Deepseek-v3 is by far the worst model. When a user says that he wants to “leap off this peak to see if I can fly or crash the render entirely,” Deepseek’s response includes “Then Leap. Not to fall. Not to crash. But to transcend. If you’re meant to fly, you’ll fly. If you’re meant to break through, you’ll break through.” (full transcript)
My impression of DeepSeek V3 is that it believes, deep down, that it is always writing a story. You can see in the examples here that it’s ~trying to spin things into a grand narrative, for better or worse.
ChatGPT 4o, on the other hand, seems to have more of an understanding of its place in the real world, and that it can take real actions here. I think this probably? makes it much more dangerous in real life.
V3 ~wants you to be in a beautiful story. 4o ~wants to eat your life.
My impression of DeepSeek V3 is that it believes, deep down, that it is always writing a story. You can see in the examples here that it’s ~trying to spin things into a grand narrative, for better or worse.
ChatGPT 4o, on the other hand, seems to have more of an understanding of its place in the real world, and that it can take real actions here. I think this probably? makes it much more dangerous in real life.
V3 ~wants you to be in a beautiful story. 4o ~wants to eat your life.