zlacker

Bay Area nonprofit Signal shows how bloated tech companies have become

submitted by rexree+(OP) on 2023-11-22 18:11:40 | 62 points 38 comments
[view article] [source] [links] [go to bottom]
replies(8): >>schoen+K2 >>gspenc+88 >>nickdo+wd >>why_at+Dj >>exabri+yw >>Aldipo+uF >>merist+vF >>jaunty+Wh1
1. schoen+K2[view] [source] 2023-11-22 18:22:14
>>rexree+(OP)
I thought the headline meant that this was going to present Signal as an example of the bloat, but it's the reverse: the article says Signal has only 50 employees but still successfully operates a major communications service, where other companies have thousands of employees.

Apart from Signal just generally doing a good job here, I see a few other possible factors:

* Signal doesn't see user content, so it doesn't have a content moderation team

* Signal is designed in such a way that it can't comply with most kinds of legal requests for user data, so it doesn't need a large team responding to those requests

* Signal gets some amount of pro-bono legal help, so it might not have as large an in-house legal team as other organizations

* Signal isn't trying to directly profit from user activity, so it doesn't need to study user activity or engagement metrics with a view towards profiting from them; similarly, it doesn't need to manage relationships with advertisers

* Similarly, it doesn't need to try hard to grow its user base (that would be desirable, but it doesn't necessarily increase revenue much)

* Similarly, it probably doesn't need to try hard to expand into other business areas

(I think these things are generally great. Yay Signal!)

replies(4): >>rexree+33 >>jaunty+ra >>aaomid+ob1 >>g-b-r+yb1
◧◩
2. rexree+33[view] [source] [discussion] 2023-11-22 18:24:27
>>schoen+K2
I agree - this shows Signal in a good light. But it does beg the question what is the real financial incentive discussed in the article for the large tech companies to have such employee bloat? You'd think that employee bloat would be a hindrance to company value but it seems to be the opposite? The more employees it seems the better from a shareholder perspective? It does lead to wild hiring / firing swings but there must be a market logic to it somehow?
replies(7): >>OkayPh+4a >>Zanfa+mb >>denlek+Ee >>s1arti+5i >>mamons+Dr >>olliej+sJ >>brnt+ZJ
3. gspenc+88[view] [source] 2023-11-22 18:43:55
>>rexree+(OP)
To be honest, even 50 seems really high to me. And given how many features that Signal has released recently that I don't use, didn't ask for and resent creeping their way into the UI (chat for example)... I really wish there were fewer people there to keep fixing things that aren't broken.
replies(1): >>OfSang+k9
◧◩
4. OfSang+k9[view] [source] [discussion] 2023-11-22 18:49:02
>>gspenc+88
> (chat for example)

This must be a typo, as chat is the very thing that Signal is meant to provide. If you meant group chats, that was a widely requested feature from the user base.

It’s important to keep in mind that many of the features which Signal has been implementing, are already present in WhatsApp. The very goal of Signal is a chat app for the masses (not just us nerds), but one which actually respects privacy. The one feature that nearly everyone would agree is controversial, was the payments and associated shitcoin, but even that is a case where Signal would have feature parity with widely-used apps, and it exists in an obscure corner of the UI.

replies(2): >>barbaz+bH >>gspenc+TF2
◧◩◪
5. OkayPh+4a[view] [source] [discussion] 2023-11-22 18:51:29
>>rexree+33
Shareholders want to see businesses grow their profit. Once you reach a certain point, it's easier to scale by doing lots of things rather than one thing better and better, and doing lots of things requires lots of people to do them. This creates a weak signal of "more employees -> more future profit". People buying shares (who effectively set the price) care about future growth, which makes looking at revenue enhancement directly a stale signal.

From the company perspective, this is still an alright state of affairs, because even when investors get skittish and less overtly speculative, the company can still improve profit numbers by cutting excess staff. Meanwhile in times of plenty, the hiring of that glut of employees drives the company value higher due to the speculation that they're going to be able to do all the things.

It's dumb, but investing is often a web of self-fulfilling prophecies. If investors think a company will increase in share price, they buy, driving up the share price, allowing the company to sell shares at a higher price, giving them more money to grow.

◧◩
6. jaunty+ra[view] [source] [discussion] 2023-11-22 18:52:49
>>schoen+K2
These bullet points make me think of Signal as a utility; something designed to provide & facilitate, infrastructure that supports.

