
    "isTranslator": false,
    "mediaCount": 638,
    "statusesCount": 3632,
    "withheldInCountries": [],
    "affiliatesHighlightedLabel": {},
    "possiblySensitive": false,
    "pinnedTweetIds": [],
    "profile_bio": {
      "description": "Daily posts on the latest dev news in AI • Founder @AlphaSignalAI (250k users) • ML Eng since 2017 • Ex-Mila",
      "entities": {
        "description": {
          "user_mentions": [
            {
              "id_str": "0",
              "indices": [
                51,
                65
              ],
              "name": "",
              "screen_name": "AlphaSignalAI"
            }
          ]
        },
        "url": {
          "urls": [
            {
              "display_url": "alphasignal.ai",
              "expanded_url": "https://alphasignal.ai",
              "indices": [
                0,
                23
              ],
              "url": "https://t.co/AyubevadmD"
            }
          ]
        }
      }
    },
    "isAutomated": false,
    "automatedBy": null
  },
  "extendedEntities": {
    "media": [
      {
        "display_url": "pic.twitter.com/84a5YsXXR0",
        "expanded_url": "https://twitter.com/LiorOnAI/status/2012555843087073541/photo/1",
        "ext_media_availability": {
          "status": "Available"
        },
        "features": {
          "large": {},
          "orig": {}
        },
        "id_str": "2012555839593197568",
        "indices": [
          270,
          293
        ],
        "media_key": "3_2012555839593197568",
        "media_results": {
          "id": "QXBpTWVkaWFSZXN1bHRzOgwAAQoAARvuCN+FlzAACgACG+4I4FXXgQUAAA==",
          "result": {
            "__typename": "ApiMedia",
            "id": "QXBpTWVkaWE6DAABCgABG+4I34WXMAAKAAIb7gjgVdeBBQAA",
            "media_key": "3_2012555839593197568"
          }
        },
        "media_url_https": "https://pbs.twimg.com/media/G-4I34WXMAAKiiZ.jpg",
        "original_info": {
          "focus_rects": [
            {
              "h": 1136,
              "w": 2028,
              "x": 0,
              "y": 0
            },
            {
              "h": 1326,
              "w": 1326,
              "x": 0,
              "y": 0
            },
            {
              "h": 1326,
              "w": 1163,
              "x": 78,
              "y": 0
            },
            {
              "h": 1326,
              "w": 663,
              "x": 328,
              "y": 0
            },
            {
              "h": 1326,
              "w": 2028,
              "x": 0,
              "y": 0
            }
          ],
          "height": 1326,
          "width": 2028
        },
        "sizes": {
          "large": {
            "h": 1326,
            "w": 2028
          }
        },
        "type": "photo",
        "url": "https://t.co/84a5YsXXR0"
      }
    ]
  },
  "card": null,
  "place": {},
  "entities": {},
  "quoted_tweet": null,
  "retweeted_tweet": null,
  "isLimitedReply": false,
  "article": null
}