[sjtindell]

I’ve become convinced that hiring more people is generally a warning sign. A team of between a few and about twenty people seems to produce all the good stuff. Anything beyond that, your value per employee drops drastically, and your cost per employee goes up as you need to offer higher and higher salaries to stay in competition. A company going on a hiring spree is now a negative indicator for me. I think FB, Goog, etc. could lay off thousands with no adverse effects and in fact an increase in velocity, quality, and quantity of new features/products. Sometimes one person can be much more productive than a whole team. This is obviously unscientific, just a feeling I’m getting lately.

[asien]

> I think FB, Goog, etc. could lay off thousands with no adverse effects and in fact an increase in velocity, quality, and quantity of new features/products.

You can add Airbnb,Netflix,Uber they often attend conference to describe their architectures.It’s obvious most of these people have no idea what they are doing have no clear direction. They are just havin’ fun will trying to navigate corporates politics. Even the stuff that is published online it’s scary to see their is no technical leadership what so ever.

To be fair I’ve worked in fortunes 500 as well, 60% of the workforce can be replaced with automation.

Since it’s cheaper and less risky they just keep hiring people for repetitive tasks , it compensate the technical debt.

[jiggawatts]

There was a post here recently where a Netflix employee was proudly showing off their log processing system. Which was collecting the equivalent of nearly 2 MB of logs per minute of user streaming time.

In my mind, that's just bonkers, and no amount of handwaving could justify it.

[mistrial9]

> Which was collecting the equivalent of nearly 2 MB of logs per minute of user streaming time.

to clarify could you say the same thing, in a different way?

[Petersipoi]

For every minute that someone streams Netflix, 2 MB of data is logged. So if 1,000 people are using Netflix simultaneously, they're generating 2 GB of logs per minute.

[exdsq]

Warning: NPM packages are out of date x 1000000

[asien]

> 2 MB of logs per minute of user streaming time.

2MB/minute is 33KB/second.

How is that impressive?

[jaywalk]

I think it's impressive that they somehow found 33KB/second worth of data to log for each stream. I can't even imagine the amount of useless shit that must be logged to get to that number.

[westpfelia]

This is where I'm at. Like thats honestly not much log data. But what are they actually logging? I imagine there is a LOT of repetitive data.

[seti0Cha]

Detailed logging can function as an on-demand APM. Not a bad idea if you have the bandwidth and storage for it.

[zwily]

I think the impressive thing is how much data that is for each user-minute. What could they possibly be storing in 33KB for each second of Netflix you stream?

[bpicolo]

That's per user. So a million (or ten, 50...) active users means a lot more per minute.

[schmeckleberg]

i think you and the above poster are in vehement agreement. ingesting 2 MBs of logs per minute is impressive in its pluperfect unimpressiveness.

maybe the presentation was called "Timmy's first named pipe" or "Sally explores /etc/logrotate.d"

[FredPret]

That’s a LOT of text to describe me sitting on my couch. 2MB per minute is far more than the most detailed biography in existence.

[exdsq]

220m users. Let’s imagine 50m are streaming concurrently. That’s 100TB an hour in logs lol. They could be storing an entire petabyte of logs a day. My friend did some data center stuff for the large hadron collider and wasn’t hitting these data ingestion states, and these are just to record me binging the office.

[ummonk]

The comment said "log processing system". Sounds more like it's a stream and not stored logs.

[eightysixfour]

2 MB/per minute/per stream at Netflix scale is crazy.

[lbriner]

There is a lot of tokenism in hiring. A Fortune 500 might be marketing a new "push to AI" or whatever and want to seem legit by hiring loads of people, quickly realising that it doesn't work like that but at least it looks good on paper.

FANNG type companies are more likely to do a big hire after doing a big raise. Imagine someone has given you $300M, what do they expect? Now you have the money, we want more features = more sales = more ROI. How do we do that? By hiring a load more people and again learning that it doesn't work like that. Leave it a year or 2 and the same investors complain about burn rate so you lay them off.

