As part of a University research study, I’ve been exploring a game engine design optimized for LLM collaboration, while still being easy to verify for anyone with basic programming skills. I think this will be a common pattern going forward.
Based on the research it avoids DSLs and custom editors, focusing instead on clarity, LLM throughput, and auditability. The result is surprisingly nice. Didn’t expect that. :)
Working on a research project for my old University I came across this masterpiece. True, and made me laugh. Thank you Alvinius et al.
“If you have poor data, it doesn’t matter if you are a world champion in qualitative analysis. The outcome will still be shit!" (Alvinius, 2023)
track by Matt Gray