[nottorp]

Q1: Why would round trip latency matter so much when the modern web 45.0 adds so much overhead?

Q2: Why don't we just force all the SV techbros to test their sites - AND their browsers - on a $200 notebook with eMMC storage from 2 years ago?

[noitpmeder]

Because most sites aren't interested in their users who use bad hardware as a market share. The most attractive client base are the ones with money to spend.

[est31]

Rich people tend to be old, and not all, but many old people have bad hardware, even if they are rich. Meanwhile, 87% of US teenagers have iPhones.

[hutzlibu]

Fair point, but even high end users have bad connection sometimes.

[strken]

It matters a lot. I just spent three days on a 130kbps 1000+ms latency high packet loss connection, and RTT was one of the problems that made lots of sites I tried (practically) unusable, due to launching dependent queries in series without any retries. The other big one was gigantic interaction-blocking assets served from CDNs with aggressive timeouts.

SV techbros are a fun pinata to smash, but that's a load of bullshit intended to conflate individual workers (easily hateable and zero power to fix anything) with ad-tech/paywalls/lead acquisition/overoptimisation driven by profit-seeking (too diffuse to hate, too little understood for lawmakers to regulate effectively...yet), and make slow websites seem like a technical or competency problem instead of an auctioning-your-eyeballs-and-spamming-you-with-modals.

[nottorp]

Pretty sure the guilt is shared. The profit seeking personal data peddlers don't choose overengineered solutions that only make sense at google scale.

I also have a feeling hardware designers could use being forced to live somewhere that's not a dust less constant temperature air conditioned office in California, but that's another matter for a different rant.

[midoridensha]

The SV techbros should have to test their sites on 2010-era notebooks running through two VPNs connecting through Australia and Finland.