Meta CTO Tells Employees Higher Headcount Has Led To 'Untenable' Slow Movement 65
An anonymous reader shares a report: Meta CTO Andrew Bosworth has one of the toughest jobs in tech this year. On one hand, he has to deliver on CEO Mark Zuckerberg's grand metaverse ambitions as Apple and ByteDance are entering the space. At the same time, he's also attempting a dramatic cultural reset within Reality Labs, the sprawling division responsible for those ambitions. In an internal memo I obtained that he sent to employees just before the holidays, Bosworth acknowledged a sentiment I've been hearing from current and ex-employees for a while: "We have solved too many problems by adding headcount. But adding headcount also adds overhead. And overhead makes everything slower."
"Every week I see documents with 100+ editors," he wrote to the roughly 18,000 people in Reality Labs. "A meeting with 50+ people that took a month to schedule. Sometimes there is even a 'pre-meeting' with its own document. I believe the current situation is untenable."
"Every week I see documents with 100+ editors," he wrote to the roughly 18,000 people in Reality Labs. "A meeting with 50+ people that took a month to schedule. Sometimes there is even a 'pre-meeting' with its own document. I believe the current situation is untenable."
Re: (Score:3)
Perhaps the problem is in not organizing them. If they needed a meeting with 50 people, then something's gone wrong. Organize into smaller groups, those groups are put into divisions, and then the CTO in charge only has to meet with the division heads, etc. Otherwise too much time is wasted on micromanagement.
If you can't arrange a meeting with 50 people so that all are there, then just meet with 40 of them. It's almost like this CTO has never worked at a corporation before.
Re: (Score:2)
18,000 people?? Doing what?? There shouldn't be 18,000 people in all of Meta/Facebook, let alone one division.
Actually, 18K is not a bad figure for a multinational. For a division, though? Yes, I would agree with you.
Re: And they keep ignoring the REAL problem (Score:2)
Firing is the solution? (Score:2)
Ok, so a meeting with over 50 people takes a month to schedule, but how is firing people the solution? Sounds like it can be fixed with some role adjustment. This is like dictators thinking they have too many people to feed so the solution is eliminating them. How about figuring out things for them to do? If a company can't figure out how to use its resources you can't blame it on the mere existence of people.
Re: (Score:1)
Ok, so a meeting with over 50 people takes a month to schedule, but how is firing people the solution? Sounds like it can be fixed with some role adjustment. This is like dictators thinking they have too many people to feed so the solution is eliminating them. How about figuring out things for them to do? If a company can't figure out how to use its resources you can't blame it on the mere existence of people.
Where does it say he's going to fire people? He's pointing out that the situation is untenable with the overhead. He might be considering Dilbert's classic Battlin' Business Units or a restructure towards hard interfaces with an engineering hierarchy.
Yes firing is the solution (Score:1, Troll)
Ok, so a meeting with over 50 people takes a month to schedule, but how is firing people the solution? Sounds like it can be fixed with some role adjustment.
Adjust to do what?
If there are too many people, there are too many people. Even if you adjust those roles, you have to have something for them to do, people to oversee that many people, on and on.
Have you ever read "The Mythical Man Month"? Sometimes a goal just has a certain number of people you need to accomplish it, and more people just drags you d
Re:Yes firing is the solution (Score:5, Insightful)
Have you ever read "The Mythical Man Month"?
Admittedly, it's been a while. I forgot the chapter about firing people until the team delivers fast enough.
Re: (Score:2)
Has it been so long you forgot the basic premise of the entire book?
You lack reading comprehension.
Re: (Score:1)
You lack reading comprehension.
You should probably re-read the book to understand the main points [8thlight.com].
The summary is good but really the book is very good, and anyone who has not read it should read the whole thing.
Re: (Score:3)
Or, and I know this depends on the project, break the project of 50 into 5 projects of 9, and a 5 person team to handle the connecting interfaces.
Re: (Score:2)
Yeah, it's tough to write stuff like that in Jupyter notebooks.
ECMAScript finally has support for modules though, so maybe there's hope.
Re:If you believe one way, why not reverse (Score:4, Insightful)
"If you can bring on enough people you start making a project take longer (which is what the book covers), can you deny it makes sense that letting some people go could make a project go faster?"
Of course I can. The point of the book was that bringing new people on board generates more work because those people start completely ignorant of the project. Therefore, letting people go doesn't help, because you are letting people go who know the project.
