R1 1776
A week ago, Perplexity released a version of DeepSeek R1 that is un-censored. The examples in their blog post are pretty compelling. I had previously been using R1 via OpenRouter, but was always uncomfortable with using such a censored model. 1776 likely still has a good amount of censorship in it (after all, they only targeted things known to be censored by CCP), but it’s clearly much better.
It’s good enough to become the third default slot in my Msty “Thinking” split chat, alongside OpenAI’s o1 and Google’s Gemini 2.0 Flash Thinking Experimental. (Msty itself is worth mentioning as an excellent UI for both remote and local models!)
How I interact with AI for all but the most basic of questions