[matwood]

Just a note, that ABNB did a huge layoff at the start of the pandemic which allowed them to come out the other side a much stronger company. Actually highlights your point.

[meowtimemania]

Was it the layoff that made them a stronger company or was it the market improving?

[matwood]

Both. When money is free flowing it's easy to avoid hard decisions (in business money hides many mistakes). Companies may continue to fund projects that should be cut or hire instead of optimizing a process. Prioritization alignment meetings end with everything is a priority.

In ABNBs case, business going to almost zero overnight was a forcing function to a level not often seen. After being forced to lean up and prioritize, they were well positioned for the market to improve.

[Apocryphon]

Seems the problem is that these massive companies hit a threshold in size and then everything is about self-perpetuation by creating large moats, even ones that don't make sense, hence you have teams and entire departments engaged in boondoggles that are wastes of time and resources.

Imagine Facebook pouring untold manpower and money into developing original content such as cloning HQ Trivia, for its also-ran streaming content that no one watches. Or even Facebook Reel, which mostly just reposts TikTok and Instagram material. Or the entire hopeless arena that is cloud gaming, where all of these tech companies are involved in with no service that has really taken off yet.

I suppose if the regulatory environment was to correctly deter these companies from staying so big and content and engaged in wasteful behavior, there would be actually more companies, and all of those people in the companies you mention would be distributed across smaller, nimbler, more customer-focused firms, with more competition and thus better choices for consumers. That's the theory, anyhow.

[captaincaveman]

Yeah but which thousand ...

[jen729w]

I’m fascinated by the idea that, once a company reaches a certain size – let’s say 100 – then it is, by definition, average. Any claims to be above average may be summarily dismissed.

This exhibits most clearly in large IT integrators. I know, I work for one. The marketing spiel makes me feel sick.

If I were ever to start a company – I probably don’t want the stress – I’d want to keep it small.

It is for this reason that I find Apple fascinating. They have their issues but if there is a company that seems to have somehow escaped this jinx, they are it. (I know, I know, YMMV. Please don’t turn this in to an Apple thread.)

[samwillis]

They way I had it described to me was that the optimal number of direct reports to one person is 7 (hence why military squads tend to be about 8 people). All companies go through growing pains as they reach ~7^n people and have to put in another layer of management. So about 7, 50, 340, 2400…

In reality most companies will have a 2-3 founders running it and so those numbers are more like, 14-20, 100-150, 700-1000.

I have never worked somewhere with larger teams so I may well be wrong, however there is definitely a step in small businesses with two founders at about the 20 people point.

[encrux]

This seems surprisingly reasonable for my personal experience, so I'm gonna use it as a general rule now /s

[ethanbond]

Inquiry like this should be left to the people who experience it, and of course all they have is their personal experience and then opportunities to share their personal experience.

Do we really need some outside authority to tell us what we’re seeing and how to talk about it?

[influx]

The term you are looking for is Span of Control.

https://en.wikipedia.org/wiki/Span_of_control

[jpgvm]

Netflix for many years was an outlier because of the stupidly high bar they kept for engineering hires. Obviously that all fell apart a few years ago and have since diluted their talent pool into mediocrity but I still think it's a counter-point for the argument that you can't maintain a high quality team into the ~200 engineers range.

[highwaylights]

I can't speak to whether your assertion is true or not (I've no experience of Netflix or it's hiring), but it kind of makes sense that this would happen if you think of the Netflix journey to date.

They transitioned from a mail order DVD system to online streaming, building that as they went, then moved the lot to the cloud while keeping everything up during it's exponential growth period.

There are a lot of really hard engineering challenges just in that sentence that would need some really good people, but now that they've built out a planetscale video delivery system their needs are a lot different (they just need to keep the lights on and the engine running).

