One of my current client is a big institution whose decisions affect the life of probably 700+ millions people a year.
A lot of critical processes are just excel sheets that people open and run once in a while, maybe copy / paste or click a button in there, then close.
Some processes are automatized, but there is no way to check if the result is good. Best you can do is to have a human look at the numbers and say "LGTM" with some confident face and call it a day.
Said client handle billions of euros every year for other entities. They basically influence the entire world economy on a daily basis.
Now, get audited from top to bottom, from internal and external teams, and all audits pass eventually. I met mostly good people there that take their job seriously. But still, every time I dive deeper into their rabbit whole, I just can't believe what humanity stands on.
There was one French artist that made a comic with 2 characters once, saying:
- computers are amazing, they can calculate 20 millions operation every second, with only one mistake every 10 million of operations
- 2 mistakes a second, you call that progress?
Today, the fun part is, the human unreliability is now a small part of a big automated chain that have large, numerous and instant results. So if the excel file is not opened, and the calculations updated but the human, the rest of the system will work with stale values. And who knows what it will do to the cascade of events handling the billions of euros?
In fact, my client has a failsafe for that: a designated chain of people that are supposed to take over if one cannot do their task, for each critical task.
Many government and academic institutions do in fact publicly publish all salaries, so the question of what the effect would be is not entirely theoretical.
Governments and most education institutions are not part of the free market. Their income also does not depend on their efficiency or value they create so its not the same comparison.
yup. my salary and everyone else at my organization is published every year by the local newspaper. We aren't required to publish it ourselves, but if anyone asks, we are required to reveal it, so the local paper files a request every year and publishes it.
- Use Android without a google account or a play store
- Use google products as little as possible
- No whatsapp, no Insta, no FB
- Use pcloud encrypted solution for online storage
But even with all that, I'm pretty sure the impact is limited, real privacy would require a lot more. Not to mention banking, power companies and so on are going to sell your data no matter what.
Besides, most people would not even do 10% of this. Not even switch to FF.
And if they are transparent and knows you, they give a lot of data on you anyway.
Had to deepl "自主可控、纯中文国内社区" to understand that your comment meant it translated to "pure Chinese domestic community", which implies if we are not part of the gang, we can't enjoy the project.
Well, we've been doing this implicitly with English for a long time, so I guess Karma is a bitch.
On the other hand, I wrote a French technical blog for years, and it was way harder to reach a lot of people, so if I had to do it again, it would be in English.
They don't have this problem, since 1B people are Chinese.
This doesn't seem like an appropriate conclusion. There's a massive semantic difference between "China #1" and "Sorry, I only know English." Karma doesn't come into play here.
Because the most watched youtube videos are music videos, and people listen to the same music regularly.
Also because some videos have been popular enough that people wanted to show that to their friends, and regular users don't know how to use the history, so they want to find it in their home page again.
I assume the algo just doesn't know the difference between videos that are rewatchable and ones that are not, and does that on some statistical metrics that don't make sense in a lot of contexts.
When you think something is stupid, the answer is usually there is something we don't know about it.
An algo that does not know what the user wants but tries to do something anyway, based on "some [..] metrics that don’t make sense [..]" is by definition "stupid".
I'm sure that many of old videos has been removed. And the more I use the Internets the more I consider to save as much of stuff as possible (webpages, videos, everything) maybe on another machine with a really giant disk.
Google: "we have determined that drivers only use their brakes 5% of the time versus the accelerator 60% of the time and therefore have decided to remove brakes from the car"
So if most people supposedly don't use the feature it probably doesn't cost that much to have it. Also now people who actually want to view videos starting from oldest will use a magnitude or two more resources just by needlessly scrolling for 5 minutes.
Seems to me that it's more likely that google just has too many employees who just do idiotic/random changes to show that they are actually doing "actual" work...
The first amazing result would not be for you to program with this, but for somebody with a phone to be able to automate a few small tasks with just the voice.
A lot of critical processes are just excel sheets that people open and run once in a while, maybe copy / paste or click a button in there, then close.
Some processes are automatized, but there is no way to check if the result is good. Best you can do is to have a human look at the numbers and say "LGTM" with some confident face and call it a day.
Said client handle billions of euros every year for other entities. They basically influence the entire world economy on a daily basis.
Now, get audited from top to bottom, from internal and external teams, and all audits pass eventually. I met mostly good people there that take their job seriously. But still, every time I dive deeper into their rabbit whole, I just can't believe what humanity stands on.