5 ms·

Lets hope so; computer use with AI is currently absolutely terrible. It is something I expected to see far larger progress in this year but it's no better than

• hypoxia 2y ago
  Yeah, +1. Looking back to the WebVoyager [1] and GPT4V generalist agent [2] papers from last January, it feels like we haven't come that far. But there are now several major technical unlocks - fine tuning for cursor locations (in Claude), better reasoning with o3, and RL fine-tuning so we can learn based on task success. That gives me significant hope. [1] https://arxiv.org/abs/2401.13919 https://arxiv.org/abs/2401.13919 [2] https://arxiv.org/abs/2401.01614 https://arxiv.org/abs/2401.01614