DeepSeek Says Service Degraded Due To 'Large-Scale Malicious Attack' 34
Chinese AI firm DeepSeek said Monday it had degraded the service, only accepting registration of new users with China-code phones numbers, amid a "large-scale malicious attack."
Poof into a cloud of vapor (Score:1)
slashdotting? (Score:5, Insightful)
Or maybe millions of people heard about it for the first time today (as I did) and wanted to try it out. Remember when being slashdotted was a thing?
Re: (Score:2)
Re: (Score:2)
Genies don't go back into bottles. Nvidia and especially Microsoft ought to already know that very well.
Re: (Score:2)
Sometimes, insider trading doesn't need genies to go back into the bottle, just a 24h break.
Re: (Score:2)
staging a denial of service attack to save their asses?
I can't get into half the websites I visit without a 20 minute cloudflare verification. Which stands to reason, why does an independent US IP need to go through that shit? WTF does cloudflare actually do?
Re: slashdotting? (Score:2)
Re: (Score:2)
Remember when being slashdotted was a thing?
Ahh, the good old days!
Re: (Score:2)
Those were the days. I know I had one server get slashdotted.
Re: Poof into a cloud of vapor (Score:2)
Source? (Score:3)
Any news agencies links for this info?
Re: (Score:2)
Even without "overhype" they will have seen a massive influx of new users globally and with this lots of additional traffic due to it being in the news everywhere.
Re: (Score:2)
Probably the result of a million people bombarding it with requests for Tank Man and Xi Jinping Winnie The Pooh photos all at once. The Chinese censorship filter was probably never stress tested that hard before?
malicious users (Score:1)
I don't suppose it blasting to the top of the download charts would indicate a massive influx of users, exposing scaling issues, would it? Sounds pretty malicious to me.
Slashdot effect (Score:2)
I already got a local model running (Score:2)
Re: (Score:2)
What will happen soon.... (Score:2)
2) Other AI based systems start using it
3) New AI system is created by AI systems of 2)
4) Goto 1)
In other words, very soon there will be the equivalen of Kessler's syndrome [wikipedia.org] for AI systems...
Re: (Score:2)
Not this year.
AIs aren't yet capable enough for that. Give it a few years. (I still don't expect a real AGI until 2035 +/- 5 years. There will be unexpected obstacles that need to be overcome. And there will probably be unexpected breakthroughs.)
Source? (Score:1)
Re: (Score:2)
Some people report having a local copy running. But I suspect that it's binary and they don't have source.
are you there? (Score:1)
Re: are you there? (Score:2)
Aam Saltman (Score:2)
If they're willing to suicide OpenAI whistleblowers than a DDOS is nothing
Re: (Score:2)
Oops, then*
oh the irony (Score:2)
They don't like their stuff getting stolen ?
They don't like it when they are just going about their business and someone gets them ?
This is RICH
I hope the people in Taiwan see this.
Re: (Score:2)
There are lots of companies in China that are just commercial endeavors. And it's quite reasonable for them to not like being attacked. (OTOH, I don't know about this one, but they well could be innocent of the crimes to which you allude.)
I don't like being accused for the crimes of the US govt, but that in justice has to mean that I don't accuse people of other countries for the crimes of their government.
It's to restrict queries (Score:2)
That is why there is an "attack" going on and why only people in China, who use a Chinese phone number, are allowed to register.
slashdootted (Score:2)
They got slashdotted.