tweetindex

Marvin Tong @marvin_tong · 04 Sep 2026

modern inference’s quality-efficiency frontier increasingly comes down to cache hit rate. Infra should be built around maximizing it. Encryption makes cache reuse extremely hard in private inference—but we’re breaking through that barrier. https://t.co/EuYTZ1ZL3L
836Views
15Likes
1Reposts
0Replies
0Quotes
1Bookmarks

Is that a lot?

0.61×vs this author's median1 372 views is typical
0Percentile for this authorof 5 recent posts
0.66×vs 10K–100K median1 266 views is typical
5.72%Reachviews ÷ followers
1.91%Engagement rateof viewers reacted

Compare with the benchmark table →

Open on X →