I've been using DeepSeek Chat as an alternative to Western-centric AI tools, and while the analysis is quite interesting and up-to-date, I've run into some frustrating issues. I keep hitting dead-end loops when I try to discuss topics like literature or hypotheticals. The chatbot responds with things like, "Sorry, that's beyond my current scope. Let's talk about something else." I've even tried rephrasing my prompts based on the suggestions given by DeepSeek, but the same problems keep cropping up. Has anyone found effective workarounds for these limitations?
5 Answers
If you're using DeepSeek for serious work and not tapping into their API, you might be making things harder for yourself. As far as I know, the API only has censorship for harmful content and sensitive topics related to China. Honestly, the chatbot's limitations can be pretty ridiculous.
Keep in mind that the responses you're getting can be really limited by the way it's been censored. If you ask similar questions using R1 1776 on Labs.perplexity.com, you might have a better experience.
You might want to try accessing DeepSeek through OpenRouter. They use uncensored versions of the DeepSeek models, which might solve your problem.
Not sure if you think all Western models are completely uncensored. Just because there are some restrictions doesn’t mean they’re all the same.
Could you provide some specific examples? When you mention dead-end loops triggered by discussions on literature or hypotheticals, what exactly do you mean? Like, if I said, "My dinner sucks," how would that conversation go wrong? Should I clarify my thoughts more, like saying, "The dinner I had at the restaurant down the road was terrible because the steak was overcooked and it was noisy with kids running around"?
Is American censorship really more restrictive than others? I wonder how different it really is.