"Has it been so long you forgot the basic premise of the entire book?"
The person who has not grasped the basic premise of the book would appear to be you.
Re: (Score:2)
The point of the book was that bringing new people on board generates more work because those people start completely ignorant of the project.
Erm, no? If people being ignorant of a project is to the detriment of every project, then no project would ever get started, because when it starts, everybody is "ignorant" of it.
On the contrary; the point of the book was that adding people to a team does not make the team work faster, because a larger team increases the overhead of communication between members of the team. (It's actually more complicated than that, but your statement simplifies it to an absurd degree.)
Re: (Score:2)
Of course I can. The point of the book was that bringing new people on board generates more work because those people start completely ignorant of the project. Therefore, letting people go doesn't help, because you are letting people go who know the project.
Partially related: I got a kick out of Elon Musk firing practically the entire team behind Twitter on day one (not just managers but technical people as well). Who does he think is going to train all the "new" people he brings into the company to fix everything? Platforms that large and complex can't just be tweaked overnight by a new field of yes-men, and all programmers know it's a nightmare to work on someone else's code.
Sigh... managers.
Plainly False (Score:1, Insightful)
I got a kick out of Elon Musk firing practically the entire team behind Twitter on day one (not just managers but technical people as well). Who does he think is going to train all the "new" people he brings into the company to fix everything?
Since Twitter has not really broken, and indeed has been handling record traffic, AND they have also been rolling out new features like Pay-to-blue, what you have said is obviously completely wrong. Otherwise how would you explain any of what Twitter has done?
Twitter
Re:Yes firing is the solution (Score:5, Insightful)
I assume that they have some specific projects(like kicking out the next revision of VR headset) that are well defined and require a nontrivial number of people; but the 'reality labs' division's problem in general seems to be its inability to answer "why would I possibly want to be trapped with cartoon Mark Zuckerberg in what looks like a garbage Unity asset flip?". It's not clear that there is an answer to that question, and I certainly hope that they go bankrupt trying to find it; but in that position telling small groups to go do interesting things seems like the in-house equivalent of buying small competitors who do interesting things; which is their current principle source of novel products.
This isn't to say that they shouldn't just get rid of some people; but the 'Mythical Man Month' case was one where people actually knew what they wanted and were trying to execute on it(which simply cannot be parallelize beyond a certain point); while Facebook seems to have the problem of not actually being at the point of having a desired end product that is moving slowly; and the process of fishing through the search space looking for something interesting to do, while not necessarily hopeful, is a lot easier to parallelize by just telling people to go off and do their thing and see what happens.
Re: (Score:2)
The growth uber alles mentality fueling the egos of corporate khans and their capitalist hordes rarely bodes well outside their limited circle. The results typically leading to over expansion, declining product quality and less competitive markets. Allowing far flung conquered corps to keep their cultures rather than undergoing forced assimilation simply postpones the inevitable as the tributes they pay aren't worth the
That is a LOT of overhead with no result (Score:1)
the obvious off-the-cuff suggestion would be to have a lot less of them talking to one another in giant groups and a lot more of them off skunkworksing anything that seems even modestly interesting.
Is the burn rate of millions (or tens or hundreds) of millions of dollars really worth the VERY slim chance those people who have produced nothing of worth in the past five years, will suddenly strike gold?
What you describe is the process that brought us the Amazon Fire Phone...
Re: (Score:2)
Bullshit Jobs might be more apropos.
Re: (Score:1)
Dictators are often right, given their position. You can't control a sprawling bureaucracy with thousands of people. The ACTUAL work that needs to be done is often minimal. Facebook has a few dozen tech products, managing a team of >100 programmers for 1 product is impossible, even Microsoft knows this, the kernel team is ridiculously small in comparison with the overall Windows product.
So you have perhaps 1000-2000 people that do work to get a website up and running and scaling it, perhaps a few hundred
Obligatory attempted joke (Score:2)
Nope, the problem with head count at Meta was only one. The one at the top of the dead fish.
(Hey, don't look at me. My 12-year-old Facebook identity was murdered. Without warning or reason. No idea why which includes no reason it couldn't happen to you.)
Re:Twitter leads the way (Score:4, Informative)
Twitter isn't paying its bills [axios.com], including for office space.