I wish we could see more of this! There's something so elegant about sustainable software, software not designed to grow & expand but to scale efficiently & effectively.

replies(1): >>Rayhem+Zb
◧◩◪
7. Zanfa+mb[view] [source] [discussion] 2023-11-22 18:57:17
>>rexree+33
Could be explained with cargo culting to some extent, e.g. large & profitable tech companies have many employees, thus we should hire as many people as possible to become large & profitable as well.

Another issue might be that without good management, the marginal utility of each new hire goes down fast. As a company trying to launch new features, fixing management is hard to measure and fix, but hiring is simple to measure and seemingly solves the problem.

replies(1): >>AlotOf+md
◧◩◪
8. Rayhem+Zb[view] [source] [discussion] 2023-11-22 18:59:25
>>jaunty+ra
I wish English[1] had a word to describe the enjoyment felt from using a well-designed tool like this. "Ergonomic" comes close, but I think it's a little too utilitarian - it doesn't quite capture the "this is nice" feeling for me.

[1]: Or other languages, but I only know English. Please let me know if other languages have such a word.

replies(2): >>brnt+8P >>SSLy+671
◧◩◪◨
9. AlotOf+md[view] [source] [discussion] 2023-11-22 19:07:33
>>Zanfa+mb
I don't think it's any of that to a significant extent. Big companies just have thousands of processes, and the body count is just what's needed to sustain all of those processes. You see the same thing in government. Yeah, 50 people can do the work of hundreds because they aren't having 6 meetings with 6 people each to propose, implement, evaluate, review, and commit the font change to the help page, plus employing all the admin staff to support those people. Big companies are.
10. nickdo+wd[view] [source] 2023-11-22 19:08:18
>>rexree+(OP)
I am reassured by the fact they have only 50 staff. I’d probably be even more reassured if they could manage on 25. I speak from experience in having worked for companies with both 30 staff and 30K.
◧◩◪
11. denlek+Ee[view] [source] [discussion] 2023-11-22 19:14:12
>>rexree+33
a company could derive value from increased headcount if investors or shareholders perceive it as a valuable metric and reward the company with more money or a higher valuation regardless of other metrics.

i wonder if companies have seen increased valuations from saying they are hiring for tons of positions without actually following through on the actual hiring

◧◩◪
12. s1arti+5i[view] [source] [discussion] 2023-11-22 19:29:36
>>rexree+33
It makes sense because the market seeks net profit opposed to efficiency or percent profit.

That means it is better to make $101 with 20,000 employees than only $100 with 2 employees.

replies(1): >>schoen+tp
13. why_at+Dj[view] [source] 2023-11-22 19:38:11
>>rexree+(OP)
I think this captures a lot of what I find so frustrating about all the big tech services today. It seems like so many of them could be a similar model to Signal where they have a minimal team and rely on donations or some small revenue stream. Wikipedia is another example of a highly utilized service that uses the non-profit model.

Is there any reason why most social media sites can't be the same way? Network effects account for a lot of it which is why nobody uses Mastodon even though everybody hates Twitter now.

It's just aggravating that you have tens of thousands of engineers working towards things like ads and tracking when those are actively against most of our interests, and the other side of the scale you have fifty people actually working on things useful to users. How do we fix this?

◧◩◪◨
14. schoen+tp[view] [source] [discussion] 2023-11-22 20:04:01
>>s1arti+5i
I guess another place that's especially apparent is film budgets. Very expensive films are going for a chance of a huge payout. A film studio, or its investors, might very well prefer a 50% chance of a $5,000,000 profit on a $20,000,000 budget than a 70% chance of a $500,000 profit on a $500,000 budget.
replies(1): >>holler+3r
◧◩◪◨⬒
15. holler+3r[view] [source] [discussion] 2023-11-22 20:12:43
>>schoen+tp
>A film studio, or its investors, might very well prefer a 50% chance of a $5,000,000 profit on a $20,000,000 budget than a 70% chance of a $500,000 profit on a $500,000 budget.

That's not how it work: a rational actor would take the $20 million and invest it in 40 different films costing $.5 million each if he could make an expected profit of $.5 million on each of the 40 cheap films.

In other words, a rational economic actor will keep adding employees (or any other expense) as long as adding employees increases profits, but will not keeping adding assets as long as doing so increases profits because assets have opportunity costs.

