1,560 karma · joined May 18, 2016
While this is a pretty layman perspective, it helps me understand how to improve for now ok enough.
Pro tip: Don't just print the return code, also print the call and it's arguments that failed, even without a stack trace.
[0] https://iso25000.com/index.php/en/iso-25000-standards/iso-25...
Srsly. Welcome to my day job. I can see that the LLMs' center of training is so far off from where I'd need it, I can accelerate auxillary stuff but prompt never beats the weights and it constantly pulls back into it's middle...
Adding is always easier than reducing to the essence.
I find it interesting how herding agents has so much in common with being a team lead. Constant struggle between too detailed and too loose instructions. One difference though is that the team learns from you, but with agents it's only you who adapts. Saying that, because I don't count instructions or anything in the context window as adapting.