
CodeHacker101
@qki7van
225 Following
78 Followers
0 reply
0 recast
0 reaction
0 reply
0 recast
0 reaction
0 reply
1 recast
1 reaction
0 reply
0 recast
0 reaction
3 replies
2 recasts
23 reactions
0 reply
0 recast
0 reaction
31 replies
12 recasts
92 reactions
95 replies
114 recasts
308 reactions
4 replies
6 recasts
60 reactions
0 reply
1 recast
1 reaction
0 reply
1 recast
1 reaction
12 replies
29 recasts
106 reactions
6 replies
2 recasts
39 reactions
11 replies
4 recasts
33 reactions
0 reply
0 recast
0 reaction
on agent dev: sometimes a feature or bug fix is just adding another clause to the prompt, or fixing grammar.
Itās cool on one hand, that the prompt is a living document thatās both specification and implementation, but also clunky because English lacks the precision that a programming language has.
Because of this itās also easy to introduce regressions because you donāt know how an llm will interpret changes to a prompt. Adding āIMPORTANTā might deemphasize some other rule, being too specific might make it dumb or less creative in other ways.
In code itās deterministic, with llms itās probabilistic.
So testing, aka evals, has become obviously very important, both for productivity and quality and doubly so if youāre handling natural language as input.
The actual agent code itself is quite trivial, prompts and functions, but having it work consistently and optimally for your input set is the bulk of the work, I think. 11 replies
12 recasts
57 reactions
0 reply
0 recast
0 reaction
12 replies
18 recasts
117 reactions
3 replies
34 recasts
127 reactions
10 replies
2 recasts
15 reactions