Thinking Machines @thinkymachines · 27 Aug 2026
LLMs with scaffolds have lagged on text-to-SQL, a task that relies on human judgment. By folding expert judgment into every part of RLVR on Tinker, @maxYuxuanZhu and @ddkang (UIUC and Bridgwater) trained the first text-to-SQL model to beat the human mark. https://t.co/JxEQC31rX9
199 719Views
361Likes
46Reposts
4Replies
13Quotes
0Bookmarks
Is that a lot?
0.91×vs this author's median219 383 views is typical
25Percentile for this authorof 4 recent posts
9.12×vs 100K–1M median21 891 views is typical
109.60%Reachviews ÷ followers
0.21%Engagement rateof viewers reacted