Blazing Fast Page Navigation With Speculation Rule
summary
Chromium’s speculation rules API uses an inline JSON script to pre-render or prefetch targeted links (two slots, five-minute cache) with three eagerness modes: conservative (mouse-down ≈200ms), moderate (hover ≈400ms), immediate (scan-and-load). Real demos show LCP dropping from ~680ms to ~62ms and other pages from ~600ms to ~100ms (~90% improvements). Built-in devtools expose speculative loads and headers, and pairing the rules with dictionary compression yields large bandwidth savings (examples cited: ~85% and ~96% gains).
AI-written summary · about
chapters (6)
- 0:00Introduction to speculation rules
- 7:07How speculation rules work
- 15:32Demo of speculation rules in action
- 20:55Key takeaways and limitations
- 27:00Downsides and optimization tips
- 29:47Q&A and closing
related talks
25mSpeculation Rules Load Pages By Thinking
46mSpeculations about Web Performance
28mBetter than loading fast… is loading instantly
14mMaking the Web Instant
36mBreaking The Bloat: Achieving The Fastest Navigations on The Web
39mWhat’s new in performance?
12mBeyond fast with new performance features
01h 03mSpeed up your Web Application