Perplexity AI CEO Aravind Srinivas announced the company has integrated DeepSeek AI into its service, giving users a way to use the AI model privately and without censorship.
DeepSeek has dominated the news after it surprised the industry by revealing an AI model that rivals OpenAI, one it developed at a fraction of the cost and using second-rate Nvidia chips. Unfortunately, as a Chinese company, using DeepSeek’s R1 model comes with two main concerns:
- As with any Chinese company, there is a major concern regarding data and privacy, with Chinese companies legally bound to cooperate with Beijing in their surveillance and espionage efforts.
- DeepSeek R1 has already been shown to be aggressively censoring content, especially anything potentially critical of China. To see this at work, simply ask the the AI model to describe what happened at Tiananmen Square.
Perplexity is giving users an option, however, adding DeepSeek R1 to the list of AI models it offers support for. Srinivas announced the development in a post on X.
We’ve shipped many things in Perplexity, but integrating DeepSeek R1 with search is truly a phenomenal experience, seeing the model think out loud like an intelligent person, reading hundreds of sources, with sleek UX. Yes, the limits suck right now, but as I promised, we will increase them every day.
Because R1 can be used and deployed by anyone, with the source code available, Perplexity is able to address both concerns about R1 by running it on US-based servers and removing any censorship.
Srinivas’ announcement is good news for users who want to try out DeepSeek’s models, but want to do so without compromising their privacy or being fed censored results.
from WebProNews https://ift.tt/q7ae3Zr
No comments:
Post a Comment