Jake Brukhman @jbrukh · 02 Sep 2026
No one wants the success of useful proof of work, especially for AI inference, more than I do. But I think strong verification will be a non-negotiable component of that. An independent empirical study (Abhinaba Basu, arXiv:2606.04819, “The Usefulness Gap in
8 434Views
58Likes
1Reposts
20Replies
3Quotes
19Bookmarks
Is that a lot?
1.35×vs this author's median6 251 views is typical
60Percentile for this authorof 5 recent posts
4.33×vs 10K–100K median1 948 views is typical
15.76%Reachviews ÷ followers
0.97%Engagement rateof viewers reacted