I'm sure some really clever engineering still happens at Netflix but mostly the hard problems appear to be solved at this point, so you don't need incredible resources. Also, it would follow that most of those people would want to seek out new challenges rather than stick around for maintenance over the long term, but I've no idea if this actually happened there.

[jcims]

I'm curious why Zoom doesn't get more shine.

In my opinion they were tested like few other companies in history at the start of the pandemic and seem to have pulled it off. My former employer had 60,000 people with technical roles and we all went to zoom in the summer of 2020 after our existing conferencing solution completely shit the bed. I won't say there were zero issues but the number definitely rounds to zero.

Obviously there are the privacy questions, but that's orthogonal to the engineering challenge of delivering realtime voice/video to millions of people.

[ecshafer]

I think Zoom doesn't get more shine for the same reason Alibaba, Baidu, or ByteDance don't. They are large companies doing complex engineering, but they are Chinese and not western.

[sargun]

I'd say that in systems engineering, alibaba gets a lot of love. Alicloud folks have actually done a lot of impressive stuff with k8s

[throwaway744678]

Not sure I am getting your point... Zoom is a US company.

[hardware2win]

> My former employer had 60,000 people with technical roles and we all went to zoom in the summer of 2020 after our existing conferencing solution completely shit the bed

Do i understand it correctly?

60k ppl left ur job and went to zoom?

[jcims]

I could have stated that more clearly. 60k people started working from home full time and we switched completely away from our existing solution to Zoom as the primary service provider for teleconferencing.

[TheJoeMan]

What’s interesting is in my opinion Netflix engineers made the best UI and best reliability with distributed edge-node storage and all those hard problems.

Disney+ still lags ridiculously, but guess which one has the blockbusters?

[jpgvm]

Yeah I completely get why it happened it's just sad that it did.

Because of how markets work in the US their management come under immense pressure to find other sources of "growth" now that the main streaming business has capped out.

TLDR engineering hiring got a lot less selective in pursuit of accelerating "new growth areas". Once that happens your A class engineers leave and you can't ever rehire at that same level again without majorly taking out the trash.

[hhmc]

> Obviously that all fell apart a few years ago and have since diluted their talent pool into mediocrity

This could well be true, but it seems far from obvious to me. Do you have any evidence? Clearly netflix stock is doing poorly in recent months, but this seems much more of a consequence of content issues rather than engineering talent.

[bsenftner]

Netflix has a strange technical management philosophy: they have zero interest in any technology other than their existing streaming infrastructure and traditional media production tools. I've had conversations with members of their tech management, and their ability to investigate new technology directions (such as a Netflix Game Service, or a Netflix Online University) are absolutely zero. One of their identified sources of success is a narrow focus, and they push that commitment.

[kderbyma]

they need product help badly. I have never seen such a stagnant tech company. its like watching paint dry

[kderbyma]

they hired the wrong engineers. they focused on infrastructure....no product

[jpgvm]

Tell me more about how they didn't manage to create a compelling streaming product that failed to gain any market share.

Oh wait. That isn't what happened.

Is their product now stagnant? Yes. Are they facing market saturation problems and increased competition, withdrawal of content from competitors that were previously partners etc, sure.

The engineers they hired were 100% the right engineers, they made streaming work on an Internet that was much less capable than it is now (over 10 years ago), were the first to successfully use cloud computing properly and laid the foundation for and/or contributed to much the of tech the rest of large distributed systems would be built on.

The fact they blew the transition and haven't successfully branched out doesn't in any way discredit the achievements of the "real" Netflix team that changed media delivery on the Internet forever.

[hardware2win]

>Netflix team that changed media delivery on the Internet forever.

Genuine question

Whats the diff between them and yt?

[jpgvm]

If you are talking exclusively about delivery systems. Bitrate.

At it's peak Netflix accounted for ~35%+ of traffic on the Internet (around 2014-2015 era) alone. Times have obviously changed since then. Not because it had more users than YouTube or more playtime but because it was delivering long-form content in high resolution. Which at the time was a considerable technical challenge, from bandwidth and distribution itself but also to encoding infrastructure.

