TL;DR
- Perplexity integrates DeepSeek's R1 model into its platform, enhancing AI-powered research capabilities.
- This comes after DeepSeek's rapid rise in popularity and challenges from cyberattacks.
- To address security concerns, Perplexity assures users that DeepSeek operates under strict data protection measures in the U.S. and Europe.
- The development reflects the increasing demand for advanced generative AI amidst geopolitical tensions.
Perplexity has announced the integration of DeepSeek’s R1 model alongside OpenAI’s o1 into its platform as part of its Pro subscription.
This development comes on the heels of a triumphant yet tumultuous period for DeepSeek, which made waves by climbing to the top of U.S. iPhone charts and subsequently faced cyberattack challenges.
Perplexity’s decision to add DeepSeek R1 to its platform reflects the growing demand for cutting-edge generative AI models that can provide thorough insights and research capabilities, especially in complex fields.
As the announcement noted, users will experience an enhanced Pro Search reasoning mode, offering a transparent chain of thought into the model’s reasoning process — similar to what we see on OpenAI’s native ChatGPT interface.
DeepSeek’s rise has been nothing short of meteoric.
Just yesterday, it surged to the top of the U.S. iPhone charts, pushing ChatGPT down the list.
This shakeup in the generative AI field has not gone unnoticed, particularly given the stringent U.S. export controls aiming to curb such advancements from China.
DeepSeek is made by an obscure Chinese startup, a fact that is not lost on the world amid these geopolitical tensions, with worries (and memes) over data security.
These concerns have led Perplexity CEO Aravind Srinivas to publicly clarify.
“All DeepSeek usage in Perplexity is through models hosted in centers in the USA and Europe,” Srinivas said in a X post.
“DeepSeek is open-source. None of your data goes to China.”
“We don’t even need to use their chat interface or their API,” Srinivas further explored the topic in a X space. “The weights are all out there, you could use Hugging Face.”
DeepSeek has trained its model using Nvidia’s H800 chips at a fraction of the cost typically required, an achievement particularly noteworthy against the backdrop of international trade restrictions.
Despite these achievements, DeepSeek has faced significant hurdles. Reports of ‘large-scale malicious attacks’ forced the company to suspend new signups globally on Tuesday, highlighting the challenges that come with rapid growth and popularity.
Perplexity is valued at $9 billion and primarily aims to serve the use-case of doing AI-based research using citations.
It relies on models developed by others, including OpenAI, Anthropic and now — DeepSeek, as part of its own product, essentially serving as a competitor to Google.
The company is backed by the likes of Amazon and has alleged that Google is engaging in unfair practices to stifle competition.
Planning your financial journey can be daunting but it doesn't have to be. Fire Fast by Dzambhala helps you understand and plan effectively.
Join the vibrant privacy-ensured Dzambhala community on
Want to give feedback on this story? Write to us at [email protected]