About two days ago, when evervone started having problems, DeepSeek dearaded incredibly for me. While almost everyone else's proxies simply didn't work or returned errors, for me everything worked, only much less efficiently. I used deepseek-chat and was more than satisfied. Especially for how cheap it is.The responses were long and detailed. with character internal conflicts... the bot remembered everything, and the answers seemed organic and realistic the bot suggested interesting plot twists and thought processes, each response made me smile like wierdo and I was lowkey having a TIME OF MY LIFE. But then the responses became short GPT-like, and simply bad compared ta what they had been. I have a very long prompt that worked great for me, but now it seems to not be helping. I tried fiddling with the settings and temperature, but it only becomes worce. It's KILLING me. I don't know if this is related to the J.ai server itself or DeepSeek (perhaps both), but I'm worried... I thried switching yo the knew DeepSeek model V4-pro but this is so, so bad. It really is much worce compared to what i used it's crazy. And longer i chat the nore idiotic it becomes what🥲🥲🥲 I don't know what news there are about it. If you
have any info please share.