* $19 million for 50 staff
- That's $338k/head on average. At face value for a nonprofit, I'd like these costs broke down as this seems excessive. There is far cheaper IT labor available outside SV.
- I'd drop these features if possible, or give them to donors.
- I was actually expecting this to be far higher
- Long term storage should probably be donor-only
- Servers could likely be optimized by going hybrid cloud with colocation and owning own hardware, but again, was surprised how "little" they're spending on this.
- Stop contributing and supporting the "Your phone number is your identity" problem.
- Move towards helping educating society and establishing a set of encryption keys as their long term identity
So it was the best of all the available options practically, if they wanted to grow and retain the users.
- That's $338k/head on average. At face value for a nonprofit, I'd like these costs broke down as this seems excessive. There is far cheaper IT labor available outside SV.
there’s just a bunch of nonprofit employees or personnel that play on the pauper perception because its convenient, but “nonprofit” and no money is not correlated to anything
so if those employee costs were excessive for any organization, saying non profit doesn’t make them more or less excessive
I think tech talent is undervalued and should at least compete directly with FAANG, for many organizations this is not possible, for organizations with other liquid assets they create (like Signal) it is possible. All employment hasnt risen with cost of living, I’m not familiar with other sectors.
That $19M/year was total employee costs which, as best I understand these things, can often work out to be double the raw salaries which would bring the average down to a slightly less excessive $170k/year.
Oh come on. Just because the organization is non-profit, meaning that it's not out to make a profit for shareholders, is no justification for the staff to be paid below their market worth. In fact, they could definitely earn more by quitting and working at for profit companies. And that is especially true for those who are getting the higher end of the compensation.
And say that staff number was like, $5m/year less? It doesn't change the fact that costs of running are substantial and more donation is needed from those who want it to remain viable.
How about they pull their socks up and use peer to peer technology instead? Messages are asynchronous so they need to be temporarily stored but routing real-time audio and video is a technology problem that they have chosen the expensive way to solve.
This is a product that solves some of the harder problems of engineering, and has a staff of 50. Cheaper isn’t going to get you the best. If you had a staff of 1000, you could make that argument. Besides that’s not a lot of money to begin with. 340k is a senior engineer salary and I am sure the people running the company are far more capable than senior engineers.
> drop those features
That’s a valid argument, but 1.7M for that 20PB of bandwidth is not a lot of money. Dropping or making the features paid, defeats the purpose. If you’re trying to be the privacy first app that competes with WhatsApp and others, this would make it harder to be a viable alternative.
> sms registration fees
Education is a harder problem to solve, but offloading some of the costs to users may make sense here.
How did you compute this? 19/5 is 3.8
https://signal.org/blog/sms-removal-android/
edit: wording, forgot the word remove
To give one example of a (not that cheap) market, outside of London average developer salaries are probably under $50k in the UK. Even accounting for additional costs like taxation and equipment, that's likely to be under $100k fully loaded.
They can't really do that, it deters adoption of something with a network effect.
The real issue here is that direct connections have privacy implications (maybe you don't want the other party to know your IP address), so they relay everything. If they could solve that they could save a lot of money.
For example, detect if the user is connected via a known VPN service (which is likely given Signal's user base) and then let the VPN hide the user's IP address instead of Signal having to pay for it. Or make a deal with popular VPNs to put the relay servers in their data centers, which gives a similar advantage and they might be able to get better pricing from them in general because the VPNs already have a lot of bandwidth, are sympathetic to what Signal does and could use it as PR.
Actually it does usually. Because when people see real meaning in their work, as opposed to find yet another way to manipulate people on other peoples behalf, then you don't have to buy their consciousness as well.
So sure, it is awesome, that signals employers get to have meaning and money. But I would bet, you would find competent people working for less. (And maybe somewhere else)
But .. they do have a working app and organisation right now and drastic changes could destroy that.
For top-notch security developers, I call bullshit. Signal would be worthless if it started offshoring development to nickel and dime.
I don't think there's industry numbers for that set of people in the UK, as it's not a big enough set. However I'd be surprised if they were 150K plus though, that's a very rare salary in the UK.
Also there are cheaper countries than the UK who have great devs.
The people behind Signal pioneered end-to-end encryption, and as is pointed out in the blog post, there's still a lot of novel cryptography development involved in building a privacy-first messenger. You can't do that without top-notch talent.