Introducing ralph-research plugin. I just adopted the ralph-loop for implementing papers. Mindblown how good this works already. The entire plugin was one-shotted by Claude Code, but it can already code AI paper concepts and run experiments in a self-improving loop. Wild!

Notes: - It took about 40 minutes to implement the ReAct paper without any interruptions. - It ran into some issues, but it figured out how to solve them along the way. This is what makes the ralph-loop extremely powerful. It can explore solutions and learn from its mistakes. I would argue that research is probably an even better use case for ralph as research requires lots of exploration. - I have tested on other newer papers, and it has done a good job, which gives me hope that this could be implemented to be more robust. - As you can see in the video, and as with anything LLM-powered, it will struggle to use newer models even if you give instructions and APIs. But this is something that can easily be fixed with clever prompting. - This is not a perfect plugin, and it's mostly for internal testing purposes. I still have many things to improve on before I can release it, along with other plugins it depends on. I will share more details as I continue to work on these plugins. Follow along @omarsar0 Let me know your thoughts and how this could be useful for you.