{"id":244577,"date":"2026-09-28T17:05:15","date_gmt":"2026-09-28T22:05:15","guid":{"rendered":"https:\/\/lifeboat.com\/blog\/2026\/09\/unslopping-ai"},"modified":"2026-09-28T17:05:15","modified_gmt":"2026-09-28T22:05:15","slug":"unslopping-ai","status":"publish","type":"post","link":"https:\/\/lifeboat.com\/blog\/2026\/09\/unslopping-ai","title":{"rendered":"Unslopping AI"},"content":{"rendered":"<p><a class=\"aligncenter blog-photo\" href=\"https:\/\/lifeboat.com\/blog.images\/unslopping-ai2.jpg\"><\/a><\/p>\n<p>Claim: we\u2019ve solved the AI slop problem (!) \ud83d\udca9\ud83e\uddf9\u2728<\/p>\n<p>Blog post: <a href=\"https:\/\/facebookresearch.github.io\/RAM\/blogs\/unslop\/\">https:\/\/facebookresearch.github.io\/RAM\/blogs\/unslop\/<\/a> by: Jason Weston.<\/p>\n<p>Key idea: take *expert* human writing and learn rubrics that find the gap between experts and models. Train with those rubrics.<\/p>\n<p>We train with RL-XAR (RL with eXpert Aligned Rubrics) &amp; see large performance gains on writing scientific paper sections, Pulitzer prize novel continuations and high quality Wikipedia pages.<\/p>\n<p>First: The failure of standard LLM Judgements \ud83d\udc80<\/p>\n<p>On paper writing tasks, strong judges (GPT-5.6 or Opus-4.8) think current \u2018slop\u2019 models are better than humans on selected high quality papers (using either pairwise, or using standard rubrics).<\/p>\n<p>Our method can learn rubrics where the human is considered better by the grader (right in fig) \u2013 the key to training.<\/p>\n<div class=\"more-link-wrapper\"> <a class=\"more-link\" href=\"https:\/\/lifeboat.com\/blog\/2026\/09\/unslopping-ai\">Continue reading \u201cUnslopping AI\u201d | &gt;<\/a><\/div>\n","protected":false},"excerpt":{"rendered":"<p>Claim: we\u2019ve solved the AI slop problem (!) \ud83d\udca9\ud83e\uddf9\u2728 Blog post: https:\/\/facebookresearch.github.io\/RAM\/blogs\/unslop\/ by: Jason Weston. Key idea: take *expert* human writing and learn rubrics that find the gap between experts and models. Train with those rubrics. We train with RL-XAR (RL with eXpert Aligned Rubrics) &amp; see large performance gains on writing scientific paper sections, [\u2026]<\/p>\n","protected":false},"author":709,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[],"class_list":["post-244577","post","type-post","status-publish","format-standard","hentry","category-robotics-ai"],"_links":{"self":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/244577","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/users\/709"}],"replies":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/comments?post=244577"}],"version-history":[{"count":0,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/244577\/revisions"}],"wp:attachment":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/media?parent=244577"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/categories?post=244577"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/tags?post=244577"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}