Yes you are right. It is brought up a lot. But the crazy thing is that despite it coming up in every other thread, I have yet to see a solid explanation. Some theories (it looks good for VCs!!!!!), some half-assed justification (akshually, running web apps does need thousands of engineers), and a lot of people agreeing that it's crazy.
I think it's probably true that most companies are overstaffed. He-Who-Shall-Not-Be-Named managed to cut headcount 80% at a fairly prominent tech employer and drama aside, the business has not severely suffered. Could Ford Motor Company or US steel or Delta Airlines lay off 80% of their workforce and manage to maintain and grow operations as well as the bird app has managed to? I doubt it.
I mean... according to him, that company's business HAS suffered. Granted, he blames it on a conspiracy out to "cancel" him with advertisers rather than technical problems, but still.
The advertiser exodus has been caused by Elon's contentious personality and his strategic decisions on how to structure Twitter as an experience, with regard to re-platforming controversial users, broadening the definition of acceptable content, and so on and so forth.
It has not been caused by laying off 6,000 project managers and react developers, and will not be fixed by bringing back on 6,000 people at six figure salaries.
The app itself is functioning well even as the number of users has increased and number of staff decreased, and not only is it being maintained, but they seem to be regularly releasing new features without protracted delays or major technical bungles.
Many advertisers have backed off because laying off a lot of people has led to a worse product.
Many more advertisers who might have otherwise been persuaded to stay by sales reps they had contact with have backed off because those sales reps themselves were let go.
One could genuinely make an argument that most SV tech companies have way more employees than they need. And one could almost trivially prove that to be true (many Tech companies were hiring developers not because they needed them, but to keep them out of the hands of competitors). That does not make Elon’s slash and burn destruction of Twitter sensible by any measure.
Meta is also offloading vast number of employees and it’s managing to do it without making it sound like their company is a dumpster fire.
Elon’s approach to Twitter is highly reminiscent of the Stephen Elop burning platform memo to Nokia, but orders of magnitudes worse.
Hyperbole. Advertisers backed away from Twitter because Musk was upsetting a group of people whom advertisers tend to tiptoe around. Twitter is terrible, but it's no more terrible now than it was when Musk bought it.
The irresistible allure of Elon's drama is fascinating. I kept my comment as vague as I could manage yet somehow people cannot help themselves but show the world their opinion about him.
Aside from Trump, I don't think any other person can so successfully and reliably cast this attention grabbing spell on the world's netizens.
What Elon has been doing, and how much of an asshole we may think he is, is actually completely unrelated to the fact that Twitter was overstaffed for essentially the entire duration of it's life as a public company. Which is the sole point of my initial comment.
To be fair, you opened with something that by your own words is not relevant, which is what this person was responding to. It's understandable that people respond to what you say and write, if you do not want to talk about it you should not say it.
> He-Who-Shall-Not-Be-Named managed to cut headcount 80% at a fairly prominent tech employer and drama aside, the business has not severely suffered
We need to stop repeating this lie.
They have reduced site reliability to reddit levels, had multiple breaches (probably far more than the ones we know about) and managed to get in trouble in multiple countries.
And yes, they were overstaffed like any other company hiring during the pandemic but not by 80%.
this! it's a radically different website, and has been through some really intense changes over the years, from the people, the methods of influencing the platform, and the platform's own position on the content and it's moderation.
Depends on time scale. You can only piss off a certain percentage of users for a short time. If that percentage goes up, and people stay angry, that takes longer, but then you die.
People forget exponential growth in a social network always ends with exponential collapse. It takes a while to get there, but sites can clear out in a matter of months.
Our society and more specifically the finance class of investors, have structured society to prioritize getting a “positive exit” as more important than literally anything else in society.
when you raise capital you've got to give it a return, and an IPO is the classic exit if the amount of capital you've raised is huge. Early investors get their payday.
If you have invested in a company, selling it on the stockmarket is a good way to turn the stake you previously bought into money.
