Viewing enriched Twitter post
can we trust decision models for evals? @HamelHusain and I tested top decision models like Jev, d1-ultra, clef, pplx-decider, openai decisions, etc. on our dataset we explored failure detection, probability calibration, threshold optimization, and the practical tradeoffs of using them for evals. give it a read: https://t.co/q3LikO8uQ9
{
"media": [
{
"url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2112915743511904569/media_0.jpg",
"media_url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2112915743511904569/media_0.jpg",
"type": "photo",
"filename": "media_0.jpg"
},
{
"url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2112915743511904569/media_1.jpg",
"media_url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2112915743511904569/media_1.jpg",
"type": "photo",
"filename": "media_1.jpg"
}
],
"processed_at": "2026-10-21T15:04:00.783161",
"pipeline_version": "2.0"
}{
"type": "tweet",
"id": "2112915743511904569",
"url": "https://x.com/doesdatmaksense/status/2112915743511904569",
"twitterUrl": "https://twitter.com/doesdatmaksense/status/2112915743511904569",
"text": "can we trust decision models for evals?\n\n@HamelHusain and I tested top decision models like Jev, d1-ultra, clef, pplx-decider, openai decisions, etc. on our dataset\n\nwe explored failure detection, probability calibration, threshold optimization, and the practical tradeoffs of using them for evals.\n\ngive it a read: https://t.co/q3LikO8uQ9",
"source": "",
"retweetCount": 1,
"replyCount": 0,
"likeCount": 6,
"quoteCount": 0,
"viewCount": 357,
"createdAt": "Wed Oct 21 14:35:43 +0000 2026",
"lang": "en",
"bookmarkCount": 2,
"isReply": false,
"inReplyToId": null,
"conversationId": "2112915743511904569",
"displayTextRange": [
0,
276
],
"inReplyToUserId": null,
"inReplyToUsername": null,
"author": {
"type": "user",
"userName": "doesdatmaksense",
"url": "https://x.com/doesdatmaksense",
"twitterUrl": "https://twitter.com/doesdatmaksense",
"id": "1060588105361633280",
"name": "Antaripa Saha",
"isVerified": false,
"isBlueVerified": false,
"verifiedType": null,
"profilePicture": "https://pbs.twimg.com/profile_images/1909265588557398016/WD3zxDCa_normal.jpg",
"coverPicture": "https://pbs.twimg.com/profile_banners/1060588105361633280/1724836968",
"description": "",
"location": "",
"followers": 15448,
"following": 443,
"status": "",
"canDm": true,
"canMediaTag": true,
"createdAt": "Thu Nov 08 17:41:23 +0000 2018",
"entities": {
"description": {
"urls": []
},
"url": {}
},
"fastFollowersCount": 0,
"favouritesCount": 26324,
"hasCustomTimelines": true,
"isTranslator": false,
"mediaCount": 513,
"statusesCount": 5418,
"withheldInCountries": [],
"affiliatesHighlightedLabel": {},
"possiblySensitive": false,
"pinnedTweetIds": [
"1842811263296176132"
],
"profile_bio": {
"description": "consulting companies in evals, search, retrieval",
"entities": {
"description": {},
"url": {
"urls": [
{
"display_url": "antaripasaha.com",
"expanded_url": "http://antaripasaha.com",
"indices": [
0,
23
],
"url": "https://t.co/K59fHS4rWm"
}
]
}
}
},
"isAutomated": false,
"automatedBy": null
},
"extendedEntities": {
"media": [
{
"allow_download_status": {
"allow_download": true
},
"display_url": "pic.x.com/iQb1eHi82I",
"expanded_url": "https://x.com/doesdatmaksense/status/2112915743511904569/photo/1",
"ext_master_playlist_only": [],
"ext_media_availability": {
"status": "Available"
},
"ext_playlists": [],
"features": {
"large": {
"faces": [
{
"h": 58,
"w": 58,
"x": 62,
"y": 12
}
]
},
"orig": {
"faces": [
{
"h": 58,
"w": 58,
"x": 62,
"y": 12
}
]
}
},
"id_str": "2112914049214369792",
"indices": [
277,
300
],
"media_key": "3_2112914049214369792",
"media_results": {
"id": "QXBpTWVkaWFSZXN1bHRzOgwAAQoAAR1SlCIQWtAACgACHVKVrIxbgTkAAA==",
"result": {
"__typename": "ApiMedia",
"id": "QXBpTWVkaWE6DAABCgABHVKUIhBa0AAKAAIdUpWsjFuBOQAA",
"media_key": "3_2112914049214369792"
}
},
"media_url_https": "https://pbs.twimg.com/media/HVKUIhBa0AAAiKx.jpg",
"original_info": {
"focus_rects": [
{
"h": 776,
"w": 1386,
"x": 0,
"y": 580
},
{
"h": 1356,
"w": 1356,
"x": 0,
"y": 0
},
{
"h": 1356,
"w": 1189,
"x": 0,
"y": 0
},
{
"h": 1356,
"w": 678,
"x": 0,
"y": 0
},
{
"h": 1356,
"w": 1386,
"x": 0,
"y": 0
}
],
"height": 1356,
"width": 1386
},
"sizes": {
"large": {
"h": 1356,
"w": 1386
}
},
"type": "photo",
"url": "https://t.co/iQb1eHi82I"
}
]
},
"card": null,
"place": {},
"entities": {
"hashtags": [],
"symbols": [],
"urls": [
{
"display_url": "hamel.dev/blog/posts/dec…",
"expanded_url": "https://hamel.dev/blog/posts/decision-model-evals/",
"indices": [
316,
339
],
"url": "https://t.co/q3LikO8uQ9"
}
],
"user_mentions": [
{
"id_str": "825766640",
"indices": [
41,
53
],
"name": "Hamel Husain",
"screen_name": "HamelHusain"
}
]
},
"quoted_tweet": null,
"retweeted_tweet": null,
"isLimitedReply": false,
"communityInfo": null,
"article": null
}