Viewing enriched Twitter post
Best possible task accuracy after fixing constraints on inference process? Tokens across all generations, depth of generation chain (“latency”), total context length and compute. PDR achieves pareto frontier: "draft in parallel → distill to a compact workspace → refine". https://t.co/jWoGu8EVdj
{
"media": [
{
"url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2000630338285089155/media_0.jpg?",
"media_url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2000630338285089155/media_0.jpg?",
"type": "photo",
"filename": "media_0.jpg"
}
],
"processed_at": "2025-12-16T14:43:40.454233",
"pipeline_version": "2.0"
}{
"type": "tweet",
"id": "2000630338285089155",
"url": "https://x.com/anirudhg9119/status/2000630338285089155",
"twitterUrl": "https://twitter.com/anirudhg9119/status/2000630338285089155",
"text": "Best possible task accuracy after fixing constraints on inference process? Tokens across all generations, depth of generation chain (“latency”), total context length and compute. \n\nPDR achieves pareto frontier: \"draft in parallel → distill to a compact workspace → refine\". https://t.co/jWoGu8EVdj",
"source": "Twitter for iPhone",
"retweetCount": 21,
"replyCount": 5,
"likeCount": 158,
"quoteCount": 0,
"viewCount": 18034,
"createdAt": "Mon Dec 15 18:13:37 +0000 2025",
"lang": "en",
"bookmarkCount": 130,
"isReply": false,
"inReplyToId": null,
"conversationId": "2000630338285089155",
"displayTextRange": [
0,
277
],
"inReplyToUserId": null,
"inReplyToUsername": null,
"author": {
"type": "user",
"userName": "anirudhg9119",
"url": "https://x.com/anirudhg9119",
"twitterUrl": "https://twitter.com/anirudhg9119",
"id": "2816636344",
"name": "Anirudh Goyal",
"isVerified": false,
"isBlueVerified": false,
"verifiedType": null,
"profilePicture": "https://pbs.twimg.com/profile_images/564397385439141888/8WnILtyS_normal.jpeg",
"coverPicture": "",
"description": "",
"location": "London, UK",
"followers": 5796,
"following": 530,
"status": "",
"canDm": true,
"canMediaTag": true,
"createdAt": "Wed Oct 08 19:47:25 +0000 2014",
"entities": {
"description": {
"urls": []
},
"url": {}
},
"fastFollowersCount": 0,
"favouritesCount": 6828,
"hasCustomTimelines": true,
"isTranslator": false,
"mediaCount": 90,
"statusesCount": 946,
"withheldInCountries": [],
"affiliatesHighlightedLabel": {},
"possiblySensitive": false,
"pinnedTweetIds": [],
"profile_bio": {
"description": "Thinking about thinking. \n\nSenior Staff RS, Spent time at @Berkeley_EECS, @MPI_IS, @GoogleDeepMind Gemini ♊.",
"entities": {
"description": {
"user_mentions": [
{
"id_str": "0",
"indices": [
58,
72
],
"name": "",
"screen_name": "Berkeley_EECS"
},
{
"id_str": "0",
"indices": [
74,
81
],
"name": "",
"screen_name": "MPI_IS"
},
{
"id_str": "0",
"indices": [
83,
98
],
"name": "",
"screen_name": "GoogleDeepMind"
}
]
},
"url": {
"urls": [
{
"display_url": "anirudh9119.github.io",
"expanded_url": "https://anirudh9119.github.io/",
"indices": [
0,
23
],
"url": "https://t.co/yNjmNYUH0Z"
}
]
}
}
},
"isAutomated": false,
"automatedBy": null
},
"extendedEntities": {
"media": [
{
"allow_download_status": {
"allow_download": true
},
"display_url": "pic.twitter.com/jWoGu8EVdj",
"expanded_url": "https://twitter.com/anirudhg9119/status/2000630338285089155/photo/1",