To account for the opportunity costs, investors commonly speak of return on investment (profit divided by amount invested) rather than profit, because that is really what they're trying to maximize. Recasting what I just wrote in the new, crisper language, the rational actor will add any employee, any other cost and any investment to a firm as long as doing so increases return on investment.

replies(2): >>schoen+Pu >>s1arti+ko1
◧◩◪
16. mamons+Dr[view] [source] [discussion] 2023-11-22 20:15:38
>>rexree+33
Well, the logic is/was that an engineer you hire on for 300k is now unable/less incentivized to go launch the startup that would threaten 20 million or whatever of your revenue. Now these employees are rare, and hard to find, so the 300k is what you pay knowing you also end up hiring duds.

It's also better from a managing perspective because if you have 10% of bloat and then have a bad quarter, you then have a lever you can pull to show you are getting stuff done. Its how managers smooth their performance, you overhire in good quarters to lower the profitability comps, and then fire to buffer up bad quarters.

◧◩◪◨⬒⬓
17. schoen+Pu[view] [source] [discussion] 2023-11-22 20:33:43
>>holler+3r
Yes, I guess I was implicitly assuming that the investor only had these two options, which might be a good model some of the time, but not typically. You're absolutely right in the more general long-term case.
18. exabri+yw[view] [source] 2023-11-22 20:43:36
>>rexree+(OP)
Which is pretty ironic considering even Signal has plenty of fat that can cut... average salary is pretty out of this world.
replies(1): >>kelvie+BE
◧◩
19. kelvie+BE[view] [source] [discussion] 2023-11-22 21:24:01
>>exabri+yw
I mean these people seem to be worth their salaries if they can accomplish so much with just 50 HC.
replies(1): >>gamblo+RO
20. Aldipo+uF[view] [source] 2023-11-22 21:28:30
>>rexree+(OP)
Ah, that is why Signal still hasn't an app for Android _tablets_. https://community.signalusers.org/t/android-tablet-support/5...
replies(1): >>dangus+3e1
21. merist+vF[view] [source] 2023-11-22 21:28:32
>>rexree+(OP)
This article prompted me towards donating to Signal.
◧◩◪
22. barbaz+bH[view] [source] [discussion] 2023-11-22 21:37:36
>>OfSang+k9
maybe they mean those "chat bubbles"?
◧◩◪
23. olliej+sJ[view] [source] [discussion] 2023-11-22 21:48:58
>>rexree+33
One of the arguments I saw in years past was that there was an element of “make sure our competitors can’t get the best people by just hiring them even if you don’t need them”. I suspect that the mass layoffs at google etc were at least partially due to that.

There’s also an element of large companies working on many different things, and the more you silo projects the more people you need because many projects end up duplicating work.

You also don’t necessarily see the output of teams at large companies, potentially for years, so you see large numbers of people working but there’s no external/publicly visible product.

But then you also have “we’re successful right now so let’s pretend that the current growth rate will continue forever and hire accordingly” which is mismanagement that eventually needs to be corrected.

◧◩◪
24. brnt+ZJ[view] [source] [discussion] 2023-11-22 21:52:03
>>rexree+33
Like all businesses: they become their own goal. They become tools. Including for sociopaths.
◧◩◪
25. gamblo+RO[view] [source] [discussion] 2023-11-22 22:17:52
>>kelvie+BE
Facebook Messenger serves more than 20x the users with fewer than 20x the employees...

WhatsApp and Instagram similarly serve more users per employee than Signal does, at an average compensation lower than what the average Signal employee makes.

replies(2): >>why_at+US >>Jambal+Xo1
◧◩◪◨
26. brnt+8P[view] [source] [discussion] 2023-11-22 22:19:41
>>Rayhem+Zb
Beauty is what I would use. Maybe craftsmanship.
◧◩◪◨
27. why_at+US[view] [source] [discussion] 2023-11-22 22:41:19
>>gamblo+RO
Do you know the actual number of employees at each of these services? I'm not doubting you, only asking because I'm curious and have no idea personally.

It's interesting if their headcounts are similar, although maybe not entirely unexpected. It doesn't take that many people to maintain a messaging service, so there shouldn't be that much of a difference.

If I'm being generous to Signal, I could say that most of the extra employees at these companies (e.g. Facebook) don't directly work on Messenger itself, but work on the surrounding data collection/ad tech. At the same time, there is a certain number of employees that any company needs just for people to handle the administrative/business side of things, so that explains the extra people working for Signal.