More broadly Netflix also had a profound impact on the CDN landscape. They made use of all existing CDNs (Akamai, Cloudfront, old stuff I forget) while also building out their own POPs, they paved the way for modern CDNs in many ways. The style of which would eventually be cloned by many players but most importantly Fastly. Fastly would then go on to provide this service to new players like Bytedance to rapidly deploy massive delivery capability with very little ramp-up. They did some less-cool things ofc, they bowed to the HDCP mob and did dumb things like build a Silverlight based client etc. By and large however Netflix has had a positive impact on the industry.

YouTube on the other hand (after acquisition) was largely built with Google tech on Google infra. The only major project to come out of it isn't delivery related is Vitess - a sharding system for MySQL. Its largest impacts on the industry are unfortunately much less savory. Mostly pioneering inventions like pre-roll ads and targeting based on content in which the video is embedded, heinous recommendation algorithms that send people into procrastination spirals etc.

Well that isn't entirely fair, we can thank Youtube for pushing for better encoding standards and tangentially through Chrome making HTML5 video not suck (arguably though, Safari/Webkit was also influential there).

[KingOfCoders]

I also found working in smaller companies more effective (worked in 20 people, 500 people and 30.000 people companies).

DHH recently said at a conference on how people always tell him they were happy when the company was small, and asked "Why scale then?" (the panel agreed that some things, like building a commercial airliner might take large teams).

This has been to my heart for years https://www.radicalsimpli.city/

Looking at HN, enjoying the single founder SaaS threads, this is the future for a lot of business models.

[abnercoimbre]

Do you believe we're favorably headed that direction already? Or would it require a lot more conferences highlighting this, getting mindshare buy-in, etc?

[KingOfCoders]

No not yet.

People don't want to have meaningless jobs [1] and want jobs that make sense to them. But the other force is VC driven startups where the mantra is "Scale, scale, scale" by hiring hundreds of people - with the goal to make people rich. I do assume for more and more people money is no longer a substitue for sense though, several of my coachees want new jobs with less pay but more sense. I do believe this becoming a mega trend for the next decades (AI, 3d-printing, ... will accelerate this trend). We're at the very very beginning.

[1] https://yougov.co.uk/topics/lifestyle/articles-reports/2015/...

[xtracto]

Back when i was testing so setup a startup I proposed a crazy idea to my cofounder: Let's setup an equity programme where every time a new person is hired, a chunk of equity would be taken directly from the existing employees(including founders!!), and would go to the new hire.

So that 2 founders would start with say 80% of the company (20% was for VCs) , and after hiring the first hire, he will get 20%, and founders will be left with 60%. After the 2nd hire 1st hire would give up 5%, and founders 10% to get 15% ...

My idea was that this would disincentivise hiring new people, so that new people would be hired only when absolutely necessary .

I'm still planning on experimenting with this at some point in my life.

[yardie]

Doesn’t this already happen at most startups. Each round dilutes the founders share and dilutes the shares of existing employees. The VC might have targets like, acquire 10m users or generate $40m in revenue, and that might not be possible with a team of 5.

[lbriner]

That's only if you raise another round and don't have enough equity for the founders to just gives some of theirs. OP is talking about something more automatic for every new hire.

[DavidPeiffer]

I could see that working while hiring is exclusively by the founders, but what would be the plan after there are departments with mostly autonomous hiring functions? Would the allocation continue to come from all employees, the manager of that department, or everyone upstream?

It's an interesting concept, but it feels like there would be significant pitfalls any way I think if it.

[xtracto]

Not only that... there's the issue with stock options already bought: once the employee bought the stock then it wouldnt be possible to take them from her. And when do you "freeze" the stock amount of an employee? When they leave?

There are lots of details that I thought about, and surely many others I didn't. That's why I'm still ruminating the idea in my head, until I feel it is complete enough to apply it.

