vaishak2future

4 ms·

1. 1.▲
   by vaishak2future 18d ago

   I spent way too much time on this. Super engaging! Great job!
2. 2.▲
   by vaishak2future 4mo ago

   Amazing
3. 3.▲
   Multimodal Coding Agents as In-Context Policy Learners for Robot Manipulation(arxiv.org)

   1 points by vaishak2future 7mo ago|1 comments
4. 4.▲
   by vaishak2future 7mo ago

   Author here. We’ve reached the point where frontier models (along with agent harnesses) can perform verbal reinforcement learning over multimodal data to reason over failures and correct policies expressed as code to fix them in a targeted
5. 5.▲
   by vaishak2future 13y ago

   Wow