Obviously you want to maximize the selling price, as that is the actual profit you are making. If they are IPOing, they either need funding or believe its current evaluation will give them a good price.
I sometimes wonder is layoff just for show for the investors. A company the size of reddit couldn't possibly care about saving salary of 90 people to take bad PR. Instead they could have simply moved people around and fire 90 over period of a year or so easily as part of performance review. Similarly I know a company whose yearly attrition rate is 30%, and they fired 10% of the people with 4 months reverence.
Is it really bad PR? People here do discuss it but it doesn't change anything regarding their users, they are just there to mindlessly consume. And for investors as we have seen it can come off as a good signal.
Only people working in tech mind care but do what in the big picture?
I've heard anecdotally from people I trust that social media is dying.
Content is struggling for views and shares.
Reddit isn't completely social media, but if people aren't responding to SM like they were it will hurt Reddit. If you're an insider it's probably the best opportunity you're going to get to juke the stats and max your IPO. Only going to go down from here.
I don't generally have a good sense of where gen pop is heading, but recently the few social media type sites I still use are losing what little appeal they had. It's all just so boring and recycled.
I don’t think social media broadly is going away, but they are changing.
Facebook kinda killed its own activity by pushing growth of their social graph.
Over time it stopped being a place to connect with friends and just became… I don’t know. But it discourages user activity because family, friends, old classmates, guys you bumped into on a night out, ex’es all can see your activity.
It’s not fun checking-in somewhere, posting funny pictures from last nights party or whatever because it is practically public.
And my impression is young kids don’t like it either. You do your highlights reel on Instagram, but other than that, everything happens in private chat groups.
Every meetup group I go to, every Facebook group I visit wants us to connect in a private chat group and have notifications every time someone says something.
For this old fart, it’s exhausting to keep up with.
I feel like there is only so much "interesting" stuff going on in the lives of the people I care about.
The need for SM companies to get more of my time requires them to shower me with increasingly irrelevant content to keep my engagement going up. At some point there is literally no additional time people can spend in SM (you can't get a time credit card).
So more ads it is until the plattform is barely usable for someone not interested in the passively consumed avalanche of polished medicare.
That's why chat groups are increasingly popular I think. You get to have a two way conservation with people who respond and care about you (at least a little bit). Meeting them in real live (or one on one online) also enforces the "authenticity" of the interaction.
Therefore I expect Facebook to introduce lots of ads to WA very very soon.
It's all been a lot of smoke and mirrors for ages now. There's nothing really that can prop up an empty ecosystem of encouraging people to work furiously for no reward or payout while only scammers and the rich win...
The platforms have been allowing their own employees and leveraging bots and contracted employees as fake "winning" accounts/influencers while deeply manipulating genuine (unaffiliated) users to pay for visibility through making them buy ads to profit and/or look successful... Even sites that aren't really considered as social media, like YouTube and reddit, are guilty of these practices.
I really struggle to understand how ads are still as big as they are. With the number of bots, adblockers, and straight up lies about the age of users, how are people not realizing that ads are not what they were when google/facebook were brand new.
Adds are bigger then ever and the more restricted the devices of people are the better for the adds. Addblockers do not work on the YouTube app.
Also adds have gotten far better, they are now paying to be integrated directly into content, making it far more effective by using the trust the viewer has built up towards the creator.
I think it's going to get worse for social media. Lately I've been seeing more and more "content" suggested to me that is AI generated. Architecture that doesn't exist. People that don't exist in locations that don't exist wearing clothes that don't exist. I'm able to notice the slight unreality of it, its too-perfect look, and I check the tags and sure enough it's #midjourneyai or something. And I have this out of body experience. What am I doing with my time? It's vapid, empty content. I can never go to this place or see this thing or person, but it's presented as real as possible. It just kind of makes the whole thing seem absurd to me.
At least in the past even if it was an influencer, I knew it was a real person or location, if a little Photoshopped. Now it just might as well be me daydreaming?
It's hard to explain. It's kind of the end of meaning, I guess.
There is going to be an AI/AR/VR existential crisis in some people.
