I am trying to improve my approach to Research Agents and could use some practical advice.
I am particularly interested in evaluation and performance. I would like to compare approaches that people have actually used rather than just theoretical suggestions.
What trade-offs have you found...