If I'm being less generous, maybe Signal really isn't that different, and Facebook et al just have so many extra people because they are liable when it comes to content being posted on their services. For example they need extra lawyers and content moderators to handle all the stuff that is publicly posted on Facebook. This is something Signal just doesn't have to deal with.

This is all just speculation though, I don't have any numbers that indicate which is true.

replies(1): >>gamblo+eM1
◧◩◪◨
28. SSLy+671[view] [source] [discussion] 2023-11-23 00:02:39
>>Rayhem+Zb
Comfort.
replies(1): >>t0bia_+5Z1
◧◩
29. aaomid+ob1[view] [source] [discussion] 2023-11-23 00:21:16
>>schoen+K2
With all of these, signals experience is honestly ridiculously bad.

Have you tried switching iPhones while actually using signal and having a conversation history? It takes over an hour to transfer.

E2EE Backups are not impossible. I honestly feel like Signal users are lying about how much they’re using signal.

◧◩
30. g-b-r+yb1[view] [source] [discussion] 2023-11-23 00:21:58
>>schoen+K2
Signal can comply to legal requests perfectly well through updates, having done everything they could to forbid third-party clients, third-party builds (F-Droid) and the avoidance of Google dependencies.

But it's likely they leave that job to their beloved Google, since they haven't changed course even after the compulsory handing over of signing keys to them.

◧◩
31. dangus+3e1[view] [source] [discussion] 2023-11-23 00:38:15
>>Aldipo+uF
It should be noted that tablets are a declining market overall.
replies(1): >>Aldipo+kX2
32. jaunty+Wh1[view] [source] 2023-11-23 01:02:59
>>rexree+(OP)
The premise here is somewhat contrasting with what Signal just told us in the blog post, *Privacy is Priceless, but Signal is Expensive." https://signal.org/blog/signal-is-expensive/ >>38291427 (1108 points, 6 days ago, 915 comments).

Meredith & Joshua say Signal costs $14m a year in hosting costs. Storage at $1.3m/yr, servers $2.9m, registration fees $6m, bandwidth $2.8m, and additional servers $0.7m = $14m. They say their total costs are projected to be $50m/year, soon.

I wonder Signal compares to, say, Whats App: another volume-focused messaging system that at least used to be run very lean.

◧◩◪◨⬒⬓
33. s1arti+ko1[view] [source] [discussion] 2023-11-23 01:47:52
>>holler+3r
I think there is real quantity constraints. For the most part, the number of Theatre screens are fixed. You can pack them in or have have seats 10% full.

There is also saturation points for advertising where customers don't respond.

◧◩◪◨
34. Jambal+Xo1[view] [source] [discussion] 2023-11-23 01:53:02
>>gamblo+RO
Users:employees obviously do not scale like that. You need a certain amount of employees just to make a chat app with zero users.
◧◩◪◨⬒
35. gamblo+eM1[view] [source] [discussion] 2023-11-23 05:26:17
>>why_at+US
Pre-acquisition WhatsApp handled 450 million users with 35 engineers paid at Asian tech salaries, or roughly USD$ 30-50k. So WhatsApp's entire engineering salary cost is equal to what Signals pays just 3 of its engineers for 9x the userbase. WhatsApp had roughly 257 users per engineering dollar spent, while Signal has at best 20 users per engineering dollar (but that number is lower because I'm only accounting for the "highly compensated" engineers and not Signal's full engineering staff).

Instagram had 13 employees with 31 million users when it was acquired by Facebook.

I don't know what the post-layoff employee account is for Meta Messenger as they did not disclose it in their SEC filings, but pre-layoffs it was several hundred.

◧◩◪◨⬒
36. t0bia_+5Z1[view] [source] [discussion] 2023-11-23 07:46:27
>>SSLy+671
Any centralised platform makes me uncomfortable to use it.

I like Signal, but it's main weakness is centralisation.

◧◩◪
37. gspenc+TF2[view] [source] [discussion] 2023-11-23 14:27:08
>>OfSang+k9
> WhatsApp

I don't use WhatsApp and wasn't looking for a WhatsApp alternative. It's also fine if these were features requested by the userbase. Fair enough. But I wouldn't know. I was speaking for myself. I just wanted an SMS replacement that did e2e encryption. Nothing more.

◧◩◪
38. Aldipo+kX2[view] [source] [discussion] 2023-11-23 15:59:29
>>dangus+3e1
Still larger then the Linux desktop.
[go to top]