zlacker

It's an interesting theory but the DDoS predates the decision to disable anonymous access. In fact, that decision was taken in order to mitigate an ongoing DDoS[0][1]. So while it's possible that the questionnable web front end retry logic is not helping things, it's not the root cause.

[0] https://twitter.com/elonmusk/status/1674865731136020505

> Temporary emergency measure. We were getting data pillaged so much that it was degrading service for normal users!

[1] https://twitter.com/elonmusk/status/1674942336583757825

> This will be unlocked shortly. Per my earlier post, drastic & immediate action was necessary due to EXTREME levels of data scraping.

> Almost every company doing AI, from startups to some of the biggest corporations on Earth, was scraping vast amounts of data.

> It is rather galling to have to bring large numbers of servers online on an emergency basis just to facilitate some AI startup’s outrageous valuation.

>>olalon+(OP)
Honestly it probably has more to do with the decision to basically shut down API access without exorbitant fees than it has to do with AI startups or whatever.

Surely there are some colossal archives of all tweets before whatever date floating around that those startups are using instead.

>>olalon+(OP)
That is a post-backlash claim by Elon Musk, so it should be taken with a huge grain of salt.

>>tsimio+jf
It seems more plausible than the alternative. Let's say that Twitter availability issues were really caused by disabling anonymous access and that Elon wants to back out of his decision due to public backlash, wouldn't he just reverse the decision? That would fix both of his issues at once. He could even claim to have solved the scraping issues through a clever trick that doesn't require disabling anonymous access. The fact he hasn't done so indicates that there is a genuine issue that he can't just fix by restoring anonymous access.

>>olalon+2j
The theory is more that Elon wants to prevent public access to Twitter in hopes of making it more profitable, but he doesn't like the backlash to such cash grabs and has invented this story which makes the Auth requirement and non-paying account view quotas more palatable.

It's important to note that Twitter's desires are fundamentally at odds with the desires of the content creators on Twitter on this point. Twitter, ideally, wants to only show it's tweet to real people who will click on an ad, since it costs money to show tweets and bots or people who don't look at ads are not paying that money back. Creators on Twitter want their tweet to be shown to anyone at all, and will leave if they feel people just aren't seeing their tweets.

So, Twitter must walk a fine line when trying to restrict who can see their tweets, lest they alienate content creators. Inventing a common foe is a tried-and-true tactic for getting people to accept changes that hurt them.

>>tsimio+jk
My initial contention was directed towards the claim that Twitter was accidentally DDoSing itself. However, it seems you are proposing another theory - that there is no DDoS at all and Twitter is "faking" its availability issues to introduce quotas and a login requirement. While I personally perceive this as veering into conspiracy theory territory, I concede that it isn't entirely beyond the realm of possibility.

>>olalon+4m
My theory is more:

1. Twitter decides to go account only for monetization reasons, and implements this feature with some bugs.

2. People start complaining about the new Twitter policy.

3. Availability issues start being observed, caused by the bugs in the implementation of 1 - possibly a self-DDoS.

4. Elon, in response to 2, lies about 1 being a temporary response to an external pre-existing DDoS attack. People start associating 3 with the claimed DDoS attack.

5. In a hasty attempt to fix 3, whose exact cause they have not yet determined, Twitter starts implementing stringent view quotas. Since 3 was not caused by an external DDoS attack, this actually only makes the problem worse.

I'm not claiming this is definitely what happened, just that it is a plausible time-line of events. The one Elon presented is also plausible, of course.

The best question that could help us distinguish these two cases from the outside is whether Twitter's availability issues started being observed before or after the authentication policy change. I would say that Elon's claim is far less likely to be true if the noticeable availability issues only appeared after the policy change. Conversely, my version of events (which is more or less the same as TFA's, I think) is far less believable if the availability issues happened before the Auth change.

I will note that I didn't use Twitter at all in the last few days and thus don't know which is the case. On HN, I definitely saw the Auth policy change story at least a day before seeing significant complaints about other availability issues, though.