It is about how companies spy on us.
Some people claim Facebook, Google, NSA and such are collection so much information about us and it is dangerous.
Is information power?.
If the information is transformed in knowledge then yes. Otherwise, it is the opposite. And it’s funny because if we have so much information then we could be unable to process it all.
First, let’s say facebook spy on us (and it does) but, can they use this information for some evil scheme?. Well, let’s see.
Question: Do I play golf?
Let’s say I have 3 friends on facebook and nothing more. Facebook knows that. Now, let’s say 3 of my friends play golf. This information could be converted into knowledge. Facebook could assume that I play golf too, so it could offer me a golf club.
So, we have information -> knowledge -> power. The information is limited but it’s pinpoint some situation. Maybe it’s wrong but there is a big chance that I play golf.
Now, let’s say I have 500 friends, 70 of them play golf, 80 of them fake that they like running a marathon and 40 play sky. So, what Facebook knows?. How it could turn this information into knowledge?
The information is not conclusive so it needs more information.
Let’s say Facebook read my conversations. One of my friends talks to me about Golf. But I have thousand of conversation about many topics.
So, the more information they collect, then more information could be misleading or simply false.
Facebook knows some information about us such as age, sex, etc. but the rest of the information is guessed by facebook.
However, let’s say Facebook knows all about me, everything. But, to achieve it, it needs to process it.
- Facebook’s revenue is $55.8B (per year)
- Facebook has 2.3B active users.
- It means that its revenue $20 per user (revenue is not profit, btw).
In gross terms, let’s say Facebook collects every bit of information about the users. It costs money, a lot. Storing 1tb of information per user costs around $10-$20 per year (It could be cheaper for Facebook). Right now, Facebook doesn’t have 1tb per user but it’s trying to go there. So, let’s say it costs $5 per year.
$20 (earning)-$5 (storage) = $15 profit.
But storing is cheaper in comparison with processing. Let’s say it process all of this information. It costs a lot and it requires constant processing. Processing an image could cost $0.01 per image (why? CPU, energy, etc.), and with 1000 images, the cost is $10. And of course, we aren’t counting to process every post and information.
$20 (earning)-$5 (storage)-$10 (processing)= $5 profit.
But Facebook needs to pay the salaries. Facebook has 39,651 employees and let’s say $70k per employee, around $1 per user (this doesn’t include contractors).
But also Facebook needs to pay infrastructure (the electric bill is already mentioned in the costs) but it also needs to support FACEBOOK (as the visible front) and again, it is expensive. It could cost $5 per user easily. Why? Because it is not a simple service, it is a service where the user is constantly using resources of facebook.
$20 (earning)-$5 (storage)-$10 (processing)-$1(salaries)-$5(facebook)= -$1 profit.
AFAIK, Facebook doesn’t pay taxes so it saves a lot of it.
Is Facebook burning money?
Not really. Simple, it is not doing all of this.
What they really do?. They read the last posts or comments, the put the user in some profile and (with luck) they process your images (and audio). But they say to the marketer: “we do the best to obtain the best profiles”, IA, Deep-IA, Ultra-IA, Big Data and such, however, it is bs! Why? Because technically the technology is available but it is a hell of expensive.
It is the reason why we see some ads that are totally off marks. Simple,to process all those information is incredibly expensive, so what is the solution of the technicians of Facebook?, they don’t, but they say that they did it.
Facebook works a micro-level while it fails completely at macro-level (but macro-level keeps the business running).
For example, Facebook could spy on some specific individual and it could eavedrop and analyze all his/her information (including the use of human analysts) but for the masses, Facebook is hard guessing.
Google is different, every website tries to do it’s best to give the best to Google-Bot(and precise information). So Google-bots simply collects it. It’s pretty simple, it is also expensive but it works. However, it is not AS expensive as IA or a deep-analysis.
However, Google also spies on us. And it also fails too. A good example is Youtube. The suggestions of youtube are clearly off-marks also the ads.
They are public employees. It says everything. While there are good public employees but it is not the trend. Most public employees can’t be fired, so they get lazy. Also, the State doesn’t pay for the best, if they fit for the job, then they are hired. There are a lot of studies on it.