[peyton]

What you’re describing is already standard practice. You have to shuffle option pools as you hire more people and raise more money.

[xtracto]

No, when I was hired as a 1st engineer at a startup, I was given 2.5% of the shares.

When we hired the next 10 people, the value/% of my shares did not get affected. So, I in a way I, as an IC with 2.5% of shares could push to hire more people to do more of the work I did without any negative direct impact to me (I know the impact comes as the company spends more money, has to raise more and then dilute the value of the stocks in theory, but that's something employees rarely see).

What I want is being able to sit down with say, the team of 10 people in the startup and ask them: Hey, do we REALLY need to hire this 11th person? we all are going to have to give X% of our share of the pie to him. And particularly, when starting the project, the first 5 or 6 should have a sizeable chunk of the pie, and not 2.5% ... ideally, a 1st employee would get to 2.5% once there are about a 100 people in the company (i.e. after hiring the other 97 people)

[dlubarov]

> When we hired the next 10 people, the value/% of my shares did not get affected.

Where did the new hires' shares come from, if not dilution?

They might have come from an option pool, but that just adds a layer of indirection. When the option pool runs low, typically it is replenished with new shares, diluting existing ones.

So I think the dynamic you're describing already exists in most startups.

[peyton]

> Hey, do we REALLY need to hire this 11th person?

This conversation happens when fundraising. Usually the lead wants to see the option pool replenished at the same time.

[lbriner]

As others have said, there are some gotchas with the idea but you could probably do something similar that doesn't involve stock like with a profit-sharing fund or a fixed amount to spend on your socials or something!

[thrwyoilarticle]

>Anything beyond that, your value per employee drops drastically

True but not fatal. Being equal to the first employees isn't a requirement: as long as the employee creates more value than they cost, it's worthwhile. Facebook is a simple premise that could be (and was) created with a skeleton team but it's still worth hiring the person to work on Linux's networking stack if that enables them to make it more efficient.

[FredPret]

Here’s another thought: Decreasing communication costs decreases the optimal organization size.

In the days of horse-delivered mail, you wanted all your employees in one big building, and you wanted to cover all the functions inside your own org so you don’t have to send many slow, expensive letters.

Fast forward to email and zoom and automated SaaS businesses - there is very little friction in engaging a third party to do the things you don’t specialize in. You don’t need any employees beyond your core team anymore.

[bigpeopleareold]

I say I will agree with this first.

The worst company-level experience I saw in my career was seeing over-hiring only with a massive downsizing later on. It was during that time, that first employee I was directly involved in hiring was part of the layoff. A company that goes acquisition crazy and trying to boost staff rapidly seemed at first like a good thing. In my case, I saw sales staff grow a lot. However, executive leadership, with whatever responsibilities they have, will always have different agendas. Hiring sprees can probably build credentials for managers, whether or not it is a good investment in the first place.

Hiring sprees and acquisitions towards some business goal that doesn't match the current product goal can also cloud judgement on what can actually be delivered. If your product is optimized for a certain set of things, but it is being sold for other use-cases because "meh, we need to compete", you create a multi-year issue when that business plan fails (in an R&D perspective, since that's only what I am interested in.) I bet this is exactly going to happen to Coinbase. Just jumping on that NFT craze instead of developing out their core business more is probably an example that can be relatable (but I am speculating here.)

[ackbar03]

Failure of the nft platform is cited as a negative by the employees here though

[xiphias2]

The worst thing in having too many people working on the product is increased code size / code change velocity.

Compile time per person and code understanding time per person grows linearly with the number of people (which grows exponentially as revenue grows), which means that the total time people are spending with understanding the code base & total compilation time for the code server grows quadraticly with number of engineers (+ exponentially with revenue).

Total code size should be kept under square root of number of engineers in an organization probably to keep product velocity OK, and engineers should spend significant time minimizing the number of changes in the code base after they have the first proof of concept.

[kilroy123]

I too think this and I have no data or anything to back this up. Just a cut feeling as well.