Twitter is refusing to pay bills run up before Musk took over, which means its legal bills are going through the roof [abovethelaw.com].
Twitter has reneged on its promise to pay appropriate severance [forbes.com] for people it let go. As a result, it is facing a multitude of lawsuits and arbitration hearings. When it finally did offer something, it was far below what was promised [cnn.com] and included clauses the people could not say a word about their time at Twitter, except if Twitter called them in for a lawsuit to defend them.
Try again.
Re:You see problems, I see cost reduction (Score:4, Insightful)
Those are all reasons why Twitter will be profitable later this year, reduction in needless expense.
Apparently office space is a needless expense. The same office space Musk ordered employees to return to [techgig.com]. You'll note in the article, as a direct of firing thousands of people, those remaining are working 12-hour days for weeks on end.
Yes legal expenses are up but that is small compared to extravagant expenses made by previous management they do not want to pay.
In other words, the company incurred the expenses but now isn't going to pay them. Sounds like a great plan.
As for your severance link, it says nothing like what you claim. Just that they are slow sending out agreements.
That was the first link. Read the second link. Only one month, no accelerated stock options, and pretty much nothing what Musk said he would do. Which is typical for him. He likes to hear his own voice.
Re: (Score:1)
Promised how exactly? Just because a CEO says they have an idea, doesn't mean that you get a new contract that says that. Also, Twitter is now private and doesn't have stock options, anyone who had them got paid out.
Re: (Score:3)
At what point will all the evidence of his massive failure be enough to overcome you blind devotion to a guy who will never care that you exist?
His incompetence as already cost him far more than he even paid for Twitter. The guy is stumbling around drunk and you think he's playing 4d chess.
Re: (Score:2)
The guy is stumbling around drunk and you think he's playing 4d chess.
He might be better off if he were stumbling around drunk. At least then nobody would let him sign contracts.
Re: (Score:1)
Twitter is refusing to pay bills run up before Musk took over
Those are all reasons why Twitter will be profitable later this year, reduction in needless expense. Yes legal expenses are up but that is small compared to extravagant expenses made by previous management they do not want to pay.
As for your severance link, it says nothing like what you claim. Just that they are slow sending out agreements.
Under California law, full payment of all monies owed is due at the time of termination. After a 7 day grace period, a penalty of 1 day additional pay is due for each day of delay.
Expenses are going UP.
Re: (Score:2)
who's gonna blink first. the company who will eventually get this money but has their own bills and services to pay for, or musk who has billions in the bank.
it looks like an open an shut case, but with money to pay lawyers to drag things out, it may very well be faster and cheaper to settle out of court for less than the total amount owed.
methinks you don't really business much, you just like playing one on tv from the comfort of your arm chair
Re: (Score:3)
I think you miss the obvious sarcasm in the parent post.
Re: (Score:2)
It's a bit early to say what will happen at Twitter. It's easy to squeeze a few pennies of profit by cutting costs to only things that are immediately generating profit. But, cut those costs too much and the whole ship may sink.
However, Twitter itself as a web application I imagine 100 solid developers(this might be an over estimate) could rewrite in a year with skilled project management. The business itself is selling ad space and promotions and that's where the bulk of the spending on headcount should
Inundated how? (Score:2)
And finally, curation, so that people don't get traumatized or inundated with the darkest stuff on the internet or scam ads,
Curation of content on Twitter beyond obeying legal requests is pointless, because everyone on Twitter follow people they like. You can only be "harmed" if you actively seek harm. Trying to curate Too Deep is what lead old Twitter into the abyss.
Old excuse (Score:3, Insightful)
Reality is they ramped up for that "Metaverse" nonsense and now they're ramping down, but nobody wants to tell somebody as rich & powerful (and meanspirited) as Mark Zuckerberg that his ideas are stupid and everyone hates them and hates him too.
Re: (Score:2)
Yep, that nicely sums it up.
Let me get this straight (Score:2)
Meta has a headcount problem, and they haven't reassigned these people to work on Zuck's dream?
Re: (Score:2)
Reality Labs is the part that works on VR.
The more the merrier (Score:2)
Re: (Score:2)
2 men build a house in 1 year. 4 men in 0.5 years 63072000 men build a house in one second.
And 2 men with 63072000 heads have a headcount problem?
Re: (Score:2)
Do 9 women have a baby in a month?
