Speculation Rules Load Pages By Thinking
summary
Claim: speculation rules make multi-page navigation feel instant by prefetching or pre-rendering pages ahead of clicks. Demo shows a link drop from ~267 ms to 63 ms when ready. API offers conservative eagerness (mouse down ~100–200 ms head start), moderate (hover ~200 ms), and immediate (full speculative cost). Works today in Chromium only, cache lasts about five minutes, scriptable via JavaScript for dynamic targeting, and pairs well with dictionary compression; SPAs and WebKit iOS remain unsupported.
AI-written summary · about
chapters (17)
- 0:00Speculation rules intro
- 1:54Why performance matters
- 3:43Traditional page loading
- 5:29What speculation rules fix
- 7:37Speculation rules API
- 9:03Eagerness levels explained
- 11:53Accuracy vs head start
- 13:37Measuring navigation perf
- 14:40DevTools for speculation rules
- 15:01Live demo
- 17:52Impact on navigation speed
- 18:34Browser support
- 19:17When not to use
- 20:42Common pitfalls
- 21:46Real-world implementations
- 22:31Dictionary compression
- 23:55Next steps
related talks
30mBlazing Fast Page Navigation With Speculation Rule
46mSpeculations about Web Performance
14mMaking the Web Instant
28mBetter than loading fast… is loading instantly
39mWhat’s new in performance?
36mBreaking The Bloat: Achieving The Fastest Navigations on The Web
47mMore Than You Wanted to Know About Resource Hints
28mThe Browser Hackers Guide To Instantly Loading Everything