I heartily agree, having spent months jankily recreating MRKL and ReAct on a much smaller scale before realizing those papers have existed for months already.

I heartily agree, having spent months jankily recreating MRKL and ReAct on a much smaller scale before realizing those papers have existed for months already. How can anyone keep up with the sheer volume of new papers and concepts here? Even Two Minute Papers is now lagging by two weeks.