"ext_media_availability": {
"status": "Available"
},
"features": {
"large": {
"faces": [
{
"h": 229,
"w": 229,
"x": 12,
"y": 1050
},
{
"h": 334,
"w": 334,
"x": 1033,
"y": 304
}
]
},
"orig": {
"faces": [
{
"h": 229,
"w": 229,
"x": 12,
"y": 1050
},
{
"h": 334,
"w": 334,
"x": 1033,
"y": 304
}
]
}
},
"id_str": "2000628979863035910",
"indices": [
278,
301
],
"media_key": "3_2000628979863035910",
"media_results": {
"id": "QXBpTWVkaWFSZXN1bHRzOgwAAQoAARvDqXUYFqAGCgACG8OqsWBaMYMAAA==",
"result": {
"__typename": "ApiMedia",
"id": "QXBpTWVkaWE6DAABCgABG8OpdRgWoAYKAAIbw6qxYFoxgwAA",
"media_key": "3_2000628979863035910"
}
},
"media_url_https": "https://pbs.twimg.com/media/G8OpdRgWoAYclti.jpg",
"original_info": {
"focus_rects": [
{
"h": 794,
"w": 1418,
"x": 0,
"y": 0
},
{
"h": 1418,
"w": 1418,
"x": 0,
"y": 0
},
{
"h": 1617,
"w": 1418,
"x": 0,
"y": 0
},
{
"h": 1784,
"w": 892,
"x": 0,
"y": 0
},
{
"h": 1784,
"w": 1418,
"x": 0,
"y": 0
}
],
"height": 1784,
"width": 1418
},
"sizes": {
"large": {
"h": 1784,
"w": 1418
}
},
"type": "photo",
"url": "https://t.co/jWoGu8EVdj"
}
]
},
"card": null,
"place": {},
"entities": {},
"quoted_tweet": {
"type": "tweet",
"id": "2000581380733030703",
"url": "https://x.com/dair_ai/status/2000581380733030703",
"twitterUrl": "https://twitter.com/dair_ai/status/2000581380733030703",
"text": "NEW Research from Meta Superintelligence Labs and collaborators.\n\nThe default approach to improving LLM reasoning today remains extending chain-of-thought sequences.\n\nLonger reasoning traces aren't always better. Longer traces conflate reasoning depth with sequence length and inherit long-context failure modes.\n\nThis new research introduces Parallel-Distill-Refine (PDR), a framework that treats LLMs as improvement operators rather than single-pass reasoners.\n\nInstead of one long reasoning chain, PDR operates in phases:\n- Generate diverse drafts in parallel.\n- Distill them into a bounded textual workspace.\n- Refine conditioned on this workspace.\n- Repeat.\n\nContext length becomes controllable via degree of parallelism, no longer conflated with total tokens generated. The model accumulates wisdom across rounds through compact summaries rather than replaying full histories.\n\nOn AIME 2024, PDR achieves 93.3% accuracy compared to 79.4% for standard long chain-of-thought at matched latency budgets.\n\nFor o3-mini at 49k effective tokens, accuracy improves from 76.9% (Long CoT) to 86.7% (PDR), a 9.8 percentage point gain.\n\nPDR also achieves the same accuracy as sequential refinement with 2.57x smaller sequential budget by converting parallel compute into accuracy without lengthening per-call context.\n\nThe researchers also trained an 8B model with operator-consistent RL to make training match the PDR inference interface. Mixing standard and operator RL yields an additional 5% improvement on both AIME benchmarks.\n\nBounded memory iteration can substitute for long reasoning traces while holding latency fixed. Strategic parallelism and distillation is shown to beat brute-force sequence extension.\n\nPaper: https://t.co/EviERpmTu7\n\nLearn to build effective AI Agents in our academy: https://t.co/zQXQt0PMbG",
"source": "Twitter for iPhone",
"retweetCount": 68,
"replyCount": 11,
"likeCount": 448,
"quoteCount": 5,