The status quo is that we interact in 2 distinct ways. First, with a simulacra of a real object or people, but at a distance. That's video conferencing, texting, email, (mostly) social media, etc. Second, with simulations, i.e. video games, produced media (i.e. fake people played by real ones, photoshopped everything, post-produced CGI'd, etc.).
The difference now is that the wall between these is going to fall, and it will be difficult to distinguish between the two going forward.
Culture doesn't require that its basis is real. Some people are going to choose Religion 2.0, where they prefer the mediated experience to the unmediated one, and others will reject it. These two groups can't get along because even if values are shared, there will be a divergence of meaning.
> At least in the past even if it was an influencer, I knew it was a real person or location, if a little Photoshopped.
Maybe not you specifically, but society generally: why does one follow an influencer? Surely they won’t meet that person, likely won’t go to that location, can’t afford those clothes etc….
And in answering those questions, does AI generated content still fill that requirement? I suspect yes for some.
I do follow the occasional influencer and brand. Some because I want to peak into a glorious resort filled life I won’t have, others to see remarkable places I won’t go. If those were fake, maybe I’d feel better about FOMO.
Some people (especially queer people) follow influencers to experience people and social groups who are like them, when they don’t see them in their normal life. AI content can’t replace that.
>why does one follow an influencer?
From the outside it seems mostly aspirational and a way for people to fantasize about a life they will never have. (This is extremely toxic behaviour, by the way)
I also think this directly links to the "life improvement" and "course selling" genre, where millionaires blatantly shill to their audience that they too can be millionaires if they just get up before 5:30 and gamble on ridicolous activities in saturated markets (course selling, drop shipping, "book writing"). Honestly some of the more despicable people on the internet.
>Some people (especially queer people) follow influencers to experience people and social groups who are like them, when they don’t see them in their normal life. AI content can’t replace that.
Why not? I think it is far easier to make AI pander to those people. AI girlfriend/boyfriend simulators are already real, they directly target those groups.
> AI girlfriend/boyfriend simulators are already real, they directly target those groups.
That is not at all the point or use case i was referring to (but it is totally a real use of social media). Using queer people as the example, a lot of young people turn to social media to see adults being happy and queer because they don’t see any examples IRL. There are a lot of laws being passed in the US to target trans people, for example, so a Floridian trans youth may watch a YouTube channel of a trans person in a far away city living a safe, fulfilling life, as inspiration for their own future.
Similarly, but more pedestrian, I knew plenty of people in high school who watched “day in the life of <profession>” videos, or “come with me on my first day at <university>” videos. They help people visualize their future, or decide between different futures. I used YouTubers’ videos to help me choose between different universities. When I moved to a new city, I followed local influencers to get an idea for things to do and places to visit. Definitely less toxic.
These videos can be fake, but if it’s not grounded in reality then it’s really not the intended purpose of watching it.
>but if it’s not grounded in reality then it’s really not the intended purpose of watching it.
Who cares? If you can't tell whether it is real or not, where is the difference.
AI can generate homosexual couples just as well as others. I don't get what your point here is.
>They help people visualize their future
No, they help people fantasize, which is an extremely destructive behaviour.
>I used YouTubers’ videos to help me choose between different universities. When I moved to a new city, I followed local influencers to get an idea for things to do and places to visit. Definitely less toxic.
Completely unimaginable to me. Also an extremely easy way to get manipulated.
> No, they help people fantasize, which is an extremely destructive behaviour.
Fantasize about being a doctor or lawyer or engineer or really anything is destructive? Everything I have done in my life was once a fantasy. I mean is watching a resident talk about their life in a YouTube video worse than me watching ER in the 90s?
I know what you mean. There is something comforting in knowing that a real person like you is out there in world, existing and somehow making it. It's important that they are actually real too, or as real as you can verify.
I was also thinking about how, you know, fake content that I know is fake is fine. A comedy sketch is clever writing and good timing, and synthetic.
Real content that I know is real is fine. I am seeing something real and true.
