One person changed the code and crashed Twitter. Musk went crazy: Rewrite everything!

This article is reprinted with the authorization of AI New Media Qubit (public account ID: QbitAI). Please contact the source for reprinting.

"Musk is crazy"!

Just today, Twitter crashed again, and pictures and any links cannot be opened.

According to Musk himself, the reason is just "A small change in the API".

But the consequences are very serious. At least in the eyes of netizens, the scene on Twitter belongs to Aunt Jiang:

And the mood of the melon-eaters is:

It’s not that netizens are too mean, it’s just that it does feel a bit like giving back for evil .

According to Platformer, the problem occurred when third-party clients accessed Twitter’s API. Just in early February, Musk decided that this API would no longer be open for free and would be changed to a paid model.

Then the code must be modified accordingly. However, there is actually and only one engineer responsible for this matter.

The Verge said directly and mercilessly: This is all the fault of Musk’s massive layoffs, and Twitter now has too few engineers.

And netizens even mocked Laman:

Except for the children in the nursery, there are no people left on Twitter who are engaged in testing.

Internal employees revealed that Musk was indeed furious. However, Iron Man seems unwilling to take the blame for the outside world. Instead, he feels that "Twitter code is too fragile" and once again made harsh words:

Twitter's code must be heavy Write.

An engineer crashed Twitter

Let’s go over this matter from the beginning.

Netizens who were surfing Twitter early in the morning suddenly discovered that the Twitter link could not be opened.

Not only cannot it be opened, it also tells you that the reason for the error is "Your current API does not include access to this breakpoint."

A simple translation is that Twitter’s own API no longer recognizes Twitter...

As for the specific reason, according to Musk, they made some changes to the API code, which resulted in the entire Twitter crashing.

Platformer learned more details from internal employees: This API is the one that Musk announced on February 1 that it would change from free to paid. This change directly brought the end of Twitter’s third-party client. The detail that makes people feel outrageous is that internal employees told Platformer that only one SRE

(Site Reliability Engineer)

engineer was responsible for the change of Twitter API fees. It turned out that after the engineer made a "wrong configuration change", the API was "basically broken." This resulted in many of Twitter's internal tools and public-facing services being disabled.

This made Musk angry. As he said in his tweet, "For no reason, Twitter's code is too fragile."

The boss was very angry, so the internal Slack

(communication software)

was instantly filled with threads from Twitter programmers frantically putting out fires.

Apparently, Musk thinks Twitter’s code is a piece of shit. But whether it is really "for no reason", there are many different opinions from the outside.

Previously, a former Twitter employee told the Washington Post that after Musk took over Twitter and drastically laid off employees, Twitter’s code has long been in a state of tightrope walking:

Code and operations Any error in is now fatal.

People who stay on Twitter will be overwhelmed and therefore more likely to make mistakes.

Twitter crashes in various ways

In fact, in March this year, Musk’s Twitter has already crashed

6 times

on a large scale.

First of all, on January 23, some Android users were "backstabbed" and unable to view watch lists, load tweets, and send tweets. It took about 4 hours for everyone to return to normal.

△ Picture source 9to5Google

Within a few days, on February 8, many users I also found that I could not forward and send tweets because the system prompted me that "the number of tweets sent per day has been exceeded." It took about 1.5 hours to fix it.

Then, on February 15th, this time it was the turn of iOS users. The "Recommended for You" (For You) module could not be loaded.

Soon, three days later, on February 18, the problem became even more serious. Even the homepage crashed, notifications for replying to messages were not prompted, and various other error messages filled the screen, making netizens panic. I wonder if Twitter is "closing down" (the topic "#TwitterDown" went viral).

Finally, on March 1st, the watch list and recommendations for you collapsed again.

Coupled with the fact that today’s pictures and links are inaccessible, Twitter’s crash events can be described as full of tricks.

——and "common food" to the point where everyone has mastered the skill of eating melon.

For example, Twitter’s own employees:

We are all numb.

For example, foreign media:

Just wait, the next collapse is probably on the way.

Would you like to ask why this is the case? In the eyes of the outside world, no matter how hard Musk shirks the blame, he can't escape responsibility - who told him to lay off all the people.

As soon as the acquisition was completed last year, Musk launched the "extreme hardcore" culture of Twitter 2.0 and sent thousands of people away in a big move.

Later, many "small actions" came sparsely, resulting in the large factory that originally had more than 7,500 employees, eventually losing 80%, leaving only about 1,300 people.

Among them, there are less than 550 full-time engineers (data from CNBC).

An anonymous engineer from Twitter expressed concern about this:

Twitter’s code base is very large, and each part of the website requires people who know different programming languages ​​to maintain it. There are barriers between these skills and it is not easy to "transfer" between websites, so after losing so many people all at once, it is difficult for us to train new engineers.

I don’t know if Musk thought about this issue when he was crazy about layoffs.

But I may have thought about it, but I still "can't control that much" - after all, in addition to highlighting the hard-core culture, the most important thing about his operation is to cut costs and make up for the US$44 billion hole.

In addition to layoffs, there are also a series of actions such as launching paid certification services, canceling free lunches, laying off cleaning staff, defaulting on office building rent and supplier marketing payments.

There are too many flaws, so I won’t expand them one by one.

But, don’t think this is the end.

In the past two days, anonymous employees broke the news to The Insider that Musk even tried to sell office equipment to employees.

In short, Twitter’s turmoil continues.

Some netizens suggested that the explanation of "unexpected circumstances" should be pinned to the top of the tweet for a long time.

As for when this horse-catching drama will calm down, maybe it will be when Musk really overthrows and reconstructs Twitter. ……

The above is the detailed content of One person changed the code and crashed Twitter. Musk went crazy: Rewrite everything!. For more information, please follow other related articles on the PHP Chinese website!