Comment by viraptor
18 hours ago
> After the massive X/Twitter layoffs (60-70% headcount culled) with X/Twitter still standing, this assumption was clearly proven false.
Twitter at the same time removed features to have fewer things to support. And didn't implement anything new (or really fix much) for ages. It's not the same service that was standing afterwards. And the "still standing" ignores the part where they started serving empty timelines, repeated messages from broken paging, broke 2fa for days, messed up whole continent access, etc. etc. They survived (and still had fewer problems than I expected), but it wasn't smooth at all - hardly a success too.
The search functionality has been mostly broken (in several, overlapping ways) for several years now
The thing is, as an outsider to Twitter but with 20 years of experience doing software dev including some time at internet scale web and mobile, I don't think that the basic "fetch a timeline" backend plus two front end apps and a web interface is that hard, a small team (<100 engineers) definitely could do that with modern cloud infrastructure. But that's not what the Twitter product was. We've just described nothing more than the bait to lure the product, which was advertisers running ads.
Most of the effort in the original Twitter- engineering and everything else- was about getting advertising revenue. That meant 1) Having good data mining to identify user interests to match ads against 2) Having a strong user experience like Meta Ads or AdSense for the ad buyers 3) Keeping the conversations such that advertisers wanted to be associated with it, both automated and manual censorship 4) Having good relationships with advertisers, both large clients and agencies
That was where the majority of Twitter's (dev and non-engineering both) effort was going, to bringing in the revenue from advertisers. When Elon Musk purchased Twitter advertising fell dramatically immediately, at basically the same time he gutted all of the people doing the advertising. That was why he tried "buying a blue-check" and so many premium features, because he got rid of all the infrastructure necessary to have a serious ads platform. And premium doesn't work, of course, as anyone with experience in the Internet world could have told him. Which is why the value of the company- and its revenues- have declined so dramatically since the acquisition.
Bluesky is basically doing the same thing as X right down to also not running ads, which is how they also manage to run on a small team. Last I checked they'd raised less than 20m, and have basically no revenue, so they are able to operate very lean. It's for the same reason that Twitter is a lot smaller now: ads are a huge engineering and non-tech effort. As Alphabet and Meta remind us, it can be insanely profitable, but you need a lot of people to get it right.
[flagged]