Fake content that tries to be real is bad, and pollutes my mind, in a sense. I can't even tell somebody about an image I saw because I don't know if it's real or not. Balenciaga pope etc. And a lot of social media is going to fall into that. I shouldn't even mention it, for fear of looking like a fool. I shouldn't even mentally store it, because it warps my understanding of general trends that I notice online. I think social media text is already there, and images are close, video maybe less than a year off from being kind of worthless.
It's partially because of dilution. There are so many things out there seeking our attention. There aren't enough people on this planet to support it all.
Honest question from someone who has never worked at a Bay Area startup:
What do all these developers at these tech companies do all day? As a freelance developer who has to meet ridiculous timelines all the time, I don't really get how a company can have hundreds of developers and yet the product seems to languish and/or get worse or slower or both.
I suppose there's a lot more overhead with internal QA, code review, meetings, etc, but with the amount of developers these companies have as full-time staff, what are they doing all day? Is it mostly internal systems, tooling, etc?
I just find it hard to believe that there can be hundreds of developers at a company like Twitch and yet the product is largely the same as it was 5 years ago. I would think features could be cranked out so much faster than they appear to be.
This sort of question comes up a lot when people are talking about big social media sites and there are two answers, both related to scale. First is that things really do just go more slowly when you are larger. There are more stakeholders and greater consequences to change. You don't change anything without measuring it first through flagging and limited a/b releases, and often those a/b tests actually fail to show a reason to release the change since the stats show that even though you think it's an improvement, it actually has a negative impact on key metrics.
But the second one is that while the site might seem simple and straightforward, the way it has to be engineered to handle the massive scale of a Reddit (with over a billion active monthly users) makes it very, very complicated internally. I'm not familiar with Reddit specifically but having seen the architecture at other places of similar size, I can tell you there's a ton going on behind the scenes to make it possible to reliably serve that many users and that makes major changes very complicated, involving multiple teams, and lots of planning.
These companies don’t operate like bay area startups, nearly the opposite. Where a startup is low on process and very nimble, these orgs are slow and bloated. They have too much money. They’re more like a bad corporate gig with even less focus on quality because they heard Zuck say breaking things is ok.
Peel back the layers of any site that has millions of active users and you'll find all sorts of things you never imagined would be a problem, but become so at scale because, well, people (and, well, scale).
59 comments
[ 3.9 ms ] story [ 134 ms ] threadAstonishing really considering Reddit hasn't changed that much since back when they had like 100 employees.
If I had a dollar for every “X is trivial” comment on HN…
I think it's probably true that most companies are overstaffed. He-Who-Shall-Not-Be-Named managed to cut headcount 80% at a fairly prominent tech employer and drama aside, the business has not severely suffered. Could Ford Motor Company or US steel or Delta Airlines lay off 80% of their workforce and manage to maintain and grow operations as well as the bird app has managed to? I doubt it.
But that aside, your point is very good.
It has not been caused by laying off 6,000 project managers and react developers, and will not be fixed by bringing back on 6,000 people at six figure salaries.
The app itself is functioning well even as the number of users has increased and number of staff decreased, and not only is it being maintained, but they seem to be regularly releasing new features without protracted delays or major technical bungles.
Many more advertisers who might have otherwise been persuaded to stay by sales reps they had contact with have backed off because those sales reps themselves were let go.
One could genuinely make an argument that most SV tech companies have way more employees than they need. And one could almost trivially prove that to be true (many Tech companies were hiring developers not because they needed them, but to keep them out of the hands of competitors). That does not make Elon’s slash and burn destruction of Twitter sensible by any measure.
Meta is also offloading vast number of employees and it’s managing to do it without making it sound like their company is a dumpster fire.
Elon’s approach to Twitter is highly reminiscent of the Stephen Elop burning platform memo to Nokia, but orders of magnitudes worse.
Let's call it what it is: he is using his power to promote some people and some ideologies while actively censoring others.
Aside from Trump, I don't think any other person can so successfully and reliably cast this attention grabbing spell on the world's netizens.
