🐦 Twitter Post Details

Viewing enriched Twitter post

@gowthami_s

Maybe hot take - I’ve read a bunch of RL for image generation papers over last few months and honestly it’s been pretty disappointing. All of them are variations of GRPO and all of them are incremental algo changes. Tbh most of these don’t even matter for large models + large group size with good reward model setting. I see most grad students are still optimizing their projects for reviewers rather than genuinely trying to solve some of the real problems in visual generation. For example - the biggest alpha in my eyes would’ve been an artifact detection model - not just for mangled limbs, most image models produce far more artifacts which are hard to quantitatively measure, but I haven’t seen a single research paper or a model on this. So my 2c, if you are a grad student targeting a job in industry, target for impact, no one cares about your third CVPR paper, one is enough to get you in the door, building a model industry actually uses gives you all the leverage. Impact > Publications. 🫳🎤

📊 Media Metadata

{
  "score": 0.42,
  "score_components": {
    "author": 0.09,
    "engagement": 0.0,
    "quality": 0.12,
    "source": 0.135,
    "nlp": 0.05,
    "recency": 0.025
  },
  "scored_at": "2026-04-13T09:23:42.007263",
  "import_source": "api_import",
  "source_tagged_at": "2026-04-13T09:23:42.007272",
  "enriched": true,
  "enriched_at": "2026-04-13T09:23:42.007274"
}

🔧 Raw API Response

{
  "type": "tweet",
  "id": "2043562476059627967",
  "url": "https://x.com/gowthami_s/status/2043562476059627967",
  "twitterUrl": "https://twitter.com/gowthami_s/status/2043562476059627967",
  "text": "Maybe hot take - I’ve read a bunch of RL for image generation papers over last few months and honestly it’s been pretty disappointing. All of them are variations of GRPO and all of them are incremental algo changes. Tbh most of these don’t even matter for large models + large group size with good reward model setting. \n\nI see most grad students are still optimizing their projects for reviewers rather than genuinely trying to solve some of the real problems in visual generation. For example - the biggest alpha in my eyes would’ve been an artifact detection model - not just for mangled limbs, most image models produce far more artifacts which are hard to quantitatively measure, but I haven’t seen a single research paper or a model on this. \n\nSo my 2c, if you are a grad student targeting a job in industry, target for impact, no one cares about your third CVPR paper, one is enough to get you in the door, building a model industry actually uses gives you all the leverage. Impact > Publications. 🫳🎤",
  "source": "Twitter for iPhone",
  "retweetCount": 1,
  "replyCount": 4,
  "likeCount": 81,
  "quoteCount": 0,
  "viewCount": 3795,
  "createdAt": "Mon Apr 13 05:30:36 +0000 2026",
  "lang": "en",
  "bookmarkCount": 27,
  "isReply": false,
  "inReplyToId": null,
  "conversationId": "2043562476059627967",
  "displayTextRange": [
    0,
    276
  ],
  "inReplyToUserId": null,
  "inReplyToUsername": null,
  "author": {
    "type": "user",
    "userName": "gowthami_s",
    "url": "https://x.com/gowthami_s",
    "twitterUrl": "https://twitter.com/gowthami_s",
    "id": "3143049392",
    "name": "Gowthami",
    "isVerified": false,
    "isBlueVerified": true,
    "verifiedType": null,
    "profilePicture": "https://pbs.twimg.com/profile_images/1937677147390533632/UV9auMb9_normal.jpg",
    "coverPicture": "https://pbs.twimg.com/profile_banners/3143049392/1622514574",
    "description": "",
    "location": "San Francisco, CA",
    "followers": 10261,
    "following": 995,
    "status": "",
    "canDm": true,
    "canMediaTag": true,
    "createdAt": "Tue Apr 07 07:20:23 +0000 2015",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {}
    },
    "fastFollowersCount": 0,
    "favouritesCount": 15090,
    "hasCustomTimelines": true,
    "isTranslator": false,
    "mediaCount": 403,
    "statusesCount": 3716,
    "withheldInCountries": [],
    "affiliatesHighlightedLabel": {},
    "possiblySensitive": false,
    "pinnedTweetIds": [],
    "profile_bio": {
      "description": "Generative modeling @theworldlabs | SGD-ing on life | Rants, Memes my own.",
      "entities": {
        "description": {
          "hashtags": [],
          "symbols": [],
          "urls": [],
          "user_mentions": [
            {
              "id_str": "0",
              "indices": [
                20,
                33
              ],
              "name": "",
              "screen_name": "theworldlabs"
            }
          ]
        },
        "url": {
          "urls": [
            {
              "display_url": "somepago.github.io",
              "expanded_url": "https://somepago.github.io",
              "indices": [
                0,
                23
              ],
              "url": "https://t.co/eRF3iwdBQJ"
            }
          ]
        }
      }
    },
    "isAutomated": false,
    "automatedBy": null
  },
  "extendedEntities": {},
  "card": null,
  "place": {},
  "entities": {
    "hashtags": [],
    "symbols": [],
    "urls": [],
    "user_mentions": []
  },
  "quoted_tweet": null,
  "retweeted_tweet": null,
  "isLimitedReply": false,
  "communityInfo": null,
  "article": null
}