I just added a link to the firefox magazine using Fedia’s + button. Should I be doing something different? Maybe my system time is off?

Sure, but that’s after years of researching and tweaking the mitigation, and doesn’t cover other OSes. I certainly agree that some folks made it sound a lot more dire than it ever was, though.

Sometimes the hardware or software configurations of the machines running the tests changes, or a bug in the test harness itself is fixed, which can skew all of the results at once.

To be fair, given how many high-profile data breaches there have been since, and the fact that CPU performance was significantly affected, and how much Intel has struggled to remain competitive, I’m not sure we were all that far off from any exaggerations.

WebTransport is pretty exciting for folks who want to avoid the extra hassle of working with the lower-level WebSockets API, and it can also lead to better performance, so it could be a game-changer for some things; a surprising number of sites are using WebSockets, including multiplayer games.

Early Hints and the speculative connection improvements are also nice to have, as they should keep Firefox page loading performance competitive as more and more sites support them.

DNS over Oblivious HTTP is also a pretty important privacy improvement for DoH, so it’s great to see that it has a negligible performance impact and we can move ahead with it.

There was also a captive portal fix, and some of the new contributors helped out with interop2023 fixes, which is very cool to see.

And the planning for off-main-thread networking is very likely to lead to performance wins down the line, if not also security wins, so it’s worth keeping an eye on as well.