Miscellaneous News

Temstar

Brigadier
Registered Member
Please, Log in or Register to view URLs content!

While Trump spurns Canada, China opens the door​




I'm honestly quite surprised at how flexible Canada is at changing the narrative. "Aw, look at how the Chinese want to be our friends again".

View attachment 144231

Came across this image on X today
Broke: Give up your Chinese citizenship to gain citizenship of another country, particularly 5 eyes
Woke: Keep renewing PR, retain your Chinese passport
Bespoke: Gain foreign citizenship but marry a Chinese citizen, particularly a CPC member to cover all bases and gain Chinese PR through your spouse if it becomes necessary
 

lube

Junior Member
Registered Member
I'm not an expert on AI so correct me if I'm wrong but it's weird of the naysayers are somehow saying because DeepSeek doesn't answer China sensitive questions, it's a failure. That's intentional just like I've read that US models don't answers questions that are sensitive to the West. The problem is if it wasn't intentional and didn't answer correctly. It's say more how US AI is going to lie and hide the truth to everyone.

I wonder if OpenAI is offline because of DeepSeek. Why would you want to pay $200 a month for something you can get for free? Yes just like China coming out with 6th Gen fighters, the usual suspects are saying the US has more advanced ones that they aren't revealing just yet, they're saying the same over AI models. Like China doesn't have their own models they're working on in secret? I've been reading how DeepSeek came out with this two months after OpenAI's debut even though theirs is closed. People are wondering why China made this open-source. China is just showing off how they can make theirs better than they could.

Interesting the timing. Some AI conference is happening and this latest comes out around the same time. Trump announces Stargate too that they're investing $500 billion when DeepSeek costs are in the millions to OpenAI's billions. You know people are thinking about that when DeepSeek comes out throwing a sabot into their gears. It goes to the accusation US tech giants are just thinking about how much money that they can get in investments on selling a dream of the future and not what's realistic.

Thought the real reason is concern trolling. The shitlib idea is that Chinese AI labs will be crippled by censorship because the model itself is compromised and researchers get locked up for a naughty chatbot gone wild.
There's literally nothing but vibes backing that speculation up.

And we can see the easiest way to apply censorship doesn't affect the model itself. Literally just least-effort compliance by adding a layer on top to intercept prompt text. That's what openAI does with the names of people who sued them. It could be different in the future, sure, but nothing to suggest that's what's happening right now.
 

siegecrossbow

General
Staff member
Super Moderator
Please, Log in or Register to view URLs content!

While Trump spurns Canada, China opens the door​




I'm honestly quite surprised at how flexible Canada is at changing the narrative. "Aw, look at how the Chinese want to be our friends again".

View attachment 144231

Came across this image on X today
The PRC is helping Asian Americans finally grow a spine. Good for them!
 

Blitzo

Lieutenant General
Staff member
Super Moderator
Registered Member
I heard that too so I did have a close look at the photo first before posting:

View attachment 144234
View attachment 144235
Seemed plausible to me but I'm not very good with faces.

I think if one wants to associate an image to the name, using the definitively confirmed image is better, and you can't get any more confirmed than the meeting with Li Qiang.

I've seen that image of the guy in the suit on a few articles but I'm not sure where it's actually from and whether it's even the same person.
But considering how lowkey DeepSeek have been in "marketing" themselves, that image seems somewhat inconsistent with their self portrayal.
 

9dashline

Captain
Registered Member
Thought the real reason is concern trolling. The shitlib idea is that Chinese AI labs will be crippled by censorship because the model itself is compromised and researchers get locked up for a naughty chatbot gone wild.
There's literally nothing but vibes backing that speculation up.

And we can see the easiest way to apply censorship doesn't affect the model itself. Literally just least-effort compliance by adding a layer on top to intercept prompt text. That's what openAI does with the names of people who sued them. It could be different in the future, sure, but nothing to suggest that's what's happening right now.
In case anyone doesnt know, DeepSeek released two R1 models, the "zero" version used pure RL and nothing else, and thus the Zero model has no censorship whatsoever. Even here, China has beat the West.... in terms of uncensored model etc
 
Top