Open source and under control: the DeepSeek paradox.

ChatGPT and co are also banned by many the same countries that banned DeepSeek and also most companies. Everyone either uses open source self-hosted models or they use internally produced ones. It's literally no different from how other could-only solutions with embedded tracking are treated.

Also stop comparing GPT-4 model with DeepSeek hosted service. Compare model to model or service to service.

Open source models can be uncensored (which is already the case for DeepSeek R1 model), proprietary cloud-only models can't.

So yes, we should really ask which "transparency" we should be seeking. A whole article written to justify desired result instead manages to prove the opposite correct. Good job Suzannah.

1 comments

ChatGPT and co are also banned by many the same countries that banned DeepSeek and also most companies. Everyone either uses open source self-hosted models or they use internally produced ones. It's literally no different from how other could-only solutions with embedded tracking are treated.
Also stop comparing GPT-4 model with DeepSeek hosted service. Compare model to model or service to service.
Open source models can be uncensored (which is already the case for DeepSeek R1 model), proprietary cloud-only models can't.
So yes, we should really ask which "transparency" we should be seeking. A whole article written to justify desired result instead manages to prove the opposite correct. Good job Suzannah.