If they were so essential I'd expect the service to have crashed and burned long ago and never came back up.
Trading off hundreds (thousands?) of people at US salaries for only a 10% impact in uptime is a sane business decision.
Having this many people working on a system that was built and "finished" 10 years ago (there have been no major feature updates to the core Twitter experience) in a decade, it means:
1) the system is built and working well but the engineers continue building complexity for complexity's sake in order to justify their continued employment and/or polish their resume (very common in any kind of large tech company with no clear product that people pay for - I call these "engineering playgrounds")
2) if the system is not actually working and needs hundreds/thousands of engineers to tend to it, then those people are incompetent and should be let go either way
> If they were so essential I'd expect the service to have crashed and burned long ago and never came back up.
False dichotomy. There is a vast space between continuing to run normally and turning into a smoking crater overnight. This includes brief outages, "secondary" functionality being broken for longer, slowdowns, inability to handle spikes or add new features, etc. All things we've actually seen with Twitter BTW. No one of them is enough to kill the site outright, but people do notice when a site is flaky. It erodes trust, and they seek alternatives.
The fact that Twitter hasn't experienced a more serious collapse since Musk's takeover is a testament to the quality of the engineering that was done before. But there's little difference between a sudden death and a slower one, in the end. If you want to rail against "perf-review driven engineering" or churn for churn's sake, you'd do better to aim your invective at Meta or Google instead of kicking people who are already down.
Maybe some of the people he fired could have helped him understand how not to lose half of advertisers or all the people put off by the presence of actual Nazis who will soon be unblockable and if they commit 8 bucks a month boosted in threads. Also if he was going to lose people maybe it should have been more of the low performers rather than high performers who didn't want to come into the office.
It seems clear to me already that the impact is more than a 10% impact on uptime and a lot of that value lost is going to be evident over years not immediately. EVERY software product on earth could fire the majority of its staff and be OK for a while.
If they were so essential I'd expect the service to have crashed and burned long ago and never came back up.
Trading off hundreds (thousands?) of people at US salaries for only a 10% impact in uptime is a sane business decision.
Having this many people working on a system that was built and "finished" 10 years ago (there have been no major feature updates to the core Twitter experience) in a decade, it means:
1) the system is built and working well but the engineers continue building complexity for complexity's sake in order to justify their continued employment and/or polish their resume (very common in any kind of large tech company with no clear product that people pay for - I call these "engineering playgrounds")
2) if the system is not actually working and needs hundreds/thousands of engineers to tend to it, then those people are incompetent and should be let go either way