I really don't get it. Twitter has five thousand full-time employees. I have to assume at least 20% of them are in product development in some capacity. What exactly are these people doing all day?
I can't point to a single notable product innovation they've had in years. And they continue to ignore the drumbeat of user asking for an edit button, and are completely unable to come up with any kind of reasonable solution to the abuse or bot problems.
I don't know Kayvon personally, but what exactly is wrong with Twitter that it's so bad and slow at product development? This should be #5000 on their list of things to do.
They're making tons of innovative new features for their customers, who are advertisers. People who write and read tweets aren't the customer, they're the product.
What exactly are these people doing all day? I can't point to a single notable product innovation they've had in years.
Maybe Twitter is less caught up on the SV "gotta change/break something user-facing today to justify my job" treadmill than other tech companies. Or maybe the changes they make are on the back end. It's been a long time since I've seen a fail whale.
Considering the global societal impact that Twitter has, I'm surprised it has only 5,000 employees.
Maintaining the infrastructure needed to deliver ads that are targeted based on real-time event streams collected from users' interaction with the website and app. Stuff like that.
That's my guess, anyway. It's the kind of thing that can keep a lot of developers very busy, but not something they'll be talking about much in public.
But is any of this really stuff that should take 1000 engineers to build? Tweet length? GIFs?
I don't question that those engineers are working hard. I'm sure they're not sitting around twiddling their thumbs. And I don't question that there aren't some genuinely hard, complicated problems to solve at Twitter, particularly around scaling and security. But from a structural perspective, I do still kind of wonder if companies the size of Twitter and Facebook aren't just an extended, very public example of Brooks's Law.
Anecdotally, I've been in large teams and small teams and I work equally hard in both environments. But even with the same amount of work, somehow, more stuff gets done and more products get shipped from the smaller teams.
There's a ton of infrastructure to build and maintain at Twitter. Yesterday's solutions don't scale and need replacing. Yesterday's "just get it done, we'll worry about cost (or quality, or operational burden) later" have happened and it's time to fix. Lather, rinse, repeat.
This post should give you an idea of scale at Twitter. When I worked there, I spent about 2 months focused just on creating software to help automate the Clos migration mentioned. And there are just tons of things like this that are constantly being worked on.
Having made the same observation as you, I always preferred the formulation in https://en.wikipedia.org/wiki/Ringelmann_effect. Although the Wikipedia article lists "loss of motivation" as a cause, it originally focussed more on coordination problems growing with group size–sort of a reverse Metcalf'.
In that sense it avoids the lazy cynicism of writing off whole groups of people as stupid or unmotivated (i. e. all of Dilbert). Instead, it's a starting point to consider how much we can still improve what's arguably humanity's claim to fame, the ability to cooperate.
None of those things strike as 1) important or 2) indicative of a high-performing product development team given their size. But we can agree to disagree!
Ads and security are both extremely important to Twitter. It's hard to imagine any product concern being more important than those two, other than perhaps uptime.
The only comment on HN more tiresome than "lol I can build twitter over a weekend" is a comment by a throwaway implicitly claiming that fixing the twitter product problems is easy.
A lot of features never make it past experimentation and never see the light of day. The lifecycle of these experiments can last months and multiple code paths on mobile clients have to be carefully maintained. Failure to catch regressions on experiments cause unclean data and prevent experiments from re-starting until the new client reaches app stores.
Could it be that they're just bogged down in fire fighting and paying off the tech debt accumulated over years?
(Genuine question, as someone who has never worked in >50 eng company, I always wonder how those companies tackle stuff like this.)
Jack has discussed this. It won't happen. Even in 5 minutes a tweet can reach MANY people, and then it would be able to be changed entirely?
Twitter won't get an edit feature, it just doesn't work for the platform. The only way I see it happening is as an 'undo' send feature like Gmail has. You can see the tweet for 15 seconds before it actually goes out. This has also been discussed as an issue since Twitter is supposed to be real-time, but maybe certain people wouldn't care and enable the feature
I think all 5000 of them are busy "moderating" complaints. My account was suspended for 7 days for telling a youtuber (one I'd spoken with back and forth many times) that she had some clothes fluff in her armpit on pic she posted. Apparently that falls under harassment of a sexual nature and my appeal was denied within 15 minutes.
I don't know Kayvon personally, but what exactly is wrong with Twitter that it's so bad and slow at product development? This should be #5000 on their list of things to do.