Something new to mix in: Research/Paper Database

[BradKML]

Sometimes, for some of the tasks, there might be similar research done, so could OpenEvolve capture papers and extract insights pre-emptively before making any moves? https://github.com/SakanaAI/AI-Scientist-v2

But also, there needs to be something like Perplexica that also captures non-standard knowledge formats, like webpages instead of papers or documentation. We don't need papers or articles as output, but a stockpile of curated resources. Some example repos

• https://github.com/ItzCrazyKns/Perplexica
• https://github.com/dzhng/deep-research
• https://github.com/assafelovic/gpt-researcher

The key issue wouldn't be just how tech can be hot-plugged, but that there needs to be synthesis of new information across tool interactions, and that even the original tool with bounded purposes do not cover that

This discussion is an extension of the previous one #13 (comment)

Bonus points if they can develop new RAG techniques https://github.com/FareedKhan-dev/all-rag-techniques/issues/12