What Elon has been doing, and how much of an asshole we may think he is, is actually completely unrelated to the fact that Twitter was overstaffed for essentially the entire duration of it's life as a public company. Which is the sole point of my initial comment.
Source?
The most prominent example was the failed DeSantis presidential announcement.
I mean, Twitch regularly does larger livestreams with video!
We need to stop repeating this lie.
They have reduced site reliability to reddit levels, had multiple breaches (probably far more than the ones we know about) and managed to get in trouble in multiple countries.
And yes, they were overstaffed like any other company hiring during the pandemic but not by 80%.
And I blame the breaches on churn more than head count. And the 'getting in trouble' is generally because of the new things being added.
It's changed dramatically, but many of the visible changes are for the worse!
To error is human but to really screw things up requires significant manpower.
Pump the revenue however possible (crazy API fees, push for native only) assuming that people have no choice cause it’s been so long
Slash anything that isn’t directly revenue producing
All to make the P/L look really groovy but more importantly as a loyalty test to your underwriters.
>> Pump the revenue however possible
People forget exponential growth in a social network always ends with exponential collapse. It takes a while to get there, but sites can clear out in a matter of months.
So this is them getting that for Reddit
Obviously you want to maximize the selling price, as that is the actual profit you are making. If they are IPOing, they either need funding or believe its current evaluation will give them a good price.
Only people working in tech mind care but do what in the big picture?
Reddit isn't completely social media, but if people aren't responding to SM like they were it will hurt Reddit. If you're an insider it's probably the best opportunity you're going to get to juke the stats and max your IPO. Only going to go down from here.
Facebook kinda killed its own activity by pushing growth of their social graph.
Over time it stopped being a place to connect with friends and just became… I don’t know. But it discourages user activity because family, friends, old classmates, guys you bumped into on a night out, ex’es all can see your activity.
It’s not fun checking-in somewhere, posting funny pictures from last nights party or whatever because it is practically public.
And my impression is young kids don’t like it either. You do your highlights reel on Instagram, but other than that, everything happens in private chat groups.
Every meetup group I go to, every Facebook group I visit wants us to connect in a private chat group and have notifications every time someone says something.
For this old fart, it’s exhausting to keep up with.
The need for SM companies to get more of my time requires them to shower me with increasingly irrelevant content to keep my engagement going up. At some point there is literally no additional time people can spend in SM (you can't get a time credit card).
So more ads it is until the plattform is barely usable for someone not interested in the passively consumed avalanche of polished medicare.
That's why chat groups are increasingly popular I think. You get to have a two way conservation with people who respond and care about you (at least a little bit). Meeting them in real live (or one on one online) also enforces the "authenticity" of the interaction.
Therefore I expect Facebook to introduce lots of ads to WA very very soon.
The platforms have been allowing their own employees and leveraging bots and contracted employees as fake "winning" accounts/influencers while deeply manipulating genuine (unaffiliated) users to pay for visibility through making them buy ads to profit and/or look successful... Even sites that aren't really considered as social media, like YouTube and reddit, are guilty of these practices.
Also adds have gotten far better, they are now paying to be integrated directly into content, making it far more effective by using the trust the viewer has built up towards the creator.
At least in the past even if it was an influencer, I knew it was a real person or location, if a little Photoshopped. Now it just might as well be me daydreaming?
It's hard to explain. It's kind of the end of meaning, I guess.
The status quo is that we interact in 2 distinct ways. First, with a simulacra of a real object or people, but at a distance. That's video conferencing, texting, email, (mostly) social media, etc. Second, with simulations, i.e. video games, produced media (i.e. fake people played by real ones, photoshopped everything, post-produced CGI'd, etc.).
The difference now is that the wall between these is going to fall, and it will be difficult to distinguish between the two going forward.
Culture doesn't require that its basis is real. Some people are going to choose Religion 2.0, where they prefer the mediated experience to the unmediated one, and others will reject it. These two groups can't get along because even if values are shared, there will be a divergence of meaning.
