2 pointsby Aedelon9 hours ago1 comment
  • Tattva075 hours ago
    "GLM-5 beats Claude Opus on the world’s hardest benchmark" This was true a couple weeks ago. The Z.ai models are going haywire recently. It might be bad tech, it might be malicious external actors, it might be both. Regardless, I wouldn't rely on China's ai tech just yet.