"viewCount": 56803,
"createdAt": "Mon Dec 15 14:59:04 +0000 2025",
"lang": "en",
"bookmarkCount": 450,
"isReply": false,
"inReplyToId": null,
"conversationId": "2000581380733030703",
"displayTextRange": [
0,
277
],
"inReplyToUserId": null,
"inReplyToUsername": null,
"author": {
"type": "user",
"userName": "dair_ai",
"url": "https://x.com/dair_ai",
"twitterUrl": "https://twitter.com/dair_ai",
"id": "889050642903293953",
"name": "DAIR.AI",
"isVerified": false,
"isBlueVerified": true,
"verifiedType": null,
"profilePicture": "https://pbs.twimg.com/profile_images/1643277398522187778/31dedbLo_normal.jpg",
"coverPicture": "https://pbs.twimg.com/profile_banners/889050642903293953/1742055232",
"description": "",
"location": "",
"followers": 83444,
"following": 1,
"status": "",
"canDm": true,
"canMediaTag": true,
"createdAt": "Sun Jul 23 09:12:45 +0000 2017",
"entities": {
"description": {
"urls": []
},
"url": {}
},
"fastFollowersCount": 0,
"favouritesCount": 3894,
"hasCustomTimelines": true,
"isTranslator": false,
"mediaCount": 90,
"statusesCount": 2668,
"withheldInCountries": [],
"affiliatesHighlightedLabel": {},
"possiblySensitive": false,
"pinnedTweetIds": [
"2000581380733030703"
],
"profile_bio": {
"description": "Democratizing AI research, education, and technologies.",
"entities": {
"description": {},
"url": {
"urls": [
{
"display_url": "dair.ai",
"expanded_url": "https://www.dair.ai/",
"indices": [
0,
23
],
"url": "https://t.co/lkqPZtMmfU"
}
]
}
}
},
"isAutomated": false,
"automatedBy": null
},
"extendedEntities": {
"media": [
{
"display_url": "pic.twitter.com/UZX1iHAkOQ",
"expanded_url": "https://twitter.com/dair_ai/status/2000581380733030703/photo/1",
"ext_media_availability": {
"status": "Available"
},
"features": {
"large": {
"faces": [
{
"h": 229,
"w": 229,
"x": 12,
"y": 1050
},
{
"h": 334,
"w": 334,
"x": 1033,
"y": 304
}
]
},
"orig": {
"faces": [
{
"h": 229,
"w": 229,
"x": 12,
"y": 1050
},
{
"h": 334,
"w": 334,
"x": 1033,
"y": 304
}
]
}
},
"id_str": "2000581376421277696",
"indices": [
278,
301
],
"media_key": "3_2000581376421277696",
"media_results": {
"id": "QXBpTWVkaWFSZXN1bHRzOgwAAQoAARvDfimN2jAACgACG8N+Ko7aUS8AAA==",
"result": {
"__typename": "ApiMedia",
"id": "QXBpTWVkaWE6DAABCgABG8N+KY3aMAAKAAIbw34qjtpRLwAA",
"media_key": "3_2000581376421277696"
}
},
"media_url_https": "https://pbs.twimg.com/media/G8N-KY3aMAA1nwq.jpg",
"original_info": {
"focus_rects": [
{
"h": 794,
"w": 1418,
"x": 0,
"y": 0
},
{
"h": 1418,
"w": 1418,
"x": 0,
"y": 0
},
{
"h": 1617,
"w": 1418,
"x": 0,
"y": 0
},
{
"h": 1784,
"w": 892,
"x": 0,
"y": 0
},
{
"h": 1784,
"w": 1418,
"x": 0,
"y": 0
}
],
"height": 1784,
"width": 1418
},
"sizes": {
"large": {
"h": 1784,
"w": 1418
}
},
"type": "photo",
"url": "https://t.co/UZX1iHAkOQ"
}
]
},
"card": null,
"place": {},
"entities": {
"urls": [
{
"display_url": "arxiv.org/abs/2510.01123",
"expanded_url": "https://arxiv.org/abs/2510.01123",
"indices": [
1719,
1742
],
"url": "https://t.co/EviERpmTu7"
},
{
"display_url": "dair-ai.thinkific.com",
"expanded_url": "https://dair-ai.thinkific.com/",
"indices": [
1795,
1818
],
"url": "https://t.co/zQXQt0PMbG"
}
]
},
"quoted_tweet": null,
"retweeted_tweet": null,
"isLimitedReply": false,
"article": null
},
"retweeted_tweet": null,
"isLimitedReply": false,
"article": null
}