3 ms·

Genuine question: How safe is it to use Chinese models via their services? Surely Anthropic and OpenAI are ingesting what I push there as well, but they're at l

by abc42 4mo ago

Genuine question: How safe is it to use Chinese models via their services? Surely Anthropic and OpenAI are ingesting what I push there as well, but they're at least vaguely allied with my home country geopolitically. China on the other hand seems to be interested in supporting countries like Iran and Russia.

andai 4mo ago
[dead]
teyopi 4mo ago
What does China do that US does not? They are releasing open models, so at-least up until now their advancements you can run yourself. US frontier labs on the other hand keep it all to themselves. The moment they cut access you have nothing and your country will be stumped on and forced in making decisions not in your national interest.
- abc42 4mo ago
  >What does China do that US does not? Support the enemies of my country, most probably. With Trump, this has admittedly become a bit more non-obvious, but I think it's mainly still so.
- shostack 4mo ago
  This is literal whataboutism. Refute the comment if you have a valid point to make.
  - teyopi 4mo ago
    Erm, no? it is pointing out hypocrisy? If you are against the thing China does, when US does the thing also be against it. US has always done what China does, now trump is doing it vocally. So it is easier to point out the hypocrisy in it. Before there was plausible deniability. Thank you for your attention to this matter
    - abc42 4mo ago
      Well, it's more complicated than that. Let's imagine that China and USA do some equally bad thing against my privacy, let's say intercept and store all the traffic between myself and my AI provider. I am inclined to excuse it from USA and not from China because it's much more likely that China would use it to undermine the security of my home country. USA on the other hand, is literally a military ally of my home country. I don't naively think that USA is a 100% benign country because of the alliance and that my interests are aligned with them. Trump especially has made that part worse. They're just more aligned with me than China.
Alifatisk 4mo ago
I suggest to not look at how each company is expressing themselves on the media, look at how they are actually behaving. When I first tried out Z.ai last year, I too was concerned regarding where my data goes. I vaguely remember from their ToS (please verify yourself too) that they followed a zero data retention policy for its AI services. This of course applied to their paying customers. I do not know if it applies to their free users on chat.z.ai. I am at the stage of not caring about geopolitics at all, I've lost a lot of trust in the current US administration, they have burned a lot of international relationships thanks to the orange man. This is while Chinese companies and startups have almost closed the gap between open-weight models and proprietary models. Its thanks to their actions and open contribution in the research I can run local models at home. Just read the tweet posted and see where Z.ai is trying to position themselves, this competition benefits us (consumers) a lot!
- droidjj 4mo ago
  I wouldn’t trust Chinese labs’ TOS very much. It would be incredibly difficult and expensive to enforce any of those terms. edit: this is a comment about suing and enforcing judgments against Chinese companies in the US, especially software companies, not necessarily about how trustworthy the Chinese labs are.
MiscIdeaMaker99 4mo ago
I don't know what you consider to be safe, but I assume you're talking about privacy and data retention. You can pay someone like Ollama, which is US-based, which runs those models in their cloud. You can also download many/most of the models to run them locally on your own computer, although the performance depends entirely on your hardware capabilities. Ollama has their own policies about privacy and data retention here: https://ollama.com/privacy https://ollama.com/privacy That's just one third-party doing this, and there are many others.

by abc42 4mo ago

andai 4mo ago

[dead]

teyopi 4mo ago

What does China do that US does not? They are releasing open models, so at-least up until now their advancements you can run yourself. US frontier labs on the other hand keep it all to themselves. The moment they cut access you have nothing and your country will be stumped on and forced in making decisions not in your national interest.

abc42 4mo ago
>What does China do that US does not? Support the enemies of my country, most probably. With Trump, this has admittedly become a bit more non-obvious, but I think it's mainly still so.
shostack 4mo ago
This is literal whataboutism. Refute the comment if you have a valid point to make.
- teyopi 4mo ago
  Erm, no? it is pointing out hypocrisy? If you are against the thing China does, when US does the thing also be against it. US has always done what China does, now trump is doing it vocally. So it is easier to point out the hypocrisy in it. Before there was plausible deniability. Thank you for your attention to this matter
  - abc42 4mo ago
    Well, it's more complicated than that. Let's imagine that China and USA do some equally bad thing against my privacy, let's say intercept and store all the traffic between myself and my AI provider. I am inclined to excuse it from USA and not from China because it's much more likely that China would use it to undermine the security of my home country. USA on the other hand, is literally a military ally of my home country. I don't naively think that USA is a 100% benign country because of the alliance and that my interests are aligned with them. Trump especially has made that part worse. They're just more aligned with me than China.

Alifatisk 4mo ago

I suggest to not look at how each company is expressing themselves on the media, look at how they are actually behaving. When I first tried out Z.ai last year, I too was concerned regarding where my data goes. I vaguely remember from their ToS (please verify yourself too) that they followed a zero data retention policy for its AI services. This of course applied to their paying customers. I do not know if it applies to their free users on chat.z.ai. I am at the stage of not caring about geopolitics at all, I've lost a lot of trust in the current US administration, they have burned a lot of international relationships thanks to the orange man. This is while Chinese companies and startups have almost closed the gap between open-weight models and proprietary models. Its thanks to their actions and open contribution in the research I can run local models at home. Just read the tweet posted and see where Z.ai is trying to position themselves, this competition benefits us (consumers) a lot!

droidjj 4mo ago
I wouldn’t trust Chinese labs’ TOS very much. It would be incredibly difficult and expensive to enforce any of those terms. edit: this is a comment about suing and enforcing judgments against Chinese companies in the US, especially software companies, not necessarily about how trustworthy the Chinese labs are.

MiscIdeaMaker99 4mo ago

I don't know what you consider to be safe, but I assume you're talking about privacy and data retention. You can pay someone like Ollama, which is US-based, which runs those models in their cloud. You can also download many/most of the models to run them locally on your own computer, although the performance depends entirely on your hardware capabilities. Ollama has their own policies about privacy and data retention here: https://ollama.com/privacy https://ollama.com/privacy That's just one third-party doing this, and there are many others.