Not really giving a f*ck about pagespeed and lighthouse reports.

And overthinking styles…

Some like to overengineer page speed too... I guess somewhere between 70-85 is great, the rest takes way more effort lol
My current employer's website is around 20-30... thanks to the CMS Of Doom...

You get most bang for the buck by fixing outright stupid shit.

Take devrant.com: why is there no http/2 support, why no next-gen image support with the picture tag, why are the image dimensions missing, and why no caching expires for static resources? It's just dumb.

@PonySlaystation Below about 70, it’s all kind of moot. The user might think “wow, that was fast” but really all they care about is that it isn’t slow.

I’ve managed some amazing response and load times, and it just felt like bragging rights and nothing else.

Also (and kind of a tangent): at scale, the server load per client matters more anyway since it directly affects hosting costs, so you’ll find yourself offloading more to the client at this point. This usually doesn’t affect load times too much, however, since most of what you can offload happens well after the initial page load.

@Root Makes sense...
Sadly, I'm not experienced at scale yet

Lighhouse/pagespeed give contradictory and counterproductive advice often and 100% green best practice is like 100% code coverage test runs: still no guarantee for real life usability!!