Chopping off the wrong heads (Score:5, Insightful)
Over-hiring, hiring the wrong people, and failure to manage those people are NOT a reflection of the people you hired, it is a reflection of your management and executive team.
Cleaning house at the management/executive suite is where you start addressing that problem.
Re: (Score:3)
Exactly. At least if you are competent. Of course, if your C-levels are idiots, especially the CEO (as is the case here), this has to come from the board and the board of Meta is probably all yes-men or they would have done something already.
It's culture, not headcount (Score:5, Insightful)
Adding people can do amazing things for projects. Adding people can do horrible things for projects.
The problem isn't the people, it's management and the values and culture they explicitly and implicitly endorse.
If they don't take time to understand what their team is working on, and what's slowing them down from achieving their goals, the number of heads(more or less) won't change the output. Tracking performance from Agile story points or tickets closed will lead to an inflation of story points and tickets; Oftentimes, said tickets will be closed without a deliverable that maps to the ticket. When management and project management becomes obsessed with impressing their boss and loses touch with the actual work, they allow dysfunctions to propagate. Oftentimes, this is due to conflicting priorities between teams and even within teams. It's not lazy developers. It's not indicative of needed more accountability. Developers almost always LOVE delivering good products on schedule and find the managerial swamp demoralizing. And, thus, this is an issue with management itself not performing it's fundamental responsibility, which is to reshape priorities of the organization and teams to work towards the companies' global goals and reduce friction within the company.
The problem is, management HATES feedback flowing upwards and actively discourage it. The personal incentive to quash feedback, either through explicit policy or defacto is too great for middle managers. It leads to 'waterfall' type decision making, where the CEO says "I want more money," the second tier management says "ok then lets squeeze productivity from our middle management," middle management says: " Hello Peter, what's happening? Ummm, I'm gonna need you to go ahead come in tomorrow. So if you could be here around 9 that would be great, mmmk... oh oh! and I almost forgot ahh, I'm also gonna need you to go ahead and come in on Sunday too, kay. We ahh lost some people this week and ah, we sorta need to play catch up."
Re: (Score:2)
Re: (Score:2)
The real issue... (Score:5, Informative)
If you are a C-level executive complaining about your team size or productivity or quality, it should be in your resignation letter because YOU created that problem.
Actually, it isn't (Score:2)
This is a common misunderstanding by people that are at the bottom of the food chain.
In most large organizations the C-level people set the direction and goals, not how to implement those goals. That's VP/Director level stuff and below.
Too many chiefs (Score:2)
Not enough Indians.
Re: (Score:2, Funny)
that's racist and non-inclusive. They're called H1Bs now :P
you, not me (Score:2)
It's not my fault; it's your fault.
Re: (Score:2)
It's not my fault; it's your fault.
Facebook employees skew younger. Millennials have had it all worked out since middle school, where they were carefully instructed in the necessary strategy: it's not my fault, it's not your fault, therefore it's nobody's fault. There were 50 people in the meeting. It couldn't possibly be any one person's fault.
Schools have been teaching kids to be herd animals since the '90s. Safety in numbers, responsibility diffused so far there can be no accountability, and your only real priority is to make sure you
Dilbert IRL (Score:3)
This is Dilbert in real life. Fun to watch from the outside.
Meetings can be productive. A successful meeting is where people show up to the meeting and everyone walks away with something to do and then everyone does those things ASAP. Someone, ideally the lead manager in the meeting, needs to make sure the meeting closes with next steps and task assignments. The least successful meetings I've ever been a part of are those where one person is given an assignment to ask someone else not in attendance a question and then they forget to do that before the next meeting. I'm a stickler for making sure that a meeting is productive so that my time and everyone else's time isn't wasted.
I work at Meta (Score:4, Interesting)
Re: (Score:2)
It used to be, they changed it in 2014 to "Move fast with stable infrastructure"
Re: (Score:2)
"Pre-meeting"? Seriously?? (Score:2)
"Sometimes there is even a 'pre-meeting' with its own document. I believe the current situation is untenable."
https://dilbert.com/strip/1996... [dilbert.com]
Re: (Score:2)
Root Cause: "Agile" (Score:1)
The Mythical Man Month. (Score:2)
Dysfunctional AF (Score:2)
Every week I see documents with 100+ editors
Jesus Christ, this is absolutely dysfunctional. I've worked with defense contractors, who are famous for the "fat", and even then, I've never heard of anything like this.