It's interesting seeing the predictions come true. I was slightly nervous when Elon fired 90% of twitter staff and the site kept working. If stuff never broke, then empirically, firing 90% of your staff seems to be a good idea.
But now stuff is breaking each week, and it's evident that maybe it wasn't the best idea.
> I was slightly nervous when Elon fired 90% of twitter staff and the site kept working. If stuff never broke, then empirically, firing 90% of your staff seems to be a good idea.
Things continuing to work doesn't mean that firing 90% of the staff is a good idea. Decently built site will remain up even if nobody's around to take care of it.
Actually changing things is where you see an actual impact of the layoffs. There have been no major changes to the platform, only cosmetics, and they managed to fuck up every single one of them ("views" moved like six times and is still not aligned properly, black on black text, translating tweets is gone for weeks...).
I'm not the one to usually complain about UI changes, but for fuck's sake, every single week things are still the same functionality-wise, but something in the interface gets moved just because.
> There have been no major changes to the platform, only cosmetics, and they managed to fuck up every single one of them
I've been trying to explain this to some of my Musk megafan acquaintances from when Musk was making lots of noise of about adding various "features", when they (the acquaintances) were fawning over "zomg look at the improvement in new product development since Elon took over!!11one".
I think there is very little competition for Twitter except TikTok and Instagram. But they have different audience profiles. So it is safe to do nothing on average but you won't grow...
As a Twitter addict, I can remember at least 5 times total outages within 2 years before Musk took over.
There are also some serious privacy accidents that surprisingly never got any attention in Western.
For example, about 3 years ago, Twitter bugged that everyone's private lists (titles, descriptions) are publicly visible for at least half a day. It was a huge shit storm in Japanese communities but I can't find any article mentioning it in English media.
It's difficult to judge based solely on experience, but it's definitely felt worse to me recently. I don't remember having so many problems with basic functionality before — things like tweets showing up in one place and not others. Tbh, many of the problems seem to be related to poor cache management.
Well, hopefully they stop breaking “the entire reason people come to twitter”. At least when tweeting broke last week we could still scroll the timeline.
Thanks for pointing out that privacy blunder. I hadn’t heard of that. It’s interesting that stories like that end up spreading in the east but not in the west.
How long were those 5 total outages? I wonder if this one will be average.
Total outage usually didn't last long, I believe there was at least one that's more than one hour.
What is more annoying is Twitter sometimes had highly disruptive bugs that often take much longer time to fix.
About half a year ago, there was a timeline bug that Twitter will constantly showing you three random tweets repeatedly after every refresh. It took them days to fix it.
Also - keeping the stuff running, and making sure nothing ever breaks, is a different problem when you just want to maintain the current state of the product and keep it stable, vs when you want to change it.
Many tech companies are optimising for having hundreds of new features and products developing on top of the current stack, and allowing quick iterations, taking bets on things that might or might not have the market for it.
You can fire all civil engineers in the country and bridges will not collapse immediately. But you won't get any new bridges built. Also - at some point you might learn which ones had structural issues hidden by maintenance.
Apart from this, that has left the timeline completely broken, has anyone else noticed that the timeline on mobile doesn't auto-update anymore? It's often out of date and I have to refresh it manually, which I didn't use to. Maybe a hardcore measure to save server capacity?
I was pleasantly surprised in the beginning of Musks reign, because a few weeks into it the World Cup came and went and Twitter didn't even flinch.
But since then it's been quite the shit-show from a stability point of view. I'm not a heavy user so I don't really care, but it is interesting how the initial pessimistic predictions had a sort of lag before being borne out.
I don't think that's at all surprising. Companies can run with a skeleton crew... for a little bit (most web services get through major holidays on a handful of on-call employees, say). But that only works for so long.
I think if you leave a product in a steady state, then you can keep a web service running for quite a while. The problem is when you want to add new features, which is a much higher risk activity.
you could run twitter indefinitely with a crew (25-50) of competent SRE for basically ever as long as they didn't roll out new features. Nobody should take that fact and conclude that firing your employees en masse is a good idea.
But now stuff is breaking each week, and it's evident that maybe it wasn't the best idea.