Phillip Isola @phillip_isola · 07 Jul 2026
An update to the Neural Thicket paper: We avoid the K× inference cost of RandOpt ensembling and make RandOpt iterative: a single model with strong performance. Like policy gradients, but with search in weight space and stepwise supervision. Like ES, but with updates through https://t.co/GGtqjTqaaS
30 147Views
120Likes
17Reposts
5Replies
2Quotes
0Bookmarks
Is that a lot?
2.15×vs this author's median14 021 views is typical
73Percentile for this authorof 11 recent posts
25.2×vs 10K–100K median1 195 views is typical
139.24%Reachviews ÷ followers
0.48%Engagement rateof viewers reacted