Perplexity is one of my favourite Large Language, but today, I thought they did something really cool.
Many of you may have heard about DeepSeek, a Chinese model that isn’t only incredibly powerful, but startlingly cheap. It only had one downside; it was clearly a propaganda tool.
For weeks, there were endless memes in the AI space showcasing people’s attempts trying crack the code, but the AI stood firm. Props, I guess.
To counter this, Perplexity have created R1 1776, a a replica of DeepSeek, with all its reasoning capability, but trained to remove any propaganda or CCP influence.
Three cheers!