Maybe not you specifically, but society generally: why does one follow an influencer? Surely they won’t meet that person, likely won’t go to that location, can’t afford those clothes etc….
And in answering those questions, does AI generated content still fill that requirement? I suspect yes for some.
I do follow the occasional influencer and brand. Some because I want to peak into a glorious resort filled life I won’t have, others to see remarkable places I won’t go. If those were fake, maybe I’d feel better about FOMO.
Some people (especially queer people) follow influencers to experience people and social groups who are like them, when they don’t see them in their normal life. AI content can’t replace that.
I also think this directly links to the "life improvement" and "course selling" genre, where millionaires blatantly shill to their audience that they too can be millionaires if they just get up before 5:30 and gamble on ridicolous activities in saturated markets (course selling, drop shipping, "book writing"). Honestly some of the more despicable people on the internet.
>Some people (especially queer people) follow influencers to experience people and social groups who are like them, when they don’t see them in their normal life. AI content can’t replace that. Why not? I think it is far easier to make AI pander to those people. AI girlfriend/boyfriend simulators are already real, they directly target those groups.
That is not at all the point or use case i was referring to (but it is totally a real use of social media). Using queer people as the example, a lot of young people turn to social media to see adults being happy and queer because they don’t see any examples IRL. There are a lot of laws being passed in the US to target trans people, for example, so a Floridian trans youth may watch a YouTube channel of a trans person in a far away city living a safe, fulfilling life, as inspiration for their own future.
Similarly, but more pedestrian, I knew plenty of people in high school who watched “day in the life of <profession>” videos, or “come with me on my first day at <university>” videos. They help people visualize their future, or decide between different futures. I used YouTubers’ videos to help me choose between different universities. When I moved to a new city, I followed local influencers to get an idea for things to do and places to visit. Definitely less toxic.
These videos can be fake, but if it’s not grounded in reality then it’s really not the intended purpose of watching it.
AI can generate homosexual couples just as well as others. I don't get what your point here is.
>They help people visualize their future No, they help people fantasize, which is an extremely destructive behaviour.
>I used YouTubers’ videos to help me choose between different universities. When I moved to a new city, I followed local influencers to get an idea for things to do and places to visit. Definitely less toxic. Completely unimaginable to me. Also an extremely easy way to get manipulated.
Fantasize about being a doctor or lawyer or engineer or really anything is destructive? Everything I have done in my life was once a fantasy. I mean is watching a resident talk about their life in a YouTube video worse than me watching ER in the 90s?
I was also thinking about how, you know, fake content that I know is fake is fine. A comedy sketch is clever writing and good timing, and synthetic.
Real content that I know is real is fine. I am seeing something real and true.
Fake content that tries to be real is bad, and pollutes my mind, in a sense. I can't even tell somebody about an image I saw because I don't know if it's real or not. Balenciaga pope etc. And a lot of social media is going to fall into that. I shouldn't even mention it, for fear of looking like a fool. I shouldn't even mentally store it, because it warps my understanding of general trends that I notice online. I think social media text is already there, and images are close, video maybe less than a year off from being kind of worthless.
What do all these developers at these tech companies do all day? As a freelance developer who has to meet ridiculous timelines all the time, I don't really get how a company can have hundreds of developers and yet the product seems to languish and/or get worse or slower or both.
I suppose there's a lot more overhead with internal QA, code review, meetings, etc, but with the amount of developers these companies have as full-time staff, what are they doing all day? Is it mostly internal systems, tooling, etc?
I just find it hard to believe that there can be hundreds of developers at a company like Twitch and yet the product is largely the same as it was 5 years ago. I would think features could be cranked out so much faster than they appear to be.
But the second one is that while the site might seem simple and straightforward, the way it has to be engineered to handle the massive scale of a Reddit (with over a billion active monthly users) makes it very, very complicated internally. I'm not familiar with Reddit specifically but having seen the architecture at other places of similar size, I can tell you there's a ton going on behind the scenes to make it possible to reliably serve that many users and that makes major changes very complicated, involving multiple teams, and lots of planning.